Encyclopedia of Electronic Circuits by badisoldschool

VIEWS: 111 PAGES: 1268

Encyclopedia of Electronic Circuits

More Info
									T HE N ETWORK   P RESS

                   S ECOND        E DITION

                    W ERNER      F EIBEL

                    I SSUES ,
                    AND T ERMS

                    C OVERS N ET W ARE 4.1, W INDOWS 95,
                    W INDOWS NT S ERVER 3.51, U NIX W ARE 2,
                    AND OS/2 W ARP C ONNECT

                    F ULL T EXT ON CD-ROM FOR Q UICK
                                   E LECTRONIC R EFERENCE
The Encyclopedia
of Networking
The Encyclopedia
of Networking
Second Edition
The First Edition of this
book was published under
the title Novell’s® Complete
Encyclopedia of Networking

Werner Feibel

         San Francisco   s   Paris   s   Düsseldorf   s   Soest
Acquisitions Editor: Kristine Plachy
Developmental Editor: Guy Hart-Davis
Editors: Kristen Vanberg-Wolff and Maureen Adams
Technical Editor: Mary Madden
Book Designer: Seventeenth Street Studios
Technical Illustrators: Cuong Le, Heather Lewis, and Alan Smith
Desktop Publisher: London Road Design
Production Coordinator: Nathan Johanson
Indexer: Matthew Spence
Cover Designer: Archer Design
Cover Photographer: Dewitt Jones

SYBEX is a registered trademark of SYBEX Inc.
Network Press and the Network Press logo are trademarks of SYBEX Inc.

TRADEMARKS: SYBEX has attempted throughout this book to distinguish proprietary trademarks from
descriptive terms by following the capitalization style used by the manufacturer.

Every effort has been made to supply complete and accurate information. However, SYBEX assumes no
responsibility for its use, nor for any infringement of the intellectual property rights of third parties which
would result from such use.

The first edition of this book was published under the title Novell’s ® Complete Encyclopedia of Networking
©1995 SYBEX Inc.

Copyright ©1996 SYBEX Inc., 2021 Challenger Drive, Alameda, CA 94501. World rights reserved. No part
of this publication may be stored in a retrieval system, transmitted, or reproduced in any way, including but
not limited to photocopy, photograph, magnetic or other record, without the prior agreement and written
permission of the publisher.

Library of Congress Card Number: 95-72476

ISBN: 0-7821-1829-1

Manufactured in the United States of America

10 9 8 7 6 5 4 3 2 1
SYBEX warrants the enclosed CD-ROM to be free of physical defects for a period of ninety (90) days after
purchase. If you discover a defect in the CD during this warranty period, you can obtain a replacement CD at
no charge by sending the defective CD, postage prepaid, with proof of purchase to:

    SYBEX Inc.
    Customer Service Department
    2021 Challenger Drive
    Alameda, CA 94501
    (800) 227-2346
    Fax: (510) 523-2373

After the 90-day period, you can obtain a replacement CD by sending us the defective CD, proof of purchase,
and a check or money order for $10, payable to SYBEX.

SYBEX makes no warranty or representation, either express or implied, with respect to this medium or its
contents, quality, performance, merchantability, or fitness for a particular purpose. In no event will SYBEX,
its distributors, or dealers be liable for direct, indirect, special, incidental, or consequential damages arising
out of the use of or inability to use the software even if advised of the possibility of such damage.

The exclusion of implied warranties is not permitted by some states. Therefore, the above exclusion may not
apply to you. This warranty provides you with specific legal rights; there may be other rights that you may
have that vary from state to state.

 Copy Protection
None of the material on the CD is copy-protected. However, in all cases, reselling or making copies of these
programs without authorization is expressly forbidden.
[In] a certain Chinese encyclopedia…it is written that animals are divided into:

(a)   those belonging to the Emperor
(b)   those that are embalmed
(c)   tame ones
(d)   suckling pigs
(e)   sirens
(f)   fabulous ones
(g)   stray dogs
(h)   those included in the present classification
(i)   those that tremble as if mad
(j)   innumerable ones
(k)   those drawn with a very fine camelhair brush
(l)   others
(m)   those that have just broken the water pitcher
(n)   those that look like flies from a long way off

  Jorge Luis Borges
  As with the first edition, this book would never have been completed without the help of
  many people. These people deserve thanks for all their efforts and energy. Guy Hart-Davis
  convinced me that it was time for a revision and set me to work. Several people did splen-
  did work during the production process: Kris Vanberg-Wolff, a veteran of the first edition,
  worked on the revision until her planned departure for the calmer (and tastier) world of
  cooking school. Maureen Adams, Laura Arendal, and Nathan Johanson took over the pro-
  duction chores after Kris left. They did an excellent job, especially considering the short
  notice and even shorter revision schedule. My heartfelt thanks to all these folks.
     Mary Madden’s technical reviews were always full of gentle, constructive corrections
  and useful suggestions for improvements. Although I may not have been smart enough to
  act on all of them, the suggestions have improved the book immensely—for which I’m very
     Kris Vanberg-Wolff’s eagle eyes and infallible grammatical sense found and fixed my
  awkward phrasings, stylistic inconsistencies, and grammatical aberrations. I shudder to
  think what the book would have looked like without the benefit of these efforts.
     As always, I’m very grateful to all the people who worked between and behind the
  scenes to make this book, and also to those who created the compact disc. Thanks also to
  the many people who sent me information about their products and who took the time
  to answer my questions.
     Finally, I dedicate this book to my wife Luanne and my daughter Molly—for all the joy
  and fun they provide, during both work and play hours.
Table of Contents
      Introduction                                      ix
      Entries (Listed Alphabetically)                    1
      Appendix A: Acronyms and Abbreviations         1113
      Appendix B: Bibliography and Other Resources   1235
      Index                                          1251

     What You’ll Find in This Book
               As in the first edition, I’ve tried to make this Encyclopedia a comprehensive source of informa-
               tion about matters relating to networking. I’ve also tried to present the information in a clear
               and useful manner.
                  This book contains comprehensive, straightforward summaries of the major concepts,
               issues, and approaches related to networking. Networking is defined broadly to encompass
               configurations ranging from a couple of connected computers just a few feet apart to a network
               of several thousand machines (of all types and sizes) scattered around the world. You’ll find
               discussions of networking as it’s done by servers and clients, managers and agents, peers, and
               even over the telephone.
                  You probably won’t find anything here that you can’t find in other places. However, I don’t
               know of any other book or source that collects so much network-related information in one
               place. To find all the information summarized here, you would need to check hundreds of
               books, disks, articles, Web pages, or other documents.
                  Despite its hefty size, this encyclopedia just scratches the surface of what there is to know
               about networking. After all, how complete can any book be if just the World Wide Web on the
               Internet has over 10 million hypertext documents. I do think, however, that this book scratches
               deeper than most other references you’ll find.
                  This revised edition updates entries for concepts and technologies that change rapidly or
               where there have been major developments. I’ve also added considerable material about the
               Internet (and especially about the World Wide Web), since interest in this networking phenom-
               enon is growing at an astounding pace.

Concepts, Not Instructions

               As in the first edition, I’ve tried to cover concepts rather than making this a how-to book.
               Thus, you won’t learn how to install networks or run specific programs. However, you will
               learn about different types of programs and what they do. For example, you can read about
               browsers and how they make exploring the World Wide Web possible; you’ll also learn about
               programs such as network operating systems and how they differ from ordinary operating
x         Introduction

An Anchor in an Ocean of Words

    This book was obsolete from the moment it was written. That’s because nothing changes faster
    than vocabulary in a field where there is money to be made. Since major breakthroughs and
    advances are still happening in the area of networking, there are new network-related words
    and concepts to be found in almost every issue of every computer magazine. If you include
    acronyms and abbreviations, the speed with which the vernacular expands is even faster. For
    example, the first edition of this book was published under a year ago and it was no trouble
    finding almost 2,000 new entries for Appendix A.
       Given the futility of even trying to stay completely up-to-date, I’ve chosen to focus on the
    more enduring concepts and facts—those that provide the foundations and background that
    underlie the constantly changing terminology. This makes the Encyclopedia more generally
    useful and enduring.

Helping the Book Grow

    While core networking concepts change very little, the core does grow. For example, ten years
    ago there was much less need to know about wireless communications because there were
    fewer wireless products, as well as less public interest in the technology. Because of such
    progress, the body of essential fundamentals grows with each year.
        I expect to update and add to the material in the book, and hope to make the Encyclopedia
    always effective, comprehensive, and useful. Fortunately, an electronic medium makes it easier
    to grow in this way.
        If you need to find out something about networking, look for it in this book. If you find an
    entry for the topic, we hope you’ll be more informed after you’ve read it. On the other hand,
    if you can’t find the information you need, didn’t understand it, or don’t think you learned
    what you should have, please drop us a line and tell us.
        Also, if there are concepts or terms you would like to see included, please let us know. If you
    can provide references, that would be helpful. Even under the best of circumstances, there’s lit-
    tle chance that you’ll get a reply to individual queries. However, we will read your comments
    and suggestions and will try to use them to improve future versions of the book.
Symbols &

     2              & (Ampersand)

Symbols & Numbers

w                                                           w
    & (Ampersand)                                           @ (At sign)
             The ampersand is used to indicate special          The at sign is used to separate the username
             characters in HTML (Hypertext Markup               from domain specifiers in e-mail addresses.
             Language) documents—that is, documents             For example, mels@golemxiv.mit.edu
             for the World Wide Web. For example,               would indicate someone with username mels
             & specifies the ampersand character             on a computer named golemxiv at MIT.
             (&); ö specifies a lowercase o with        w
             an umlaut, or dieresis, mark (ö).              \ (Backslash)
w                                                               In some operating systems, such as DOS,
    < > (Angle Brackets)                                        OS/2, and NetWare, the backslash character
             Angle brackets are used in pairs to surround       separates directory names or directory and
             markup tags in HTML (Hypertext Markup              file names in a path statement. By itself, the
             Language) documents for the World Wide             backslash represents the root directory in
             Web. For example, <P> indicates a para-            these operating systems.
             graph break; <B> and </B> indicate the start          In various programming and editing con-
             and end of a section that is to be displayed       texts, the backslash is used to escape the
             in boldface.                                       character that follows. For example, \n is an
                                                                escape code to indicate a newline character
w                                                               in many operating environments.
    * (Asterisk)
             In several operating systems, the asterisk
                                                            // (Double Slash)
             serves as a wildcard character: to represent
             one or more characters, such as in a file           In URLs (Uniform Resource Locators), dou-
             name or extension. For example, a* matches         ble slash characters separate the protocol
             act, actor, and and, but not band.                 from the site and document names. For
                In pattern matching involving regular           example, if it existed,
             expressions, the asterisk matches the occur-         http://examplehost.ucsc.edu/
             rences of the single character immediately           filename.html
             preceding it. For example, ba*th matches
             bth, bath, and baaaaath, but not bbath.            would refer to a file named filename.html
                In e-mail and in other contexts that use        residing on the examplehost machine at the
             plain text, asterisks are sometimes used           University of California at Santa Cruz. To
             around words or phrases to indicate em-            get to this file, you would use a server that
             phasis. For example, “I *really* want              supports the HTTP (Hypertext Transport
             to emphasize the second word in this               Protocol).
                                                                              4B/5B Encoding        3

w                                                       events that have helped define the computer
µ(Mu)                                                   culture:
    Used as an abbreviation for the prefix micro,           http://www.phil.uni-sb.de/fun/jargon/
    as in µsec for microsecond and µm for                  index.html
    micrometer. This order of magnitude corre-          In this URL, the file is named index.html,
    sponds to 2−20, which is roughly 10−6, or           and it is located in the /fun/jargon directory
    one-millionth.                                      on a machine in Germany (de).
    SEE ALSO                                               In other operating systems, such as DOS,
      Order of Magnitude                                OS/2, and NetWare, a slash is sometimes
                                                        used to indicate or separate command line
w                                                       switches or options for a command.
. and .. (Period and Double Period)
    In hierarchically organized directory sys-
    tems, such as those used by UNIX, DOS,
    and OS/2, . and .. refer to the current and         The IEEE 802.3 committee’s designation
    the parent directories, respectively. In pat-       for an Ethernet network that operates at
    tern matching involving regular expressions,        1 megabit per second (Mbps) and that
    the . matches any single character, except a        uses unshielded twisted-pair (UTP) cable.
    newline character.                                  This configuration uses a physical bus,
                                                        with nodes attached to a common cable.
w                                                       AT&T’s StarLAN is an example of a 1Base5
? (Question Mark)                                       network.
    In many operating systems, a question mark
                                                        SEE ALSO
    serves as a wildcard character that repre-
                                                          10BaseX; 10Broad36
    sents a single character, such as in a file or
    directory name.                                 w
                                                    4B/5B Encoding
/ (Slash)                                               4B/5B encoding is a data-translation scheme
                                                        that serves as a preliminary to signal encod-
    The slash (also known as a forward slash or
                                                        ing in FDDI (Fiber Distributed Data Inter-
    a virgule) separates directory levels in some
                                                        face) networks. In 4B/5B, every group of
    operating systems (most notably UNIX), in
                                                        four bits is represented as a five-bit symbol.
    addresses for gopher, and in URLs (Uniform
                                                        This symbol is associated with a bit pattern
    Resource Locators). For example, the fol-
                                                        that is then encoded using a standard signal-
    lowing URL specifies the name and location
                                                        encoding method, usually NRZI (non-return
    of a hypertext version of the jargon file,
                                                        to zero inverted).
    which contains definitions for terms and
                                                           This preprocessing makes the subsequent
                                                        electrical encoding 80 percent efficient. For
4         5B/6B Encoding

    example, using 4B/5B encoding, you can          w
    achieve a 100 megabit per second (Mbps)         10BaseX
    transmission rate with a clock speed of only        The designations 10Base2, 10Base5,
    125 megahertz (MHz).                                10BaseF, and 10BaseT refer to various
       In contrast, the Manchester signal-              types of baseband Ethernet networks.
    encoding method, which is used in Ethernet
    and other types of networks, is only 50 per-    10Base2
    cent efficient. For example, to achieve a 100
                                                        10Base2 uses thin coaxial cable. This ver-
    Mbps rate with Manchester encoding, you
                                                        sion can operate at up to 10 megabits per
    need a 200 MHz clock speed.
                                                        second (Mbps) and can support cable seg-
w                                                       ments of up to 185 meters (607 feet). It is
5B/6B Encoding                                          also known as thin Ethernet, ThinNet, or
    A data-translation scheme that serves               CheaperNet, because thin coaxial cable is
    as a preliminary to signal encoding in              considerably less expensive than the thick
    100BaseVG networks. In 5B/6B, every                 coaxial cable used in 10Base5 networks.
    group of five bits is represented as a six-bit   10Base5
    symbol. This symbol is associated with a bit
    pattern that is then encoded using a stan-          10Base5 uses thick coaxial cable. This ver-
    dard signal-encoding method, such as NRZ            sion is the original Ethernet. It can operate
    (non-return to zero).                               at up to 10 Mbps and support cable seg-
                                                        ments of up to 500 meters (1,640 feet). It is
8B/10B Encoding                                         also known as thick Ethernet or ThickNet.

    A data-translation scheme related to 4B/5B      10BaseF
    encoding that recodes eight-bit patterns into
                                                        10BaseF is a baseband 802.3-based Ethernet
    10-bit symbols. 8B/10B encoding is used, for
                                                        network that uses fiber-optic cable. This
    example, in IBM’s SNA (Systems Network
                                                        version can operate at up to 10 Mbps.
    Architecture) networks.
                                                           Standards for the following special-
w                                                       purpose versions of 10BaseF are being
9-Track Tape                                            formulated by the IEEE 802.3:
    A tape storage format that uses nine parallel         10BaseFP (fiber passive): For desktops
    tracks on 1/2-inch, reel-to-reel magnetic
    tape. Eight tracks are used for data, and one         10BaseFL (fiber link): For intermediate
    track is used for parity information. These             hubs and workgroups
    tapes are often used as backup systems on             10BaseFB (fiber backbone): For central
    minicomputer and mainframe systems; digi-               facility lines between buildings
    tal audio tapes (DATs) are more common on
                                                                  66-Type Punch-Down Block          5

10BaseT                                                 cables for each direction, so that each cable
                                                        needs only an 18 MHz bandwidth.
    10BaseT is a baseband 802.3-based Ethernet
    network that uses unshielded twisted-pair           B RO A D E R C A T E G O R I E S

    (UTP) cable and a star topology. This ver-             Ethernet; Network, Broadband
    sion can operate at up to 10 Mbps. It is also       SEE ALSO
    known as twisted-pair Ethernet or UTP                  1Base5; 10BaseX
    B RO A D E R C A T E G O R Y                    56K Line
                                                        A digital telephone circuit with a 64 Kbps
    SEE ALSO                                            bandwidth, but with a bandwidth of only
       1Base5; 10Broad36; 100BaseT                      56 Kbps data, with the other 8 Kbps being
                                                        used for signaling. Also known as an ADN
                                                        (Advanced Digital Network) or a DDS
                                                        (Dataphone Digital Service) line.
    10Broad36 is a broadband, 802.3-based,
    Ethernet network that uses 75-ohm coaxial       w

    (CATV) cable and a bus or tree topology.
                                                    64K Line
    This version can operate at up to 10 mega-          A digital telephone circuit with a 64 Kbps
    bits per second (Mbps) and support cable            bandwidth. Also known as a DS0 (digital
    segments of up to 1,800 meters (about               signal, level 0) line. When the entire 64 Kbps
    6,000 feet).                                        are allocated for the data, the circuit is
       A 10Broad36 network uses differential            known as a clear channel. This is in contrast
    phase shift keying (DPSK) to convert the            to a circuit in which 8 Kbps are used for
    data to analog form for transmission.               signaling, leaving only 56 Kbps for data.
    Because of the encoding details, a
    10Broad36 network actually needs                66-Type Punch-Down Block
    18 megahertz (MHz) for each channel:
    14 MHz to encode the 10 Mbps signal and             A device for terminating wires, with the
    4 MHz more for collision detection and              possibility of connecting input and output
    reporting capabilities.                             wires. This type of punch-down block can
       In a 10Broad36 network, throughput is            handle wires with up to 25 twisted pairs.
    10 Mbps in each direction—that is, a total          The 66-type have generally been superseded
    bandwidth of 36 MHz is needed. This band-           by 110-type punch-down blocks.
    width can be provided in a single cable or in       SEE ALSO
    two separate cables. A split-cable approach            Punch-Down Block
    uses half the cable for each direction, which
    means the cable must have a 36 MHz band-
    width. A dual-cable approach uses separate
6         100BaseFX

w                                                     The main differences between fast (100
100BaseFX                                           Mbps) Ethernet and standard (10 Mbps)
    A 100BaseT basal type variant that runs         Ethernet are:
    over multimode fiber-optic cable. Nodes on         s   A 100BaseT Ethernet allows a much
    a 100BaseFX network can be up to 2 kilo-              shorter gap between signals.
    meters apart. This variant is also written
    100Base-FX.                                       s   A 100BaseT Ethernet requires either
                                                          higher-grade cable or more wire pairs.
                                                          It can run at 100 Mbps speeds on
      100BaseT                                            Category 3 or 4 cable—provided four
    C O M P A RE                                          pairs are available; Category 5 cable
      100BaseT4; 100BaseTX                                requires only two pairs.
                                                      s   Currently, a 100BaseT Ethernet can
100BaseT                                                  support a network that is only about
                                                          a tenth of the length allowed for an
    The general name for any of three 100 Mbps
                                                          ordinary Ethernet network. For net-
    Ethernet variants that have just been made a
                                                          works that use copper (as opposed to
    standard by an IEEE 802.3 subcommittee
                                                          fiber-optic) cabling: Two nodes of a
    (802.3u). 100BaseT Ethernet is one of the
                                                          100BaseT4 network can be no further
    candidates trying to become the standard
                                                          apart than 205 meters—regardless of
    100 Mbps Ethernet. This version was devel-
                                                          whether the nodes are next to each
    oped and proposed originally by Grand
    Junction, in collaboration with several other
    corporations.                                     The following variants of 100BaseT
       The term fast Ethernet is often used for     Ethernet have been defined:
    this version. This is unfortunate, since that
                                                      100BaseFX: Runs over multimode fiber-
    term is also used to refer to any Ethernet
                                                        optic cable. Nodes on a 100BaseFX
    implementation that supports speeds faster
                                                        network can be up to two kilometers
    than the official 10 Mbps standard. To add
    to the confusing terminology, a software
    product (no longer available) was also            100BaseTX: Uses two wire pairs,
    named fastEthernet.                                 but requires Category 5 unshielded
       100BaseT Ethernet retains Ethernet’s             or shielded twisted pair (UTP or
    CSMA/CD (Carrier Sense Multiple Access/             STP) wire.
    Collision Detect) media access method—in
                                                      100BaseT4: Can use category 3, 4, or 5
    contrast to the 100BaseVG variant (now
                                                        UTP cable. The T4 in the name comes
    officially, IEEE 802.12)—which is the other
                                                        from the fact that four wire pairs are
    major 100 Mbps Ethernet available.
                                                        needed: two for sending and two for
                                                                                 100BaseVG         7

       In some configurations, fast and ordinary         SEE
    Ethernet nodes can share the same network.            100BaseT
    Fast Ethernet devices identify themselves as
                                                        C O M P A RE
    such by sending a series of FLPs (fast link
                                                          100BaseT4; 100BaseFX
    pulses) at startup.
    P R I M A R Y S O U RC E S
       IEEE 802.3u committee publications
                                                        100BaseVG is a version of Ethernet devel-
    B RO A D E R C A T E G O R I E S                    oped by Hewlett-Packard (HP) and AT&T
       Ethernet                                         Microelectronics, and is currently under
    C O M P A RE
                                                        consideration by an IEEE 802.12 committee.
                                                        It is an extension of 10BaseT Ethernet that
                                                        will support transmissions of up to 100
w                                                       megabits per second (Mbps) over voice-
100BaseT4                                               grade (Category 3) twisted-pair wire. The
    A 100BaseT Ethernet variant that can use            VG in the name stands for voice-grade.
    category 3, 4, or 5 unshielded twisted pair
                                                    Differences from 10 Mbps Ethernet
    (UTP) cable. The T4 means that four wire
    pairs are needed: two for sending and two           100BaseVG Ethernet differs from ordinary
    for receiving. Two nodes of a 100BaseT4             (10 Mbps) Ethernet in the following ways:
    network can be no further apart than 205
                                                          s   Uses demand priority (rather than
    meters, regardless of whether the nodes are
                                                              CSMA/CD) as the media access
    next to each other. This variant is sometimes
    written 100Base-T4.
                                                          s   Can use ordinary (Category 3)
                                                              unshielded twisted-pair (UTP) cable,
                                                              provided that the cable has at least
    C O M P A RE                                              four wire pairs. Ordinary Ethernet
       100BaseTX; 100BaseFX                                   needs only two pairs: one to send and
                                                              one to receive.
100BaseTX                                                 s   Uses quartet signaling to provide four
    A 100BaseT Ethernet variant that uses two                 transmission channels (wire pairs)
    wire pairs, but requires Category 5 UTP or                instead of just one. All wire pairs are
    STP wire. Two nodes of a 100BaseTX net-                   used in the same direction at a given
    work can be no further apart than 205                     time.
    meters—regardless of whether the nodes are            s   Uses the more efficient 5B/6B NRZ
    next to each other. This variant is sometimes             signal encoding, as opposed to the
    written 100Base-TX.
8         100BaseX

          Manchester encoding scheme used by        100BaseVG/AnyLAN
          ordinary Ethernet.
                                                        100BaseVG/AnyLAN is an extension
      s   For category 3 cable, a VG network            of 100BaseVG, developed as a joint effort
          can be at most 600 meters from end to         between Hewlett-Packard and IBM. This
          end—and only 200 meters if all hubs           version also supports the Token Ring archi-
          in the network are connected in the           tecture, and it can be used with either Ether-
          same wiring closet. These values              net or Token Ring cards (but not both at the
          increase by 50%—that is, to 900 and           same time or in the same network). Because
          300 meters, respectively—when cate-           the demand priority access method can be
          gory 5 cable is used. For VG using            deterministic, the 100BaseVG/AnyLAN
          fiber-optic cable, the most widely sepa-       architecture could handle isochronous
          rated network nodes can be up to              data—that is, data (such as voice or video)
          5000 meters, or 5 kilometers, apart.          that requires a constant transmission rate.
Upgrading to 100BaseVG                                     The 100VG-AnyLAN Forum is the advo-
                                                        cacy group for this Ethernet variant. This
    100BaseVG is designed to provide an easy            consortium includes over 20 members,
    upgrade path from 10 Mbps Ethernet. An              including Apple, Compaq, and IBM.
    upgrade requires two new components:                100Base VG/AnyLAN is also known simply
                                                        as VG or AnyLAN.
      s   A 100BaseVG network interface card
          (NIC) for each node being upgraded.           B RO A D E R C A T E G O R Y
          This NIC replaces the 10 Mbps version            Ethernet
          in the node.
                                                        SEE ALSO
      s   A 100BaseVG hub to replace the 10                HSLAN (High-Speed Local-Area
          Mbps hub. This type of hub is plug-              Network)
          compatible with a 10 Mbps hub, so
                                                        C O M P A RE
          that the upgrade requires simply
          unplugging a node from one hub and               100BaseT
          plugging it into the 100BaseVG hub.       w
          This can all take place in the wiring     100BaseX
                                                        100BaseX (sometimes written as 100
      If you are already using twisted-pair             Base-X) is a function that translates bet-
    Ethernet cabling, you may not need any              ween the FDDI (Fiber Distributed Data
    new wiring, provided that the cable has four        Interface)-based physical layer and the
    wire pairs.                                         CSMA/CD-based data-link layer in a 100
                                                                                 3174       9

megabit per second (Mbps) Ethernet pro-       w
posed by Grand Junction Networks. The         100 Mbps Ethernet
term was used more generally to refer to a        Any of several proposed 100 Mbps imple-
100 Mbps Ethernet developed by Grand              mentations of the Ethernet network archi-
Junction, among others. This proposed spec-       tecture. Three different approaches have
ification has since become known as Fast           been proposed: 100BaseVG, 100BaseX, and
Ethernet, and has been refined into three          fastEthernet. These implementations differ
variants:                                         most fundamentally in the media-access
   s   100BaseFX, which runs over fiber-           methods and types of cable they use.
       optic cable                            w
   s   100BaseT4, which runs over
                                              110-Type Punch-Down Block
       unshielded twisted pair (UTP) cable        A device for terminating wires, with the
       rated at Category 3 or higher—pro-         possibility of connecting input and output
       vided there are four available wire        wires. This type of punch-down block has
       pairs                                      generally replaced the older 66-type blocks
                                                  originally used by the telephone company.
   s   100BaseTX, which runs over
       Category 5 UTP cable                       SEE ALSO
                                                    Punch-Down Block
   These variants all use the standard
CSMA/CD (carrier sense multiple access/       w
collision detection) medium access scheme     193rd Bit
used by classic Ethernet. (In contrast, the       In a T1 communications channel, a framing
100BaseVG variant proposed by Hewlett-            bit that is attached to every group of 192
Packard and other companies uses a demand         bits. These 192 bits represent a single byte
priority access scheme.) Specifications and        from each of the 24 channels multiplexed in
standards for the Fast Ethernet versions          a T1 line.
have been debated by the IEEE 802.3u sub-
committee, and were just approved in June         SEE ALSO

1995.                                               T1

B RO A D E R C A T E G O R Y                  w
   Ethernet                                   3174
                                                  A cluster control unit for the IBM 3270
                                                  family of display terminals.
   Fast Ethernet
10        3270

w                                                   w
3270                                                3279
    The 3270 designation is used for a line of          The designation for a color version of the
    terminals, communications controllers, and          3278 terminal used to communicate with
    printers that are used with IBM mainframes.         IBM mainframes.
    The 3270 devices use synchronous commu-
    nications protocols, either SDLC (Synchro-
    nous Data Link Control) or BSC (Binary
    Synchronous Communication), to communi-             The designation for a computer that serves
    cate with the host.                                 as a data communications controller for
       In order for a stand-alone PC to commu-          IBM’s 370-series mainframes. The 3705 also
    nicate with an IBM mainframe, it must have          has ports for asynchronous access over
    an add-in board that enables the PC to emu-         dial-up lines.
    late a 3270 terminal.

3270 Data Stream
    In IBM’s SNA (Systems Network Architec-
    ture) environment, a stream in which char-
    acters are converted and/or formatted, as
    specified through control characters and
    attribute settings.

    The designation for a cluster controller that
    can serve as a front end for an IBM main-
    frame host. Devices, such as 3270 terminals
    or printers, communicate with the host
    through this controller. The 3274 cluster
    controllers have been replaced by 3174
    establishment controllers in newer

    The designation for a popular IBM terminal
    used to communicate with IBM mainframes.
    12      AA (Auto Answer)

                                                         communications protocols, and other items.

    AA (Auto Answer)                                     For example, Abstract Syntax Notation One
     A modem feature in which the modem can              (ASN.1) was developed as part of the OSI
     automatically respond to a call and establish       Reference Model; Extended Data Represen-
     a connection.                                       tation (XDR) was developed as part of Sun
                                                         Microsystems’ Network File System (NFS).
    AAL (ATM Adaptation Layer)                       w
                                                     AC (Access Control)
     The topmost of three layers defined for the
     ATM network architecture. The AAL medi-             A field in a token ring token or data frame.
     ates between the ATM layer and the various      w
     communication services involved in a trans-     AC (Alternating Current)
                                                         AC (alternating current) is a power supply
     SEE ALSO                                            whose polarity (direction of flow) switches
         ATM (Asynchronous Transfer Mode)                periodically. AC is the type of electrical
                                                         power supplied for homes and offices.
                                                            With AC, the actual amount of power
    AAR (Automatic Alternate Routing)
                                                         being supplied at any given moment depends
     In X.25 and other networks, the process             on where in the switching process you are.
     by which network traffic is automatically            When plotted over time, a “pure” AC power
     routed to maximize throughput, minimize             supply produces a sine wave.
     distance, or balance channel usage.                    Not all countries use the same switching
                                                         rate. For example, in North America, the
    ABM (Asynchronous Balanced Mode)                     current switches polarity 60 times per sec-
                                                         ond; in most European countries, the rate is
     In the ISO’s HDLC (High-Level Data-Link
                                                         50 times per second. These values are indi-
     Control) protocol, an operating mode that
                                                         cated as cycles per second, or hertz (Hz).
     gives each node in a point-to-point connec-
                                                         Thus, electrical power in the United States
     tion equal status as senders and receivers.
                                                         alternates at 60 Hz.
w                                                           Not all devices can use AC. In some cases
    ABP (Alternate Bipolar)                              the AC power must be converted to direct
     A signal-encoding method.                           current (DC), which provides a constant
                                                         voltage level and polarity. All digital systems
     SEE ALSO                                            (such as computers) must use DC.
         Encoding, Signal
                                                         C O M P A RE
w                                                          DC (Direct Current)
    Abstract Syntax
     A machine-independent set of language ele-
     ments and rules used to describe objects,
                                                                               Access Control      13

w                                                     w
AC (Application Context)                              Acceptance Cone                                      a
    In the OSI Reference Model, AC (applica-              In fiber optics, the three-dimensional analog     b
    tion context) is a term for all the application       of an acceptance angle. The cone generated
    service elements (ASEs) required to use an            by revolving the acceptance angle 360            c
    application in a particular context.
       More specifically, in network manage-
                                                          degrees with the center of the fiber’s core
                                                          as the cone’s point.
    ment, the AC provides the ground rules that
    serve to define the relationship between two
    applications during a temporary connection.
                                                      Access Control                                       f
    These ground rules will determine the types           An operating system uses access control to
                                                          determine the following:
    of services that can be invoked during the
    connection and also the manner in which                 s   How users or resources can interact
    information will be exchanged. Such a con-
    text is important for defining the systems
                                                                with the operating system                  i
    management services provided by a CMISE                 s   What a specific user or group of users      j
    (common management information service                      may do when interacting with the
    element).                                                   operating system                           k
                                                            s   Who can access a file or directory and      l
                                                                what that user can do after accessing it
      ASE (Application Service Element);                                                                   m
      CMISE (Common Management                              s   How system or network resources can
      Information Service Element);                             be used                                    n
      Acceptable Use Policy (AUP)
                                                             At the lowest levels, hardware elements       o
Acceptable Use Policy (AUP)
                                                          and software processes can obtain limited
                                                          access to the system through mechanisms
    SEE                                                   such as interrupts or polling. For example,      q
                                                          low-level access to DOS is through IRQs
      AUP (Acceptable Use Policy)
                                                          (interrupt request lines) and through soft-      r
Acceptance Angle
                                                          ware interrupts, such as INT 21H, which          s
                                                          provide programs with access to DOS capa-
    In fiber optics, a value that measures the             bilities and to certain hardware resources.      t
    range over which incoming light will be                  Access-control measures can be associ-
    reflected and propagated through the fiber.             ated with users, files and directories, or
    The size of this angle depends on the relative        resources. When assigned to users or groups      v
    refractive indexes of the fiber core, the clad-        of users, these control measures are known
    ding, and the surrounding medium (which is            as access rights, access privileges, trustee     w
    generally air).                                       rights, or permissions. When associated with
                                                          files and directories, the access-control
14        Access Control Decision Function (ACDF)

    elements are known as attributes or flags.        w
    Resources and other system objects gener-        Access Control Information (ACI)
    ally have an associated access control list          SEE
    (ACL), which contains all the users who                ACI (Access Control Information)
    may use the resource.
       Access control is generally specified by a     w
    system administrator or by the owner of a        Access Network
    particular file or resource. Some access privi-       A network attached to the trunk of a
    leges are determined for users during net-           backbone network. This type of connection
    work configuration; others may be assigned            usually requires a gateway or a router,
    when the user logs on to a network or begins         depending on the types of networks that
    a session with an operating system.                  comprise the backbone network.
       Access-control issues can be complex,
    particularly if multiple operating environ-
                                                     Access Rights
    ments are involved, as on an internetwork.
    One reason is that operating environments            Access rights are properties associated with
    differ in the access-control measures they           files or directories in a networking environ-
    support. Because there are overlaps, omis-           ment; also known as access privileges or
    sions, and definition differences, mapping            trustee rights. Access rights determine how
    access controls between environments may             users and network services can access and
    be complicated.                                      use files and directories. All networking
                                                         environments and operating systems use
                                                         some type of access rights settings to control
      Access Rights                                      access to the network and its resources.
w                                                           Access rights are similar to security
Access Control Decision Function                         attributes, which specify additional proper-
(ACDF)                                                   ties relating to a file or directory. Security
                                                         attributes can override access rights. In gen-
                                                         eral, rights are assigned to a user for a spe-
      ACDF (Access Control Decision                      cific file or directory. Attributes are assigned
      Function)                                          to a file or directory and control access by
w                                                        any user, regardless of that user’s rights. The
Access Control Enforcement Function                      set of rights a user has been assigned to a file
(ACEF)                                                   or directory is called his or her trustee
                                                            The number of access rights is relatively
      ACEF (Accces Control Enforcement
                                                         small. The terminology and particular com-
                                                         bination of rights vary from system to sys-
                                                         tem. For example, in Novell’s NetWare 3.x
                                                         and 4.x, access rights may be associated
                                                                                        Access Rights            15

with directories or files or both, and a right                        folder does not even appear on the
may apply to all the files in a directory or                          user’s screen.                                   a
only to individual ones. In NetWare 2.x,
rights apply only to directories. See the table
                                                               s     Make Changes, which allows a user to             b
                                                                     change the contents of a file or folder.
“Novell NetWare Access Rights” for
                                                                     Even drastic changes such as deletions
descriptions of the access rights associated
with NetWare.
                                                                     are allowed.                                     d
   The meaning or effect of a specific privi-                  These AppleShare environment privileges                 e
lege may also be system-dependent. For                      may be granted to any of the following:
example, in an AppleShare environment, the
                                                               s     Owner: The user who created (and,
following access privileges are defined:
                                                                     hence, owns) the file or folder.                  g
   s   See Files, which allows a user to see,
       open, and copy files.
                                                               s     Group: The collection of users to                h
                                                                     whom the privilege is granted. This
   s   See Folder, which allows a user to see a                      may be a single user.                            i
       folder (but not necessarily the folder’s
       contents). If this privilege is not set, the
                                                               s     Everyone: All users with access to the           j
                                                                     file server.
AC C E S S R I G H T     U S AG E A L L OW E D
Access Control (A)       Allows you to modify the trustee assignments and inherited rights mask (IRM) for a           o
                         file. With Access Control rights, you can grant other users any rights except Supervi-
                         sory rights.
Create (C)               Allows you to create subdirectories or files within a directory. Also allows you to           q
                         salvage a file if it is deleted.
Erase (E)                Allows you to delete a file or directory.
File Scan (F)            Allows you to see a file or directory name when listing the parent directory.

Modify (M)               Allows you to change the name and attributes of a file or directory.
Read (R)                 Allows you to open and read a file.
Supervisory (S)          Allows you to exercise all rights to a file or directory, including the right to grant        v
                         Supervisory privileges to the file or directory to other users. (This right does not
                         exist in NetWare 2.x.)
Write (W)                Allows you to open, edit, and save a file.
16       Access Time

   In UNIX, owners, groups, and others                w
 may be granted read, write, or execute per-          Access Time
 missions for a file or a directory, as follows:           In hard-disk performance, the average
     s   Read access for a file allows a user              amount of time it takes to move the read/
         to read or display the contents of a             write heads to a specified location and
         file. Read permission for a directory             retrieve data at that location. The lower the
         means the user can generate a direc-             value, the better the performance. Currently,
         tory listing.                                    hard disks with access times of less than 15
                                                          milliseconds are common.
     s   Write access for a file means the user
         can edit the file or redirect output to it.   w

         Write access for a directory allows the
         user to create a file or a subdirectory.          A process by which network usage can be
                                                          determined and charges assessed for use of
     s   Execute access for a file allows the user
                                                          network resources, such as storage, access,
         to use the file name as a command.
                                                          and services. Accounting measures include
         Execute permission for a directory
                                                          blocks read, blocks written, connect time,
         means the user can pass through the
                                                          disk storage, and service requests.
         directory to subdirectories.
                                                             Most network operating systems include
    When a single machine or network                      an accounting utility or support an add-on
 includes more than one environment, there                accounting package. For example, NetWare
 must be a well-defined rule for assigning                 3.11 has an accounting option in its
 and determining access rights. For example,              SYSCON utility.
 in NetWare for Macintosh, the NetWare
 access rights supersede the AppleShare               Accounting Management
 access privileges.
    Similarly, there are mechanisms for ensur-            One of five OSI network management
 ing that access rights are applied only as               domains defined by the ISO and CCITT.
 broadly as intended. For example, NetWare                This domain is concerned with the adminis-
 uses an Inherited Rights Mask (version 3.x)              tration of network usage, costs, charges, and
 or Inherited Rights Filter (version 4.x) to              access to various resources.
 specify which access rights for a directory              SEE ALSO
 are also applicable in a subdirectory.                     Network Management
 B RO A D E R C A T E G O R Y                         w
     Access Control                                   Account Policy
 SEE ALSO                                                 In networking and other multiuser environ-
     Attribute; IRM/IRF (Inherited Rights                 ments, a set of rules that determines whether
     Mask/Inherited Rights Filter)                        a particular user is allowed to access the sys-
                                                          tem and what resources the user may use. In
                                                    ACF (Advanced Communications Function)         17

    Windows NT Advanced Server, the account           w
    policy determines the way in which pass-           ACF (Advanced Communications                      a
    words may be used in a domain (a group of
    servers with a common security policy and
                                                          ACF (Advanced Communications Function)
    database).                                            is the base name for several IBM software      c
                                                          packages that operate under IBM’s SNA
                                                          (Systems Network Architecture). In some
Accumaster Integrator
    A network management program from
                                                          cases, the programs are revisions or exten-    e
                                                          sions of older programs.
                                                              The following programs are included:       f
w                                                            s   ACF/NCP (Advanced Communica-
ACD (Automatic Call Distributor)
    A device that automatically switches an
                                                                 tions Function/Network Control Pro-
                                                                 gram): Resides in a communications
    incoming call to the next available line.                    controller. It provides and controls    i
                                                                 communications between the host
ACDF (Access Control Decision                                    machine and the network devices.
Function)                                                    s   ACF/TCAM (Advanced Communica-           k
    In open systems, a function that uses various                tions Function/Telecommunications       l
    types of information, such as ACI (access                    Access Method): Serves as an ACF/
    control information), and guidelines to                      VTAM application and provides mes-      m
                                                                 sage handling and other capabilities.
    decide whether to grant access to resources
    in a particular situation.
                                                             s   ACF/VTAM (Advanced Communica-
w                                                                tions Function/Virtual Telecommuni-
ACE (Adverse Channel Enhancement)                                cations Access Method): Provides and    p
    A modem-adjustment method that allows                        controls communications between a
    the modem to compensate for noisy lines.                     terminal and host programs. ACF/        q
    For example, the modem might lower the                       VTAM supersedes and adds capabili-
                                                                 ties to the older VTAM software.
    operating speed.
                                                             s   ACF/VTAME (Advanced Communi-
ACEF (Access Control Enforcement                                 cations Function/Virtual Telecommu-     t
Function)                                                        nications Access Method-Entry): An
                                                                 obsolete program that has been super-   u
    In open systems, a function that enforces
    the decision made by the ACDF (access
                                                                 seded by ACF/VTAM.
    control decision function).                           B RO A D E R C A T E G O R Y
                                                             SNA (Systems Network Architecture)
18        ACI (Access Control Information)

w                                                        workstation, and to manage the server from
ACI (Access Control Information)                         this workstation. In NetWare 3.x, the
    In the CCITT’s X.500 directory services              RCONSOLE utility provides the same func-
    model, any information used in controlling           tion across a direct connection.
    access to a file or directory.                           In NetWare 4.x, RCONSOLE was
                                                         updated to add ACONSOLE’s asynchro-
w                                                        nous capability, and ACONSOLE was
ACID (Atomicity, Consistency,                            removed.
Isolation, and Durability)
                                                         B RO A D E R C A T E G O R Y
    In transaction processing (TP), the attributes
    that are desirable for a transaction.                   NetWare

ACK                                                  ACS (Asynchronous Communications
    In telecommunications, a control character
    that indicates that a packet has been                An ACS is usually a dedicated PC or expan-
    received without an error. In certain net-           sion board that provides other network
    work architectures, ACK is the name for a            nodes with access to any of several serial
    frame that sends such an acknowledgment.             ports or modems. The ports may be con-
    The ASCII ACK character has value 6.                 nected to mainframes or minicomputers.
                                                            To access a modem or a port, the work-
w                                                        station user can run an ordinary communi-
ACL (Access Control List)                                cations program in a transparent manner.
    In some networking environments, the ACL             However, in order for this to work, one of
    is a list of services available on a network,        the following must be the case:
    along with the users and devices that are               s   The communications program
    allowed to use each service. This list pro-
                                                                must include a redirector to route
    vides one way to control access to network
                                                                the communication process to the
                                                                appropriate ACS.
        In NetWare Directory Services (NDS),
    each object in the directory has a property             s   The workstation must have a special
    called the ACL, which lists all the other                   hardware port emulation board
    objects that have trustee assignments (rights)              installed, which takes up one of the
    to that object.                                             workstation’s expansion slots. In this
                                                                case, the communications package
                                                                does not need any special rerouting
    In Novell’s NetWare 3.x, ACONSOLE is a
    utility that allows a network supervisor to
                                                            s   The user must run a redirection
    access a server through a modem from a                      program before starting the commu-
                                                                nications package. To work with a
                                                                                      Adapter      19

          software-based redirector, the commu-       w
          nications package must be able to use       Active Link                                        a
          DOS interrupt INT 14H. Unfortu-
          nately, many communications pro-
                                                          In an ARCnet network, a box used to con-       b
                                                          nect two cable segments when both cable
          grams bypass this interrupt to access           segments have high-impedance network           c
          the UART (universal asynchronous
          receiver/transmitter) directly for faster
                                                          interface cards (NICs) connected.              d
          operation.                                  w
                                                      Active Star
    B RO A D E R C A T E G O R Y
                                                          A network configuration in which the            f
       Server                                             central node of a star topology cleans
                                                          and boosts a signal.
ACSE (Association Control Service
                                                          SEE ALSO
                                                            Topology, Star                               i
    In the OSI Reference Model, an application-
    level service that establishes the appropriate    w
                                                      ACU (Autocall Unit)
    relationship between two applications, so
    that they can cooperate and communicate               A device that can dial telephone numbers
    on a task, such as exchanging information.            automatically.                                 l
w                                                     w
                                                      AD (Administrative Domain)
    When used to describe hardware or a con-              In the Internet community, a collection of     n
    figuration, active generally means that the            nodes, routers, and connectors that is man-    o
    hardware does some signal processing—                 aged by a common administrator, such as an
    cleaning, boosting, or both. For example,             organization or a company.                     p
    an active hub boosts and cleans a signal
    before passing it on.
                                                      w                                                  q
w                                                         A board that plugs into an expansion bus,
Active Hub                                                and that provides special capabilities, such   s
    In an ARCnet network, a component that                as video, fax, modem, network access, and
    makes it possible to connect additional               so on. Besides functionality, adapters are     t
    nodes to the network and also to boost                distinguished by the width of the data bus     u
    signals that go through the hub.                      between the adapter and the PC. Adapters
                                                          may have 8-, 16-, or 32-bit connections.       v
       Hub                                                                                               w
20        ADC (Analog-to-Digital Converter)

w                                                   Hardware Address
ADC (Analog-to-Digital Converter)
                                                      A hardware address, also known as a physi-
    A device that converts an analog signal to
                                                      cal address or a MAC address, is a unique
    digital form.
                                                      numerical value assigned to a network inter-
w                                                     face card (NIC) during the manufacturing
ADDMD (Administrative Directory                       process or by setting jumpers or switches
Management Domain)                                    during network installation. One part of this
    In the CCITT’s X.500 directory services           address is assigned to the manufacturer by
    model, a collection of directory system           the IEEE (Institute of Electronics Engineers)
    agents (DSAs) under the control of a single       and is common to all components from that
    authority.                                        manufacturer; the second part of the hard-
                                                      ware address is a unique value assigned by
                                                      the hardware manufacturer.
      DSA (Directory System Agent)
                                                    Network Address
Address                                               A network address is an arbitrary value that
    An address is a value used to specify a loca-     is assigned identically to each station in a
    tion. The location may be an area of local or     particular network. As long as there is only
    shared memory, or it may be a node or other       a single network, this value is automatically
    device on a network.                              unique. If two or more networks are con-
                                                      nected, each must have a different network
Network-Related Addresses
                                                      address. If a station (for example, a server)
    Several types of addresses are distinguished      connects to two networks, that station will
    for network locations. The type of address        have two different network addresses.
    used in a particular context depends partly           A network address is also known as a net-
    on which protocol or device is creating the       work number or an IPX external network
    address. Address information may be main-         number.
    tained in any of several ways, such as in
    look-up tables or directories.                  Node Address
       Some common types of network-related           In addition to a common network address,
    addresses are hardware, network, node,            each station in a network has a unique node
    Internet, and e-mail (electronic mail). There     address. This value identifies a particular
    are other types of addresses, and not all         node, or more specifically, the NIC assigned
    types of addresses are used in the same           to each node, in a particular network. This
    conceptual model. Devices that connect            address is also known as a node number or
    networks or network segments generally            station address.
    get network and/or node addresses on                 When specified as a source or destination,
    each network they connect.                        a network server or workstation may be
                                                                                     Address      21

  identified by a network and a node address             logical address. Only certain network oper-
  or by a hardware address.                             ating systems, such as NetWare, support
     The node addresses for Ethernet cards are
  factory-set, and no two cards have the same
                                                        internal addresses.
                                                           See the figure “Examples of network
  number. The node addresses for ARCnet                 addresses” for an illustration of the kinds of   c
  and Token Ring cards are set by changing              addresses discussed so far.
  jumpers or switches on the cards. If a node
  contains two NICs, the node will have two           Internet Address                                   e
  different network addresses.
                                                        An Internet address is a network-layer           f
                                                        address that uniquely identifies a node on
Internal Address
                                                        an internetwork or on the Internet. This type    g
  An internal address is a unique value that
  specifies a node with respect to a particular
                                                        of address uses four bytes of storage, and
                                                        it is generally represented as four decimal
  server in a network, which is useful in net-          values separated by decimal points, as in        i
  works that have multiple servers. This is a  Certain bits from an Internet
  E X A M P L E S O F N E T WO R K A D D RE S S E S                                                      k
 22      Address Bus

  address can be masked to identify a subnet-         The segment value represents a (usually 16-
  work that contains some of the nodes in the         byte) location that is aligned on a paragraph
  internetwork.                                       boundary. The offset value represents the
     Special protocols, such as the Address           number of bytes to shift from this segment
  Resolution Protocol (ARP), are used to con-         address. DOS uses segmented addresses.
  vert from an Internet to a hardware address;
  other programs, such as the Reverse ARP         Paged Address
  (RARP), convert from a hardware to an               Certain types of address space actually con-
  Internet address.                                   sist of two types of values. For example, in
                                                      expanded memory, locations in a special
E-Mail Address
                                                      set of chips, and hence, in a special set of
  An e-mail (electronic mail) address is              addresses, are mapped into special memory
  an application-layer address that identifies         buffers. These buffers are broken into pages
  a user’s mailbox location in a message-             of a specific size.
  handling system. These addresses have                  Virtual memory also uses paged
  little in common with the other types of            addresses.
  addresses mentioned; however, the e-mail
  address must be associated with the station’s   Address Bus
  network and node address or with its hard-
  ware address in order for messages to be            An address bus is the electrical signal lines
  transferred from a sender to a receiver.            over which memory locations are specified.
                                                      Each line carries a single bit, so the number
Memory-Related Addresses                              of lines on the bus determines the number of
                                                      possible addresses:
  Several different formats are used for mem-
  ory addresses in personal computers: flat              s   20 lines allow access to 1 megabyte
  address space, segmented address, and                     (MB) of memory. Examples include
  paged address.                                            Intel’s 8086 and 8088 processors.
                                                        s   24 lines provide access to 16 MB.
Flat Address Space
                                                            Examples include Intel’s 80286 and
  An address in a flat address space is a simple             Motorola’s 68000 processors.
  numerical value in the range between 0 and            s   32 lines provide access to 4 gigabytes
  the highest address value. For example, in a              (AB). Examples include Intel’s 80386,
  machine with 1 megabyte of memory, the                    80486, and Pentium; and Motorola’s
  addresses range from 0x00000 to 0xfffff.                  68020 and later processors.

Segmented Address                                       s   64 lines provide access to 16 exabytes
                                                            (EB). (An exabyte is a billion billion,
  An address in a segmented address space
  consists of a segment and an offset value.
                                                                                 Administration   23

         or a quintillion, bytes.) Digital Equip-       (Consultative Committee for International
         ment Corporation’s Alpha APX chip is           Telegraphy and Telephony) or a national
         an example of a 64-bit address bus.            PTT (Post, Telegraph, and Telephone). Spe-
                                                        cific examples of ADMDs include MCImail
Address Mask                                            and AT&Tmail in the United States; British     c
                                                        Telecom Gold400mail in Britain.
    In the IP (Internet Protocol) addressing               ADMDs are public carriers, unlike
    scheme, a group of selected bits whose val-
    ues identify a subnetwork; also known as a
                                                        PRMDs (private management domains),            e
                                                        which are run by private organizations or
    subnet mask. All the members of that sub-           companies. In accordance with CCITT            f
    network share the same mask value. Using
    an address mask makes it easier for the sys-
                                                        guidelines, ADMDs handle any interna-          g
                                                        tional connections; PRMDs communicate
    tem to reference a member of a particular           through a local ADMD. ADMDs can con-           h
    subnet.                                             nect PRMDs, but a PRMD cannot connect
                                                        ADMDs. Because all ADMDs run under the
Address Resolution                                      auspices of CCITT, the conglomeration of       j
                                                        ADMDs in the world forms the backbone
    The process of mapping one type of address
    to another; specifically, mapping a network
                                                        for a global X.400 network.                    k
    (local) address to a hardware-dependent             B RO A D E R C A T E G O R I E S               l
    address. The most widely used method                   MD (Management Domain); X.400
    of address resolution is the Address Resolu-
    tion Protocol (ARP) or a variation of that
                                                        C O M P A RE
                                                           PRMD (Private Management Domain)
Adjacent Channel                                    Administration                                     p
                                                        Administration involves the management
    A frequency band immediately before or
                                                        and maintenance of a computer system,
    after the current channel. For example, a
    channel between 100 MHz and 500 MHz                 network, or environment.                       r
    and a channel between 700 MHz and 900
    MHz are both adjacent to the channel
                                                    Administrative Tasks                               s
    between 500 MHz and 700 MHz.                        An administrator’s responsibilities may        t
                                                        be grouped into the following general
                                                        categories:                                    u
ADMD (Administration Management
Domain)                                                    Configuration management: Handling           v
                                                             tasks such as user accounts, hardware
    In the CCITT’s X.400 Message Handling
                                                             settings, access rights, and security.
    System (MHS) model, an ADMD (Adminis-
    tration Management Domain) is a network                Data-flow management: Monitoring
    or network section operated by the CCITT                 performance, managing memory              y
 24      Administration Management Domain (ADMD)

        and resources, making sure applica-               Internetwork: Refers to multiple net-
        tions and data files are accessible, and              works. Some or all of these networks
        generally ensuring that data is flowing               may use different architectures. An
        properly.                                            internetwork administrator should be
                                                             able to assume that any subnetworks
      Hardware maintenance: Installing, main-
                                                             are under the control of network
        taining, and diagnosing hardware
                                                             administrators, so that the internet-
                                                             work administrator can concentrate
      Software maintenance: Installing applica-              on the connections between networks
        tions and other software, software                   rather than those between machines.
        version control, bug reporting and
        resolution, and so on.
                                                    Administration Management Domain
      Help: Training users, providing documen-      (ADMD)
        tation for using the system resources           SEE
        and applications, and offering other
                                                          ADMD (Administration Management
Levels of Administration                            w
                                                    Administrative Domain (AD)
  Various levels of administration are distin-
  guished, including the following:                     SEE
                                                          AD (Administrative Domain)
      System: Refers to a particular division
         in a company or a particular type of       w
         hardware, such as mainframes or data-      Advanced Function Printing (AFP)
         base servers. System administration
         responsibilities do not necessarily
                                                          AFP (Advanced Function Printing)
         involve networking issues; that is, a
         system administrator may or may not        w
         need to attend to issues relating to the   Advanced Intelligent Network (AIN)
         connections between machines, as well
         as to the machines themselves.
                                                          AIN (Advanced Intelligent Network)
      Network: Usually refers to a LAN (local-
        area network), but may encompass            Advanced Mobile Phone Service
        machines in a larger range, provided        (AMPS)
        these machines are all connected by a
        common architecture. In addition to             SEE
        the individual machines, a network                AMPS (Advanced Mobile Phone Service)
        administrator must keep track of the
        connections between the machines.
                                                                                        Agent      25

w                                                       authority, or administrator, that is allocating
Advanced Research Projects Agency                       the IDI (initial domain identifier) values. The
(ARPA)                                                  AFI also specifies the format of the IDI and
                                                        the DSP (domain specific part), which are
      ARPA (Advanced Research Projects                  other parts of the NSAP address.                  c
      Agency)                                       w                                                     d
                                                    AFP (Advanced Function Printing)
                                                        In IBM’s SAA (Systems Applications Archi-
Advantage Networks
    Advantage networks represent a networking           tecture) environments, the ability to print       f
                                                        text and images; that is, to use all points
    strategy from Digital Equipment Corpora-
                                                        addressable (APA) printers.
    tion (DEC), designed to add support for
    protocols such as the TCP/IP suite to           w
    DEC’s OSI-compliant DECnet Phase V
                                                    AFT (Application File Transfer)                       i
                                                        In the International Standardized Profile
w                                                       (ISP) grouping, a prefix that identifies FTAM       j
Adverse Channel Enhancement (ACE)                       (file transfer, access, and management) pro-
                                                        files. For example, AFT11 represents basic
      ACE (Adverse Channel Enhancement)
                                                        file transfer.                                     l
                                                    w                                                     m
                                                        In general, an agent is a program that can
    The process by which a network service
    makes its presence and availability known
                                                        perform a particular task automatically,          o
                                                        when appropriate or upon request by
    on the network. For example, Novell Net-
                                                        another program. An agent is commonly
    Ware services use the SAP (Service Advertis-
    ing Protocol).
                                                        used to provide information to an applica-        q
                                                        tion, such as a network management pro-
w                                                       gram. An agent may be machine- or                 r
AE (Application Entity)                                 function-specific.
                                                           The following are some of the agents that
    In the OSI Reference Model, an entity (pro-
    cess or function) that runs all or part of an       are found in networking-related contexts:         t
    application. An AE may consist of one or
    more application service elements (ASEs).
                                                          s   In a client-server networking model,        u
                                                              an element that does work on behalf of
                                                              a client or a server application. For       v
AFI (Authority and Format Identifier)                          example, in Novell’s SMS (storage
                                                              management system) backup architec-
    In the OSI Reference Model, part of the
    address for the network-layer service access
                                                              ture, a special backup agent, called a      x
                                                              TSA (target service agent), is loaded on
    point (NSAP). The AFI portion specifies the                                                            y
26        Aging

          every node that you want to back up        w
          from a centralized location. The agent     AIM (Analog Intensity Modulation)
          allows the central backup program to           In communications using light (rather than
          access and back up the data on that            electrical) signals, a modulation method in
          node.                                          which the intensity of the light source varies
      s   In an IBM Token Ring architecture, an          as a function of the signal being transmitted.
          element on the network interface card      w
          that monitors certain aspects of the       AIN (Advanced Intelligent Network)
          node and ring performance, and that
                                                         In telecommunications, the name for a
          reports this information to a network
                                                         sophisticated digital network of the future.
          management program or to a Ring
          Error Monitor (REM).                       w
                                                     AIS (Alarm Indication Signal)
      s   In network management and mon-
          itoring, a terminate-and-stay-resident         A signal used in the OSI network manage-
          (TSR) program that runs on a work-             ment model and also in broadband ISDN
          station to monitor activity and report         networks to indicate the presence of an
          this to a network management                   alarm or error somewhere on the network.
          program.                                   w

       The data collected by an agent is orga-
                                                     AL (Application Layer)
    nized and processed by an agent handler. In          The topmost of the seven layers in the OSI
    network management, an agent handler may             Reference Model.
    organize and analyze data concerning some
                                                         SEE ALSO
    network function or component.
                                                           OSI Layer
    A process by which old items or table entries
                                                         In various network environments, particu-
    are removed in a systematic manner, such as
                                                         larly network management, an alarm is a
    first in, first out. This process serves both to
                                                         signal used to indicate that an abnormality,
    update such tables and to speed up access.
                                                         a fault, or a security violation has been
w                                                        detected. Alarms may be distinguished by
AI (Authentication Information)                          type, such as performance, fault, or security,
    In network security, information used to             and also by the severity of the event that
    determine whether a user is legitimate and           caused the alarm.
    authorized to access the system.                        At one extreme are critical events that
                                                         represent immediate threats to continued
                                                         network operation; for example, when a
                                                         crucial LAN (local-area network) node or a
                                                                                Algorithm      27

    server goes down. In some network manage-           In most cases, the algorithms are of little
    ment environments, such critical alarms may      interest to either the casual or intense net-
    trigger automatic response by the network
    management package.
                                                     work user. However, several algorithms have
                                                     escaped from behind the scenes and have
       At the other extreme are events that are      actually become items in marketing litera-       c
    not currently serious, but that may eventu-      ture and other product discussions. The
    ally become serious enough to threaten net-      following are a few of the better-known
    work operation; for example, when network        algorithms:                                      e
    traffic is getting close to the network’s band-
    width limit. Such events generally do not
                                                       Auto-partition: An algorithm by which a        f
                                                         repeater can automatically disconnect
    require immediate correction but should be
                                                         a segment from a network if that seg-        g
                                                         ment is not functioning properly. This
w                                                        can happen, for example, when a bro-
Alarm Indication Signal (AIS)                            ken or unterminated cable causes too         i
                                                         many collisions. When the collisions
                                                         have subsided, the network segment           j
      AIS (Alarm Indication Signal)

                                                         can be reconnected.                          k
                                                       Bellman-Ford: An algorithm for finding
Alarm Reporting Function (ARF)
                                                         routes through an internetwork. The
    SEE                                                  algorithm uses distance vectors, as          m
      ARF (Alarm Reporting Function)                     opposed to link states. The Bellman-
                                                         Ford algorithm is also known as the          n
Alert                                                    old ARPAnet algorithm.                       o
    In network management, an alarm sent by            Distance-vector: A class of computation-
                                                         intensive routing algorithms in which
    an agent to the administrator. An alert
    reports that a problem has arisen or that a          each router computes the distance            q
    threshold has been reached.                          between itself and each possible desti-
                                                         nation. This is accomplished by com-
                                                         puting the distance between a router         s
                                                         and all of its immediate router neigh-
    An algorithm is a predefined set of instruc-          bors, and adding each neighboring            t
    tions for accomplishing a task. An algorithm         router’s computations for the distances
    is guaranteed to produce a result in a finite         between that neighbor and all of its
    amount of time. Algorithms are used in               immediate neighbors. Several com-            v
    many ways in networking. For example,                monly used implementations are
    there are hashing algorithms for finding file          available, such as the Bellman-Ford          w
    names in a directory and timing algorithms
    for deciding how long to wait before trying
                                                         algorithm and the ISO’s Interdomain          x
                                                         Routing Protocol (IDRP).
    to access a network.                                                                              y
28      Alias

     Hot potato: In networks, a routing algo-      w
       rithm in which a node routes a packet       Alias
       or message to the output line with the          In a computer environment, a name that
       shortest queue.                                 represents another, usually longer, name. In
     Link-states: A class of routing algorithms        NetWare Directory Services (NDS), an alias
        in which each router knows the loca-           is an object in one part of the Directory
        tion of and distance to each of its            tree that points to the real object, which is
        immediately neighboring routers, and           located in a different part of the tree. Users
        can broadcast this information to all          can access the real object through the alias.
        other routers in a link state packet       w
        (LSP). If a router updates its LSP, the    Alignment Error
        new version is broadcast and replaces
                                                       In an Ethernet or other network, an error
        the older versions at each other router.
                                                       in which a packet has extra bits; that is, the
        The scheme used to distribute the LSP
                                                       packet does not end on byte-boundaries and
        greatly influences the performance of
                                                       will have invalid CRC (cyclic redundancy
        the routers. These types of algorithm
                                                       check) values. An alignment error may be
        are an alternative to distance-vector
                                                       caused by a faulty component, such as a
        algorithms; rather than storing actual
                                                       damaged network interface card (NIC),
        paths, link-state algorithms store the
                                                       transceiver, or cable.
        information needed to generate such
        paths. The ISO’s open shortest path        w
        first (OSPF) algorithm is an example        Allocation Unit
        of a link-state algorithm.                     In Novell’s NetWare, areas that are used
     Spanning-tree: An algorithm that is used          to store information from files and tables.
       to compute open paths (paths without            Two types of storage are distinguished:
       loops) among networks. The algorithm            blocks, which are used to store data on disk,
       can generate all such paths and select          and buffers, which hold data in RAM
       one. If that path becomes inoperative           temporarily.
       because a node has gone down, the               SEE ALSO
       algorithm can find an alternate path.              Block; Buffer, Fiber-Optic Cable; Buffer,
       This type of algorithm is used by                 Memory.
       bridges to find the best path between
       two nodes in different networks, and        w
       to ensure that no path loops occur in       Alternate Mark Inversion (AMI)
       the internetwork. This algorithm is             SEE
       defined in the IEEE 802.1 standard.                AMI (Alternate Mark Inversion)
                                                               AMP (Active Monitor Present)      29

w                                                   w
Alternate Route Selection (ARS)                     American National Standards Institute               a
    SEE                                                                                                 b
      ARS (Alternate Route Selection)                   SEE
                                                          ANSI (American National Standards
Alternate Routing                                         Institute)                                    d
    This term describes the use of an alternative   w                                                   e
    communications path, such as a telephone        America Online (AOL)
    connection, when the primary one is not             SEE
    available.                                            AOL (America Online)                          g
AM (Accounting Management)
                                                    w                                                   h
                                                    AMF (Account Metering Function)
    In network management, a function for               In the OSI network management model, the
    gathering performance and usage informa-
    tion from a network.
                                                        function that keeps track of every user’s       j
                                                        resource usage.
AM (Active Monitor)                                 AMH (Application Message Handling)                  l
    In a token ring network, the node that is
    responsible for creating, passing, and main-
                                                        In the International Standardized Profile        m
                                                        (ISP) model, the prefix used to identify MHS
    taining the token. The performance of the           (Message Handling System) actions.              n
    AM is monitored constantly by standby
    monitors (SMs) to ensure that the token-        w                                                   o
                                                    AMI (Alternate Mark Inversion)
    passing process is not interrupted.                                                                 p
                                                        A signal-encoding scheme in which a 1 is
AME (Asynchronous Modem
                                                        represented alternately as positive and nega-   q
                                                        tive voltage, and 0 is represented as zero
                                                        voltage. It does not use transition coding,
    An AME, also known as a null modem, is a
    serial cable and connector with a modified
                                                        but can detect noise-induced errors at the      s
                                                        hardware level.
    pin configuration (compared to an ordinary
                                                        SEE ALSO
    RS-232 cable). This cable enables two com-
    puters to communicate directly; that is,              Encoding, Signal                              u
    without modems as intermediaries.               w                                                   v
                                                    AMP (Active Monitor Present)
                                                        In token ring networks, a packet issued
                                                        every 3 seconds by the active monitor (AM)      x
30        Amplifier

    on the ring to indicate that the AM is work-     w
    ing and is still in charge.                      Analog-to-Digital Conversion
                                                         The process of converting an analog signal
Amplifier                                                 (one that can take on any value within a
                                                         specified range) to digital form. An analog-
    A device for boosting an analog signal. The
                                                         to-digital converter (ADC) is a device that
    same service is provided by a repeater for
                                                         converts an analog signal to digital form.
    digital signals.
w                                                    ANF (AppleTalk Networking Forum)
                                                         A consortium of developers and vendors
    The magnitude, or level, of a signal. For
                                                         working to encapsulate AppleTalk in other
    an electrical signal, it is expressed in volts
                                                         protocols; for example, within the TCP/IP
    (voltage) or amperes (current). In computer
    contexts, current is more likely to be
    expressed in milliamperes.                       w
                                                     ANI (Automatic Number
w                                                    Identification)
AMPS (Advanced Mobile Phone
Service)                                                 In ISDN and some other telecommunica-
                                                         tions environments, a feature that includes
    A cellular telephone service. AMPS is a wire-
                                                         the sender’s identification number, such as
    less analog communications service that
                                                         telephone number, in the transmission, so
    operates in the 825 to 890 megahertz range.
                                                         that the recipient knows who is calling; also
w                                                        known as caller ID.
Analog Communication
    A telecommunications system that uses            Annex D
    analog (that is, continuous, sinusoidal)
                                                         In frame-relay technology, a document that
    signals to represent information. An exam-
                                                         specifies a method for indicating permanent
    ple of an analog communication system is
                                                         virtual circuit (PVC) status. The document is
    the classic voice-based telephone system
                                                         part of the ANSI T1.617 standard.
    (which is being replaced by the newer, digital
    systems).                                        w
                                                     Anonymous FTP
Analog Intensity Modulation (AIM)                        On the Internet, a protocol that allows a
                                                         user to retrieve publicly available files from
    SEE                                                  other networks. By using the special user ID,
      AIM (Analog Intensity Modulation)                  “anonymous” users can transfer files with-
                                                         out a password or other login credentials.
                                                         (FTP is an application-layer protocol in the
                                                         Internet’s TCP/IP protocol suite.)
                                                                       Anti-Virus Program      31

w                                                        Other anti-virus programs are intended to
Anonymous Remailer                                   be run periodically. When they are run, the
    An Internet service that can be used to hide     programs look for the tell-tale signs (known
                                                     as signatures) of particular viruses. These
    the origins of an e-mail message being sent
    to someone. The anonymous remailer               programs are minimally disruptive; on the         c
                                                     other hand, their effectiveness is directly
    removes any source address information
    from a message, substitutes any specified         proportional to the frequency with which
    pen name, and then sends the message on          they are used.                                    e
    to the specified destination.                         Because the coding for computer viruses
                                                     is constantly changing, anti-virus programs       f
ANSI (American National Standards
                                                     must also be updated regularly. It is impor-      g
                                                     tant to test anti-virus programs thoroughly,
Institute)                                           which means that every new release must be        h
    The United States representative in the ISO      tested. Make sure an anti-virus program per-
    (International Standardization Organiza-         forms to your expectations before installing
    tion). ANSI creates and publishes standards      it on a network. Some programs can eat up         j
    for programming languages, communica-            a significant amount of working memory.
    tions, and networking. For example, the              Recently, a very different (and, conse-       k
    standard for the FDDI network architecture       quently, very controversial) type of anti-
    is ANSI X3T9.5.                                  virus program has become available. InVirc-
                                                     ible, created by Zvi Netiv, is designed to        m
                                                     detect viruses that have already infected a
Anti-Virus Program
                                                     system, and to clean these up. Rather than        n
    An anti-virus program is used for detecting
    or removing a computer virus. An anti-virus
                                                     looking for virus signatures, InVircible uses     o
                                                     expert system rules to look for behavior
    program looks for suspicious activity, such      characteristic of viruses: replication, use of    p
    as unnecessary disk access, attempts to inter-
    cept a BIOS or other low-level call, and
                                                     memory, attempts to attach to the anti-virus
                                                     program, etc. InVircible will even put out
    attempts to format or delete files. In some
    cases, the anti-virus program detects a pat-
                                                     “virus bait” to get an existing virus to try to   r
                                                     infect the bait.
    tern characteristic of a particular virus.                                                         s
       Some anti-virus programs are TSR              B RO A D E R C A T E G O R Y
    (terminate-and-stay-resident) programs,             Data Protection                                t
    which monitor computer activity constantly,
                                                     RELATED AR TICLE                                  u
    looking for indications of a virus. In some
    cases, these types of programs can be
    extremely annoying and very processor
    intensive. Users have been known to remove
    an anti-virus TSR program from memory                                                              x
    out of frustration.
32        AOL (America Online)

w                                                    w
AOL (America Online)                                 APD (Avalanche Photodiode)
    America Online is a commercial online ser-           A detector component in some fiber-optic
    vice like CompuServe and Prodigy. AOL                receivers. The APD converts light into elec-
    supports both DOS and Windows users, and             trical energy. The “avalanche” refers to the
    provides a range of services (mail, news, ref-       fact that the detector emits multiple elec-
    erence, financial, entertainment, Internet            trons for each incoming photon (light
    access, etc.). Users pay a flat monthly fee,          particle).
    which allows a limited number of free hours.
    Additional hours are billed at a predeter-
                                                     APDU (Application Protocol Data
    mined rate. AOL’s graphical interface is
    highly regarded—in fact, Apple has licensed
    the interface technology for use in Apple’s          A data packet at the application layer; also
    eWorld interface. AOL provides a very com-           called application-layer PDU.
    prehensive set of access opportunities to the        SEE ALSO
    Internet.                                              OSI Reference Model
      Call AOL at 800-827-6364                       API (Application Program Interface)
w                                                        An abstract interface to the services and pro-
AOM (Application OSI Management)                         tocols offered by an operating system, usu-
    In the International Standardized Profile             ally involving a published set of function
    (ISP) model, the prefix for functions and             calls. Programmers and applications can
    services related to network management.              use the functions available in this interface
                                                         to gain access to the operating system’s
w                                                        services.
AOW (Asia and Oceania Workshop)
    One of three regional workshops for imple-
                                                     APIA (Application Program Interface
    menters of the OSI Reference Model. The
    other two are EWOC (European Workshop
    for Open Systems) and OIW (OSI Imple-                A group that writes APIs for the CCITT’s
    menters Workshop).                                   X.400 Message Handling System (MHS).

w                                                    w
AP (Application Process)                             APPC (Advanced Program-to-
                                                     Program Communications)
    In the OSI Reference Model, a program that
    can make use of application layer services.          In IBM’s SAA (Systems Application Archi-
    Application service elements (ASEs) provide          tecture), APPC is a collection of protocols
    the requested services for the AP.                   to enable executing applications to commu-
                                                         nicate directly with each other as peers
                                                         (without intervention by a mainframe host).
                                                                                   AppleTalk     33

       APPC is defined at a level comparable             C O M P A RE
    to the session layer in the OSI Reference             AppleDouble                                    a
    Model. It can be supported in various net-
    working environments, including IBM’s
                                                    w                                                    b
    SNA (System Network Architecture),                                                                   c
    Ethernet, Token Ring, and X.25.                     AppleTalk is Apple’s proprietary protocol
       APPC/PC (Advanced Program-to-                    suite for Macintosh network communica-           d
    Program Communications/Personal Com-                tions. It provides a multilayer, peer-to-peer
                                                        architecture that uses services built into the
    puters) is a PC-based version of APPC.
                                                        operating system. This gives every Macin-        f
w                                                       tosh networking capabilities. AppleTalk can
AppleDouble                                             run under any of several network operating
    In the Macintosh world, a file format that           systems, including Apple’s AppleShare,           h
    uses separate files for the data and resource        Novell’s NetWare for Macintosh, and Sun
    forks that make up a Macintosh file. This            Microsystems’ TOPS.                              i
    enables the files—or at least the data por-             AppleTalk was developed in the mid-
    tion—to be used on different platforms.             1980s with the goal of providing a simple,
    C O M P A RE
                                                        portable, easy-to-use, and open networking       k
                                                        environment. To access such a network,
                                                        a user just needs to “plug in, log in, and       l
w                                                       join in.”
                                                           A newer version, Phase 2, was released in
    A network operating system from Apple.              1989. This version provided some new capa-       n
                                                        bilities and extended others.
    AppleShare runs on a Macintosh network                                                               o
    server, providing file and printer services.
    AppleShare uses the AppleTalk protocol
                                                    AppleTalk Layers
    suite to carry out its tasks.                       AppleTalk is a comprehensive, layered envi-
                                                        ronment. It covers networking services over
                                                        almost the entire range of layers specified in    r
                                                        the OSI Reference Model. The figure “The
w                                                       AppleTalk protocol hierarchy” shows the          s
AppleSingle                                             organization of the AppleTalk layers, as well
                                                        as the protocols in the AppleTalk Protocol
    In the Macintosh world, a file format that
    stores both a file’s contents (data fork) and        Suite.                                           u
    its resources (resource fork) within a single                                                        v
    file. Because data and resources are mixed in
    a proprietary format, such a file cannot be                                                           w
    used on other platforms.
34     AppleTalk


                                                                               AppleTalk     35

Physical and Data-Link Layers                          s   FDDITalk, Apple’s implementation
                                                           of the 100 Mbps FDDI architecture.
  There are AppleTalk implementations for
  the following network architectures at the          For each of these architectures, a Link        b
  physical and data-link layers:                    Access Protocol (LAP) is defined: LLAP for
                                                    LocalTalk, ELAP for EtherTalk, TLAP for
     s   Apple’s 230 kilobit per second (Kbps).     TokenTalk, and FLAP for FDDITalk.                d
         LocalTalk architecture. LocalTalk pro-

         vides a media-access method and a        Network Layer
         cabling scheme for AppleTalk. The
                                                    All AppleTalk networks use the DDP (Data-        f
         architecture uses twisted-pair cables
         and RS-422 connections, allows nodes
                                                    gram Delivery Protocol) at the network           g
                                                    layer, regardless of the architecture operat-
         to be separated by as much as 305
         meters (1,000 feet), and can transmit
                                                    ing at the data-link layer. This protocol        h
                                                    makes a best effort at packet delivery, but
         at up to 230.4 Kbps. The term Local-
                                                    delivery is not guaranteed.
         Talk is sometimes used to refer to an
         AppleTalk network.
                                                       Note also the AARP (AppleTalk Address         j
                                                    Resolution Protocol) at this layer. The
     s   EtherTalk, Apple’s implementation          AARP maps AppleTalk (network) addresses          k
         of the 10 megabit per second (Mbps)
         Ethernet architecture. Two versions
                                                    to Ethernet or Token Ring (physical)
         of EtherTalk exist. The earlier one,                                                        m
         EtherTalk Phase 1, is modeled on the     Higher Layers
         Blue Book Ethernet 2.0 (as opposed
                                                    For reliable packet delivery, the ADSP
         to the version specified in the IEEE
                                                    (AppleTalk Data Stream Protocol) and
         802.3 documentation). Its successor,
         Phase 2, is modeled on the IEEE 802.3
                                                    ATP (AppleTalk Transaction Protocol)             p
                                                    are available. Each of these protocols is
         standard. Because these two variants
         of Ethernet define packets somewhat
                                                    appropriate under different conditions.          q
                                                       The NBP (Name Binding Protocol) and
         differently, Phase 1 and Phase 2 nodes
                                                    ZIP (Zone Information Protocol) help make
         cannot communicate directly with
         each other. EtherTalk has replaced
                                                    addressing easier. NBP associates easy-to-       s
                                                    remember names (used by users) with the
         LocalTalk as the default networking
                                                    appropriate address.                             t
         capability in newer Macintosh models.
         TokenTalk, Apple’s implementation of
                                                       ZIP is used mainly on larger networks or      u
     s                                              internetworks, which are more likely to be
         the token-ring architecture. AppleTalk     divided into zones. A zone is a logical group-   v
         supports both the 4-Mbps version           ing of nodes that together make up a subnet-
         specified by IEEE 802.5 and the 16-         work. The concept of a zone was introduced
         Mbps version from IBM. The token-          to allow for larger networks with more than      x
         ring architecture is supported only in     255 nodes, and also to make addressing and
         AppleTalk Phase 2.                         routing tasks easier.                            y
36      AppleTalk

     Applications access an AppleTalk net-       ASP (AppleTalk Session Protocol): A
  work through the AFP (AppleTalk Filing           session-layer protocol used to begin
  Protocol); they access printer services by       and end sessions, send commands
  shipping PostScript files through the PAP         from client to server, and send replies
  (Printer Access Protocol).                       from server to client.
     A few protocols make use of services
                                                 ATP (AppleTalk Transaction Protocol): A
  from more than one lower-level protocol.
                                                   transport-layer protocol that can pro-
  For example, ZIP relies on ATP and DDP
                                                   vide reliable packet transport. Packets
                                                   are transported within the framework
AppleTalk Protocol Suite                           of a transaction (an interaction
                                                   between a requesting and a responding
  The following protocols make up the Apple-       entity {program or node}).
  Talk Protocol Suite (see the figure “The
  AppleTalk protocol hierarchy,” earlier in      AURP (AppleTalk Update Routing
  this article):                                   Protocol): A transport-layer routing
                                                   protocol that is similar to RTMP
     AARP (AppleTalk Address Resolution            (Routing Table Maintenance Proto-
       Protocol): A network-layer protocol         col) but that updates the routing table
       that maps AppleTalk (network)               only when a change has been made to
       addresses to physical addresses.            the network.
     ADSP (AppleTalk Data Stream Protocol):      DDP (Datagram Delivery Protocol): A
       A session-layer protocol that allows        network-layer protocol that prepares
       two nodes to establish a reliable con-      and routes packets for transmission on
       nection through which data can be           the network.
                                                 LAP (Link Access Protocol): Works at
     AEP (AppleTalk Echo Protocol): A              the data-link layer, converting packets
       transport-layer protocol used to deter-     from higher layers into the appropriate
       mine whether two nodes are connected        form for the physical transmission.
       and both available.                         Each network architecture needs its
     AFP (AppleTalk Filing Protocol): A pre-       own LAP.
       sentation/application-layer protocol      ELAP (EtherTalk Link Access Protocol):
       used by applications to communicate         The link-access protocol used for
       with the network.                           Ethernet networks.
     ASDSP (AppleTalk Safe Data Stream           FLAP (FDDITalk Link Access Protocol):
       Protocol): A session-layer protocol         The link-access protocol used for
       that is similar to ADSP but that pro-       FDDI networks.
       vides additional security against
       unauthorized use.
                                                                             AppleTalk      37

  LLAP (LocalTalk Link Access Protocol):
    The link-access protocol used for
                                                Numbers and Zones
                                                  In AppleTalk networks, every node has an
    LocalTalk networks.
                                                  official numerical address. In addition, a
  TLAP (TokenTalk Link Access Proto-
    col): The link-access protocol used
                                                  node may be part of a named group of            c
                                                  nodes, which somehow belong together.
    for Token Ring networks.                                                                      d
                                                Network and Node Numbers
  ARAP (AppleTalk Remote Access                                                                   e
    Protocol): A link-access protocol for         Each AppleTalk network is assigned a
    accessing the network from a remote           unique network number, and each node in         f
    location over a serial line.                  that network is assigned this number. Pack-     g
                                                  ets addressed to a node on the network must
  NBP (Name Binding Protocol): A
    transport-layer protocol that associ-
                                                  include the network number.                     h
                                                     In addition to a network number, each
    ates device names with network
                                                  node has a node number that is unique
    addresses. If the NBP is successful, this
    binding process will be completely
                                                  within that network. This is an 8-bit number    j
                                                  and can be any value between 1 and 254,
    transparent to the user.
                                                  inclusive (0 and 255 are reserved as node       k
  PAP (Printer Access Protocol): A session-
    layer protocol for creating a path from
                                                  numbers). However, servers must have node
                                                  numbers within the range of 128 to 254,
    the user or application to a printer.         and workstations must have numbers in           m
                                                  the 1 to 127 range.
  RTMP (Routing Table Maintenance                                                                 n
    Protocol): A transport-layer routing
    protocol for moving packets between
                                                Zones                                             o
    networks.                                     A zone is a logical grouping of nodes. The      p
                                                  basis for the grouping can be any criterion
  ZIP (Zone Information Protocol): A
    session-layer protocol used to help           that is useful for a particular configuration,   q
    find a node; for example, in a large           as in the following examples:
    internetwork.                                       Geographical, such as all machines on

                                                        the second floor
   If installed, an AppleShare server runs
on top of these protocols at the uppermost          s   Departmental, such as all machines in     t
(application) layer. The AppleShare server
uses the AFP to provide centralized file shar-
                                                        the marketing department                  u
ing for its clients, and can use the PAP to
                                                    s   Functional, such as all machines that
                                                        can provide access to printers
provide printer sharing.
                                                     By restricting routing or searches to
                                                  machines in a particular zone, network traf-    x
                                                  fic and work can be reduced considerably.
 38       AppleTalk

  Accessing resources by zones also makes             When you are assigning number ranges,
  it easier to determine what is available for     a rough guideline is to assign one network
  specific needs.                                   number for every 25 to 50 nodes. If you
      A node may belong to more than one           expect a lot of growth, use a smaller num-
  zone at the same time, or not be part of any     ber. For example, assigning two network
  zone. A zone can cross network boundaries;       numbers for a 100-node network leaves
  that is, a zone can consist of parts of two or   room for 406 additional nodes.
  more different networks or include multiple         When a network is part of an internet-
  networks.                                        work, there are several restrictions on what
                                                   can be connected and how. These restric-
Phase 2 AppleTalk                                  tions concern routers and bridges, and the
  Phase 2, an updated version of AppleTalk,        networks they can connect, as follows:
  was released in 1989. This version provides        s   All routers connected to a particular
  several improvements over Phase 1, includ-             network must use the same network
  ing the following:                                     number range for the interface with
      s   Allows more than 254 nodes per                 that network. For example, if a router
          network                                        thinks the network uses numbers
                                                         1,000 to 1,009, another router con-
      s   Allows a network to be assigned more           nected to the same network cannot use
          than one network number                        1,002 to 1,008.
      s   Introduced the AppleTalk Internet          s   Routers must connect networks with
          Router, which allows up to eight               different number ranges that do not
          AppleTalk networks to be connected             overlap. This means that routers can-
                                                         not connect a network to itself and
Network Numbering in Phase 2                             that networks with overlapping net-
                                                         work numbers cannot interact with
  In AppleTalk Phase 2, a network can be
                                                         each other.
  assigned a range of network numbers. A
  particular node on this network can be asso-       s   A bridge must connect network seg-
  ciated with any one number in this range. By           ments with the same number range.
  providing multiple network numbers for a
                                                   The figure “Rules for connecting AppleTalk
  single network, it is possible to have more
                                                   Phase 2 internetworks” illustrates these
  than the 254 nodes allowed in a Phase 1 net-
  work, because each network number can
  support 253 (yes, 253) individual nodes.
                                                                                          Application   39

    R U L E S F O R C O N N E C T I N G A P P L E TA L K P H A S E 2 I N T E R N E T WO R K S                  a
w                                                         Stand-Alone Applications
AppleTalk Networking Forum (ANF)
                                                             A stand-alone application can execute only
       ANF (AppleTalk Networking Forum)
                                                             one version of itself at a time and can sup-      q
                                                             port only a single user at a time. This type of
w                                                            application executes on a single machine,         r
Application                                                  which may or may not be connected to a            s
    An application is a program that calls oper-             network. Single-user versions of spread-
    ating system services and performs work,                 sheet, graphics, and database programs            t
                                                             are examples of stand-alone applications.
    such as data creation or manipulation, for                                                                 u
    the user. Applications may be stand-alone,
    network-based, or part of an integrated
                                                          Network-Based Applications
    package.                                                 A network-based application executes on a
                                                             network and is aware of the network, which
 40     Application Entity (AE)

  means that it can use networking conven-             data. Microsoft Office, Lotus SmartSuite,
  tions, elements, resources (such as print            and Borland Office are examples of such
  spoolers and cache buffers), and devices             integrated applications.
  (such as printers, modems, and backup            Accessing Networks from Applications
      This type of application can be used by          Users may access networks through or for
  multiple users at the same time. Applications        applications. For example, an application
  differ in the number of allowable users and          may use a network resource or may need
  in the measures taken to enforce restrictions        to communicate with an application on
  and to make sure users do not ruin other             another machine. Or a user may log in to a
  users’ data. Network and data protection             network with the specific intention of using
  measures include the use of flags, access             an application available on that network.
  rights, and lock-outs. These serve to help              Regardless of the details, such network
  ensure that data is used correctly, only as          accesses are through the topmost layer in the
  needed, and with fair access to all users.           OSI Reference Model: the application layer.
      Network-based applications may execute           This layer provides users and programs with
  on a single machine or be distributed over           an interface to the network. At this layer,
  multiple machines. Client/server computing           both the user and the application are iso-
  is an example of a distributed arrangement           lated from the details of network access and
  in which part of an application (the front           communication.
  end) executes on the workstation to provide
  an interface for the user, and another part          S H A R I N G DATA A M O N G
                                                       A P P L I C AT I O N S
  (the back end) executes on a server to do the
  actual work, such as searching a database.           Separate applications can also communicate and
      A network-based application may be               exchange data. Using pipes, in which the output
  multiuser or multilaunch. Only one copy of           from one program is simply “piped” in as input to
  a multiuser application executes, but multi-         another program is one of the simplest ways to
  ple users can access files in this executing          share data.
  program. A multilaunch application allows
  multiple users to execute the program sepa-          OLE (object linking and embedding) is a more
  rately but at the same time. In effect, each         sophisticated method, which provides much
  user gets a private version of a multilaunch         greater flexibility. OLE makes it possible for
  application.                                         updates to be carried over automatically to
                                                       whatever applications use the updated items.
Integrated Applications
  An integrated application is part of a collec-
                                                   Application Entity (AE)
  tion, or suite, of programs. Ideally, these
  programs complement each other in their              SEE
  functionality and allow easy exchange of                AE (Application Entity)
                                                                                  Archie     41

w                                                  Architecture) environment. APPN allows
Application File Transfer (AFT)                    peer-to-peer communications between com-
    SEE                                            puters without requiring a mainframe in the
      AFT (Application File Transfer)
                                                      APPN is also supported within IBM’s            c
w                                                  SNA (Systems Network Architecture) envi-
Application Layer                                  ronment. Unlike standard SNA, however,
    The topmost layer in the seven-layer OSI       APPN supports dynamic routing of packets.         e
    Reference Model.
                                                   B RO A D E R C A T E G O R Y                      f
    SEE ALSO                                          SAA (Systems Application Architecture)
      OSI Reference Model
w                                              ARA (Attribute Registration                           h
Application Process (AP)                       Authority)
    SEE                                            In the X.400 Message Handling System
      AP (Application Process)                     (MHS), the organization that allocates            j
                                                   unique attribute values.
Application Program Interface (API)
                                               Archie                                                l
      API (Application Program Interface)
                                                   An Internet service that can find the location     m
                                                   of specified files based on the file’s name or
w                                                  description. An archie server gets its infor-
Application Program Interface                      mation by using the FTP program to do a           o
Association (APIA)                                 listing of files on accessible servers and also
                                                   by getting file description information. Cur-      p
      APIA (Application Program Interface
                                                   rently, archie servers have data about over
                                                   2.5 million files on over 1,000 servers.
                                                       Archie servers are scattered throughout       r
w                                                  the Internet, and are accessible using services
Application Protocol Data Unit                     such as telnet or gopher, through e-mail, or      s
(APDU)                                             by using archie client programs. Archie serv-
                                                   ers should be equivalent (except for minor
      APDU (Application Protocol Data Unit)        differences arising because not all servers are   u
                                                   updated at the same time), so selecting a
w                                                  server is just a matter of convenience. See
APPN (Advanced Peer-to-Peer
                                                   the table “Example Archie Servers” for a          w
                                                   list of some of the available servers.
    APPN is a network architecture defined
    within IBM’s SAA (Systems Application                                                            y
42          Architecture

  E X A M P L E A RC H I E S E R VE R S                                   followed by a regular expres-
  S E RV E R S                 L O C AT I O N                             sion, the command displays
                                                                          only the servers that match
  archie.ac.il                 Israel
                                                                          the expression.
  archie.au                    Australia
  archie.doc.ic.ad.uk          United Kingdom                servers      Displays a list of all the avail-
  archie.edvz.uni-linz.ac.at   Austria                                    able archie servers.
  archie.funet.fi               Finland
  archie.kr                    Korea                         version      Displays the version number
  archie.mcgill.ca             Canada (McGill                             of the archie server you’re
                               University)                                querying. Such information
  archie.ncu.edu.tw            Taiwan                                     will come in handy if you
  archie.rediris.es            Spain                                      need to get help with the
  archie.rutgers.edu           USA (Rutgers                               program.
  archie.sura.net              (SURAnet is a service          Various other commands and configu-
                               provider)                   ration possibilities are available to make
  archie.switch.ch             Switzerland                 archie more useful and more convenient
  archie.th-darmstadt.de       Germany                     to use.
  archie.unipi.it              Italy
  archie.univ-rennes1.fr       France                  w
  archie.unl.edu               USA (University of      Architecture
                               Nebraska, Lincoln)          Architecture is an amorphous term in the
  archie.wide.ad.jp            Japan                       area of networking. The term can refer to
                                                           both the physical layout (topology) of the
                                                           network and also the protocols (communi-
Useful Archie Commands
                                                           cation rules and data elements) used to
  Once a connection has been established with              communicate.
  the archie server, various commands are                      Architecture can also refer to the basic
  available. The following list summarizes                 structure of a networking service, such as
  some useful ones:                                        a print service architecture. Used this way,
                                                           it generally indicates the overall scheme
     help             Displays a list of available         of APIs (Application Program Interfaces),
                      commands.                            agents, and so on, used to fit different pieces
     manpage          Displays the reference manual        of the service together.
                      for archie.                              You will hear references to network
                                                           architectures, such as ARCnet, Ethernet, and
     list             Displays a list of the anony-        Token Ring, which are all defined primarily
                      mous STP servers whose con-          at the two lowest layers of the OSI model:
                      tents are listed in archie’s         the physical and data-link layers. Each
                      database. If this command is         architecture includes an implicit topology.
                                           ARCnet (Attached Resource Computer Network)       43

       In the context of hardware, the term          network by Datapoint Corporation in the
    refers to the manner in which a computer is      late 1970s. ARCnet became very popular
    constructed. The architecture includes the
    type of processor (for example, Intel 80x86
                                                     when Standard Microsystems Corporation
                                                     (SMC) developed a chip set for PCs. The
    or Pentium, Motorola 680xx, or RISC chip)        architecture has been used for years and has   c
    and the type of bus that is used to transmit     become a de facto standard. However, it has
    data and other signals to the computer’s         not become as popular as other network
    components and peripherals.                      architectures, such as Ethernet. ARCnet is     e
       In the IBM PC world, which is currently       popular for smaller networks because it is
    dominated by Intel processors, the three         relatively simple to set up and operate, its   f
    major buses are ISA (Industry Standard           components are inexpensive (street prices      g
    Architecture), EISA (Extended Industry           for ARCnet boards are among the lowest),
    Standard Architecture), and MCA (Micro-          and the architecture is widely supported.      h
    channel Architecture). However, two newer            ARCnet has a transmission rate of 2.5
    bus designs—VL (VESA Local) and PCI              megabytes per second (Mbps). ARCnet Plus
    (Peripheral Component Interconnect)—are          is a newer, 20 Mbps version. A third-party,    j
    growing in popularity and are likely to          100 Mbps architecture based on ARCnet
    become the dominant bus architectures.           is also available from Thomas-Conrad.          k
                                                     Although ARCnet Plus was developed by
                                                     Datapoint Corporation alone, current and
      Network Architecture
                                                     future development of ARCnet standards is      m
                                                     under the aegis of the ATA (ARCnet Trade
Archive                                              Association), a consortium of vendors that     n
    As a noun, a repository for data, applica-       market ARCnet products.                        o
    tions, and so forth. These materials may be          ARCnet uses token passing to control
    master copies or regular backups of the cur-     access to the network. Each node in an         p
    rent hard disk contents. As a verb, the act of   ARCnet network has a unique address
                                                     (between 1 and 255), and the token is
    backing up data files to provide a safe copy
    in case of a disaster.                           passed sequentially from one address to the    r
                                                     next. Nodes with successive addresses are
w                                                    not necessarily next to each other in the      s
Archive Site
    On the Internet, a node that provides access
                                                     physical layout.
                                                         Officially, ARCnet uses a bus topology,
    to a collection of files.                         but in practice ARCnet networks can use a      u
                                                     star or a bus wiring scheme. These two types
                                                     of networks use slightly different compo-
ARCnet (Attached Resource
Computer Network)
                                                     nents and are sometimes referred to as low-    w
                                                     impedance and high-impedance ARCnet,
    ARCnet is a baseband network architec-           respectively.                                  x
    ture originally developed as a proprietary
44        ARCnet (Attached Resource Computer Network)

    The figure “Context and properties of            ARCnet Network Components
 ARCnet” summarizes the characteristics
 of this architecture.                                The hardware components needed in an
                                                      ARCnet network include an ARCnet net-
 C O N T EX T A N D P RO P E R T I E S                work interface card, cable, connectors, hubs,
 OF ARCNET                                            active links, and baluns.

Context                                             ARCnet Network Interface Card (NIC)

Network Architecture                                  ARCnet NICs include chips to handle
  Shared-Media                                        the ARCnet protocols and packet formats,
      ARCnet                                          as well as a transceiver (usually with a BNC
      Ethernet                                        connector) on the card. Most ARCnet NICs
      Token Ring                                      have a low-impedance transceiver, which is
  Switched Media
                                                      best suited for a star or tree topology. (A tree
                                                      topology has features of both star and bus
                                                      topologies.) Cards with high-impedance
                                                      transceivers are suitable for a bus topology.
  Description      Shared-media, baseband network
                                                         ARCnet cards do not come with hard-
  Topology         Bus (high-impedance ARCnet)        ware addresses in a ROM chip. Instead, they
                   Star (low-impedance ARCnet)        have jumpers that can be set to specify an
                                                      address for the node in which the card is
  Access method    Token passing
                                                      installed. The network administrator needs
  Speed            Up to 2.5 Mbps                     to set this address (which must be between 1
                                                      and 255) for each card in the network. Each
  Cable            RG-62 coaxial (93-ohm)             node must have a unique address. The net-
                   Unshielded twisted-pair
                                                      work administrator also needs to set the
                                                      IRQ (interrupt) and I/O (input/output)
  Frame size       Up to 508 data bytes               addresses on the card. The hardware address
                                                      is network-dependent; the IRQ and I/O
  Variants         High-impedance ARCnet
                                                      addresses are machine-dependent.
                   Low-impedance ARCnet
                   Mixed-impedance ARCnet
                   ARCnet Plus                      Cable

                                                      ARCnet cable can be coaxial, twisted-pair,
                                                      or even fiber-optic. Coaxial ARCnet net-
                                                      works generally have RG-62 cable, which
                                          ARCnet (Attached Resource Computer Network)           45

  has a 93-ohm impedance. Other types of            Hubs
  coaxial cable, such as RG-59U or RG-11U,
                                                      Hubs serve as wiring concentrators. Three
  are also used.
      An ARCnet network might include
                                                      types of hubs can be used:                       b
  unshielded twisted-pair (UTP) or IBM’s                   Active hubs: Active hubs have their own     c
  special-design cables (Types 1 and 3), but                 power supply. They can clean and
  only if the NIC has the appropriate connec-                boost a signal and then relay it along
  tors or if an appropriate adapter is available.            the network. An active hub serves as      e
  If UTP cabling is used, nodes are arranged                 both a repeater and a wiring center.
  in a daisy chain and one end of the chain                  Active hubs usually have 8 ports, but     f
  is connected to a hub or to an adapter that                they can have as many as 64. The type     g
  connects to coaxial cable. Similar converters              of hub used must be appropriate for
  can convert from coaxial to fiber-optic                     the type of cable being used. Active      h
  cable.                                                     hubs can extend the maximum dis-
      The last node in an ARCnet network                     tance between nodes.
  must be terminated with a resistor of appro-
                                                           Passive hubs: Passive hubs simply relay     j
  priate strength: 93 ohm for coaxial net-
  works and 105 ohm for networks using
                                                             signals without cleaning or boosting      k
                                                             them. These types of hubs collect wir-
  twisted-pair wiring.
                                                             ing from nodes and must be connected      l
                                                             to an active hub. Passive hubs have
Connectors, Active Links, and Baluns
                                                             four ports and are used only in low-
  For coaxial cable, BNC connectors are                      impedance networks. Passive hubs          n
  used. For twisted-pair cable, the connectors               cannot be used to extend the distance
  are either the modular RJ-11/RJ-45 tele-                   between nodes.                            o
  phone type, or the D-shell type used for
                                                           Intelligent hubs: Intelligent hubs are      p
  standard serial and parallel ports.
     Active links are boxes used to connect
                                                              active hubs that use a low-frequency
                                                              signal band to monitor the status of
  two cable segments when both cable
  segments have high-impedance NICs
                                                              a link. These hubs can have up to        r
                                                              16 ports.
  connected.                                                                                           s
     Baluns are used to connect coaxial and         ARCnet Operation
  twisted-pair cabling.                                                                                t
                                                      ARCnet data transmissions are broadcast to
                                                      all nodes on the network (a feature charac-
                                                      teristic of both bus and star topologies), but   v
 46      ARCnet (Attached Resource Computer Network)

  the transmitted packets are (presumably)               bit sequence 110, so that each byte actually
  read only by the node(s) to which the desti-           requires 11 bits in an ARCnet transmission.
  nation address applies. Note that even                    ARCnet data frames consist of data,
  though all nodes can listen at the same time,          header, and trailer. Originally, an ARCnet
  only one node can transmit.                            frame could have up to 252 bytes of data.
                                                         Almost all ARCnet implementations now
Structure of an ARCnet Packet                            support an expanded frame of up to 508
                                                         bytes of data (plus a dozen or so header
  ARCnet has several different types of
  frames, or packets, which are listed on the
                                                            An ARCnet header for a PAC frame
  table “ARCnet Packets.” The figure “ARC-
                                                         includes the following:
  net frame structure” shows the makeup of
  ARCnet frames.                                            s   A start of header byte

  A R C N E T P A C KE T S                                  s   Source and destination addresses,
  PAC K E T T Y P E     FUNCTION                                with values between 1 and 255 (a des-
  ITT (Invitation to    The token, which deter-                 tination address of 0 indicates that the
   Transmit)            mines the node that is                  frame is being broadcast to all nodes)
                        allowed to transmit                 s   One or two bytes indicating the num-
  FBE (Free Buffer      The frame that is used to               ber of data bytes
   Enquiry)             ask whether the destina-
                                                           The trailer is a 16-bit CRC (cyclic redun-
                        tion node is able to receive
                                                         dancy check) value.

  ACK (Acknowledge)     The packet used to indi-       Data Frame Transmission
                        cate that a packet was
                        received as transmitted          The transmission of data frames in an ARC-
                                                         net network is controlled by a token, which
  NAK (Negative         The packet used to indi-
                                                         is a special data frame. This token, in turn, is
  Acknowledge)          cate that a packet was not
                                                         dispensed by the network’s controller, which
                        received correctly and
                        should be retransmitted
                                                         is the node with the lowest address. The
                                                         controller is determined when the network
  PAC                   The actual ARCnet data           is first activated. Each node broadcasts
                        frame                            its address and the node with the lowest
                                                         address becomes the controller. This recon-
                                                         figuration process, which takes less than a
     The data, control, or check bytes that
                                                         tenth of a second, is repeated each time a
  make up the frame are known as ISUs
                                                         new node joins the network.
  (information symbol units). ISUs are defined
                                                             The controller passes the token sequen-
  differently in ARCnet and in ARCnet Plus.
                                                         tially from one address to the next. The
     All ARCnet frames begin with a 6-bit
                                                         node with the token is the only node
  alert signal, and all bytes begin with the
                                                         allowed to transmit, with some exceptions.
                                                ARCnet (Attached Resource Computer Network)                    47

 A R C N E T F RA M E S T R U C T U RE                                                                              a
                                                                 Frame components are symbols containing            b
                                                                 the following:
               1      1      1       1      1                   SD      Starting delimiter, a special bit
                                                                        pattern of six consecutive 1bits, to        d
              Starting Delimiter (SD)                                   indicate the start of the frame
                                                                EQT     ASCII 0x04, which indicates the
                                                                        frame type                                  f
                          ITT Frame
                                                                NID     The address of the next node to             g
              SD EQT NID NID                                            get the token
                                                                ENQ     ASCII 0x85, which identifies the
                                                                        frame type
                                     FBE Frame                  DID      The address of the destination node        j
                          SD ENQ DID DID                                for the enquiry                             k
                                                                ACK     ASCII 0x86, indicating that the
                                                                        packet was recieved correctly
                                    ACK Frame                   NAK     ASCII 0x15, indicating that the
                                         SD ACK                         packet was not recieved correctly           n
                                                                SOH      ASCII 0x01, indicating the start of
                                                                        the header
                                                                SID      The address of the source node
                                    NAK Frame
                                                                        sending the frame                           q
                                         SD NAK
                                                                CP      A continuation pointer value,
                                                                        indicating the number of data bytes
                                                                SC      System code
PAC Frame                        1 or 2 1 or 2 0-508   2
SD SOH SID DID DID CP                    SC DATA FSC            DATA    Up to 508 symbols containing
                                                                        system code and data                        u
                                                                FCS     Frame check sequence, verifying the
                                                                        integrity of the frame
 48      ARCnet (Attached Resource Computer Network)

     Frame transmission is a complicated pro-          themselves, and the network begins trans-
  cess in ARCnet. A node (the source) waiting          mitting again.
  to send a message to another node (the desti-           New nodes on an ARCnet network also
  nation) needs to do several things, in the fol-      send a reconfiguration burst. This pattern
  lowing order:                                        announces their presence on the network,
                                                       and possibly establishes a new node as
      1. The source waits for the token (ITT).
      2. Once it has the token, the source sends
         an FBE packet to the destination to         Communicating with Higher Layers
         make sure the destination has room for
                                                       ARCnet’s small frame size causes compati-
         the frame.
                                                       bility problems with some network-layer
      3. The source waits for a positive reply.        protocols, such as Novell’s IPX protocol.
      4. Once the source gets a positive               IPX passes 576-byte packets (known as
         response (ACK) to the FBE packet,             datagrams) to the architecture operating at
         the source broadcasts the frame.              the data-link layer. This packet size is too
                                                       large, even for an extended ARCnet frame.
      5. The source waits for an acknowl-                 To enable IPX to talk to ARCnet, the
         edgment from the intended desti-              fragmentation layer was developed. At this
         nation. The destination node must             layer, the source node breaks an IPX packet
         acknowledge receipt of the frame.             into two smaller frames for ARCnet. At the
         Since acknowledgment is required,             destination’s fragmentation layer, the data-
         ARCnet can guarantee frame delivery.          gram is reassembled before being passed
      6. Once the frame has been received              to IPX.
         at the destination, the controller passes
                                                     High-Impedance ARCnet
         the token to the next address.
                                                       High-impedance ARCnet networks use a
Disrupting Data Transmission                           bus topology, as illustrated in the figure
                                                       “Layout for a high-impedance ARCnet
  Unless something is wrong on the network,
                                                       network.” The high-impedance NICs make
  every node gets the token at least once every
                                                       it possible to daisy chain nodes and active
  840 milliseconds. If a node has not seen the
                                                       hubs. The active hubs serve as collectors for
  token within that time, that node can dis-
                                                       other hubs and nodes.
  rupt the network and force the creation of
                                                           The following restrictions apply to high-
  a new token by sending a reconfiguration
                                                       impedance ARCnet networks:
  burst—a predefined bit pattern sent hun-
  dreds of times in succession—to destroy the             s   No single cable segment connecting
  existing token. After a period, the token is                nodes can be be more than 305 meters
  regenerated, the network nodes reannounce                   (1,000 feet) long.
                                         ARCnet (Attached Resource Computer Network)   49

L AYO U T F O R A H I G H - I M P E D A N C E A R C N E T N E T WO R K                      a
50       ARCnet (Attached Resource Computer Network)

     s   Only active (or intelligent) hubs may     with nodes. In the latter case, the active hub
         be used.                                  also acts as a wiring center.
                                                      The following restrictions apply to low-
     s   Adjacent active hubs (hubs with no
                                                   impedance ARCnet networks:
         intervening nodes) must be within 610
         meters (2,000 feet).                        s   Active hubs can be connected to
                                                         nodes, active hubs, or passive hubs.
     s   Nodes are connected to the trunk cable
                                                         The active hub must be within 610
         using BNC T-connectors. The node’s
                                                         meters (2,000 feet) of an active hub or
         NIC must be connected directly to the
                                                         a node, or within 30 meters (100 feet)
         T-connector; that is, drop cable is not
                                                         of a passive hub.
                                                     s   Passive hubs can be used only between
     s   T-connectors must be at least 1 meter
                                                         a node and an active hub; two passive
         (3.25 feet) apart on the cable.
                                                         hubs cannot be next to each other. A
     s   At most, eight nodes can be connected           passive hub must be within 30 meters
         in a series (with no intervening hubs).         (100 feet) of an active hub and within
                                                         30 meters (100 feet) of a node.
     s   Both ends of a cable segment must be
         terminated with either a BNC termina-       s   Nodes can be attached anywhere on
         tor or an active hub (or link).                 the network, provided the node is
                                                         within the required distance of an
     s   The cabling cannot loop back on itself.
                                                         active or passive hub: within 610
         For example, the cable cannot go from
                                                         meters (2,000 feet) of an active hub
         an active hub through other hubs and
                                                         or within 30 meters (100 feet) of a
         eventually connect back into the origi-
                                                         passive hub.
         nal hub.
                                                     s   Unused hub ports must be terminated
Low-Impedance ARCnet                                     on a passive hub and should be termi-
  Low-impedance ARCnet networks use a star               nated on an active hub.
  topology, in which passive hubs serve to col-      s   The cabling cannot loop back on itself.
  lect nodes, as illustrated in the figure “Lay-          For example, the cable cannot go from
  out for a low-impedance ARCnet network.”               an active hub through other hubs and
  Each passive hub is connected to an active             eventually connect back into the origi-
  hub. Active hubs can be linked with each               nal hub.
  other, and they can also be linked directly
                                        ARCnet (Attached Resource Computer Network)   51

L AYO U T F O R A L OW - I M P E D A N C E A R C N E T N E T WO R K                        a
52       ARCnet (Attached Resource Computer Network)

Mixed-Impedance ARCnet                                IBM Type 1 (shielded) cable:
                                                        200 meters (660 feet)
  A mixed ARCnet network is one that
  includes both high- and low-impedance
                                                      s   The maximum cable length for the
  components in the same network, as                      entire network is 6,000 meters (20,000
  illustrated in the figure “Layout of a                   feet)
  mixed-impedance ARCnet network, with                s   The maximum number of cable seg-
  low-impedance and high-impedance com-                   ments in a series is three. If UTP cable
  ponents.” In this type of network, all the              is used, the series of segments can be at
  restrictions for both impedance levels                  most about 130 meters (430 feet); for
  must be observed.                                       coaxial cable, the maximum length is
      Perhaps the most important constraint               about 300 meters (990 feet).
  for a mixed-impedance ARCnet is that high-
  impedance NICs can be used in place of
                                                      s   Each cable segment must be termi-
  low-impedance cards, but the reverse is not             nated at both ends by being connected
  possible. Because of this restriction, it is cru-       to an active hub or terminator.
  cial that you keep track of what kind of NIC        s   An ARCnet network can have a maxi-
  is in each node.                                        mum of 255 nodes. Each active hub
                                                          counts as a node.
Restrictions on ARCnet Networks
                                                      s   At most, 10 nodes are allowed in a
  The following restrictions apply to both                series when UTP cable is used; 8 nodes
  high- and low-impedance ARCnet networks:                if coaxial cable is used.
     s   The maximum length of a cable seg-           s   The maximum distance between any
         ment depends on the type of cable. The           two nodes on the network is deter-
         general restriction is that the signal           mined by the constraint that no ARC-
         attenuation must be less than 11 dB              net signal can have a propagation
         over the entire cable segment at a fre-          delay of more than 31 microseconds.
         quency of 5 MHz. In practice, this               The total propagation delay is deter-
         leads to the following maximum                   mined by adding the propagation
         distances:                                       delays in all the devices (nodes, hubs,
     Coaxial cable: 450–600 meters (1,500–                and cable) connecting the nodes. Net-
       2,000 feet)                                        work components generally have
                                                          propagation delays of less than 0.5
     UTP and IBM Type 3 (unshielded) cable:
                                                          microseconds, and much less in some
       100 meters (330 feet)
                                            ARCnet (Attached Resource Computer Network)   53

L AYO U T O F A M I XE D - I M P E D A N C E A R C N E T N E T WO R K ,                        a
W I T H L OW - I M P E D A N C E A N D H I G H - I M P E D A N C E C O M P O N E N T S
54       ARCnet (Attached Resource Computer Network)

ARCnet Advantages                                           small networks, the throughput is less
                                                            than 65 percent of maximum, and this
  ARCnet has the following advantages:                      value decreases as more nodes are
     s   Components are relatively inexpen-                 added to the network.
         sive. Street prices for basic ARCnet           s   The network administrator must man-
         NICs usually are less than those for               ually set a unique address by adjusting
         Ethernet or Token Ring NICs.                       switches on every NIC in the network.
     s   Because the ARCnet architecture and                If two nodes have the same address,
         the chip set have been around a long               the administrator will need to track
         time, the hardware has become stable.              down the conflicting boards by tedious
         The result is that there are few compat-           examination of each NIC.
         ibility or reliability problems with           s   Because of throughput and addressing
         ARCnet components.                                 restrictions, ARCnet is not particularly
     s   Wiring is very flexible, allowing for               well-suited for internetworking.
         lots of leeway in placing nodes.           ARCnet Plus
     s   It is relatively easy to use different
                                                      Datapoint recently released ARCnet Plus, a
         types of cabling in an ARCnet network
                                                      20-Mbps version of the ARCnet standard.
         (but adapters must be used to avoid
                                                      ARCnet Plus has the following features:
         connection incompatibilities).
                                                        s   Backward-compatibility with ARCnet
     s   A star layout makes diagnostics easy in
         low-impedance networks.                        s   Ability to communicate with both
                                                            ARCnet and ARCnet Plus nodes
     s   Except for the extra cabling a star
         topology requires, installation is rela-       s   Support for transmission rates of up to
         tively inexpensive.                                20 Mbps

ARCnet Disadvantages                                    s   Support for data frames up to 4,224
                                                            bytes long
  ARCnet has the following disadvantages:
                                                        s   Use of the same RG-62 cable as ordi-
     s   Its data transmission is inefficient.               nary ARCnet
         ARCnet sends three overhead bits
         for every byte. Also, administrative
                                                        s   New frames, with enhanced frame
         exchanges (such as ACK or NAK                      formats and command sets
         packets) between source and destina-           s   Support for up to 1 MB of buffer space
         tion are done on the data bandwidth,
         which degrades performance further.             ARCnet Plus achieves its greater speed by
                                                      cutting the time interval for a symbol in half
     s   Actual throughput is much less than          and by using phase and amplitude shifting to
         the maximum 2.5 Mbps. Even for               encode four bits in every signal; that is, the
                                     ARPAnet (Advanced Research Projects Agency Network)             55

basic symbol in ARCnet Plus is actually a          w
nibble.                                             ARF (Alarm Reporting Function)                        a
   Like its predecessor, ARCnet Plus regu-
lates much network activity by timing. The
                                                       In the OSI network management model,               b
                                                       a service that reports failures, faults, or
allowable intervals are much smaller with              problems that might become faults.                 c
ARCnet Plus, however. For example, a bit
interval is half as long in ARCnet Plus as         w                                                      d
                                                    ARM (Asynchronous Response Mode)
in regular ARCnet.                                                                                        e
   Another extension of this type of                   In the ISO’s HLDC (High-Level Data Link
architecture is TCNS, offered by Thomas-               Control) protocol, ARM is a communica-             f
Conrad, which is a 100 Mbps, copper-                   tions mode in which a secondary (slave)
                                                       node can initiate communications with a
based network.
                                                       primary (master) node without first getting         h
                                                       permission from the primary node.
   Network Architecture                                   ARM’s operation is in contrast to NRM           i
SEE ALSO                                               (normal response mode), in which the pri-          j
   TCNS (Thomas-Conrad Network                         mary node must initiate any communica-
   System)                                             tion, and to ABM (asynchronous balanced            k
                                                       mode), in which the two nodes are equal.
Keep accurate addresses. Make sure you have up-
                                                       B RO A D E R C A T E G O R Y
                                                          HDLC (High-Level Data Link Control)
to-date records of the address set for each ARC-
net node’s NIC. When you need to find duplicate
addresses or add nodes, you’ll be glad you did.
                                                    ARPA (Advanced Research Projects
If you’re the administrator, never let anyone
                                                       The agency that was largely responsible for
else change the node addresses, because you may
have to deal with the problems caused by their
                                                       what eventually became the Internet. Now           q
                                                       called DARPA (for Defense ARPA).
sloppiness.                                                                                               r
Assigning the low address is particularly impor-
tant. The network controller will be the node
                                                    ARPAnet (Advanced Research Projects                   s
                                                    Agency Network)
with the lowest address, so make sure this
                                                       ARPAnet was the first large-scale, packet-
machine is fast enough to handle the controlling
role. In general, it’s best to assign the lowest
                                                       switched, wide-area network (WAN). It was          u
                                                       originally developed in the early 1970s
addresses to servers, bridges, and routers.
                                                       under the auspices of the U.S. Department
                                                       of Defense’s Defense Advanced Research             w
                                                       Projects Agency (DARPA).
56         ARQ (Automatic Repeat Request)

       Many of the most commonly used net-                   In the OSI Reference Model, an autono-
    working protocols, including TCP/IP, were             mous system is known as a routing domain,
    developed as part of the ARPAnet project.             IGPs are known as intradomain routing pro-
    The ARPAnet was decommissioned in 1991,               tocols, and EGPs are known as interdomain
    but parts of the network have become part             routing protocols.
    of the Internet.
w                                                     AS/400
ARQ (Automatic Repeat Request)                            A minicomputer line from IBM. The AS/400
    In communications, a control code that                was introduced in 1988 to replace the Sys-
    indicates an error in transmission and                tem/36 and System/38 series.
    that requests a retransmission.
w                                                     ASCII (American Standard Code for
ARS (Automatic Route Selection)                       Information Interchange)
    In telephony, a process by which a path               ASCII is the character-encoding system used
    is selected for a transmission; also called           most commonly in local-area networks
    alternate route selection.                            (LANs). The standard ASCII characters are
                                                          encoded in seven bits and have values
                                                          between 0 and 127. The remaining 128
AS (Autonomous System)
                                                          characters form the extended ASCII charac-
    In the Internet world, AS (autonomous sys-            ter set, whose elements may be defined dif-
    tem) is a term for a collection of routers that       ferently depending on the language being
    are part of a larger network but that are             used. See the tables “Standard ASCII Char-
    under the control of a single organization.           acter Set” and “Extended ASCII Character
    The routers, or gateways as they are called           Set (IBM PC).”
    in the older Internet terminology, communi-              In common usage, ASCII is used to refer
    cate with each other using a common proto-            to a text-only file that does not include spe-
    col, known as an interior gateway protocol            cial formatting codes.
    (IGP). Currently, the two most widely sup-
    ported IGPs in the Internet community are             B RO A D E R C A T E G O R Y

    the OSPF (Open Shortest Path First) and the              Encoding
    Integrated IS-IS protocols.                           C O M P A RE
       ASs communicate using an exterior
    gateway protocol, such as EGP (Exterior
    Gateway Protocol) and BGP (Border Gate-
    way Protocol).
                                  ASCII (American Standard Code for Information Interchange)     57

DECIMAL       C H A R AC T E R         DECIMAL      C H A R AC T E R      DECIMAL    C H A R AC T E R
0             NUL (null)               21           NAK (negative         43         +                  b
1             SOH (start
                                                                          44         , (comma)          c
              of heading)              22           SYN (synchronous
                                                                          45         −                  d
2             STX (start of text)
                                       23           ETB (end trans-
                                                                          46         .                  e
3             ETX (end of text)
                                                    mission block)
4             EOT (end of
                                                                          47         /
                                       24           CAN (cancel)
                                                                          48         0
                                       25           EM (end of medium)    49         1
5             ENQ (enquire)
                                       26           SUB (substitute)
                                                                          50         2
6             ACK (acknowledge)
                                       27           ESC (escape)          51         3
7             BEL (bell)
                                       28           FS (file separator)    52         4
8             BS (backspace)

9             HT (horizontal tab)
                                       29           GS (group             53         5                  k
10            LF (line feed)
                                                                          54         6                  l
                                       30           RS (record
11            VT (vertical tab)                     separator)            55         7                  m
12            FF (form feed)           31           US (unit separator)   56         8                  n
13            CR (carriage return)     32           space                 57         9                  o
14            SO (shift out)           33           !                     58         :
15            SI (shift in)            34           "                     59         ;
                                                                          60         <
16            DLE (data link
                                       35           #
                                       36           $                     61         =
17            DC1 (device                                                 62         >
                                       37           %
              control 1)
                                       38           &                     63         ?                  t
18            DC2 (device
              control 2)               39           ’ (apostrophe)        64         @                  u
19            DC3 (device              40           (                     65         A                  v
              control 3)
                                       41           )                     66         B                  w
20            DC4 (device
              control 4)
                                       42           *                     67         C                  x
58    ASCII (American Standard Code for Information Interchange)


 68          D                  88          X                  108       l

 69          E                  89          Y                  109       m

 70          F                  90          Z                  110       n

 71          G                  91          [                  111       o

 72          H                  92          \                  112       p

 73          I                  93          ]                  113       q

 74          J                  94          ^                  114       r

 75          K                  95          _                  115       s

 76          L                  96          à                  116       t

 77          M                  97          a                  117       u

 78          N                  98          b                  118       v

 79          O                  99          c                  119       w

 80          P                  100         d                  120       x

 81          Q                  101         e                  121       y

 82          R                  102         f                  122       z

 83          S                  103         g                  123       {

 84          T                  104         h                  124       |

 85          U                  105         i                  125       }

 86          V                  106         j                  126       ~

 87          W                  107         k                  127       DEL
                                  ASCII (American Standard Code for Information Interchange)        59

DECIMAL        C H A R AC T E R      DECIMAL       C H A R AC T E R   DECIMAL    C H A R AC T E R
0                                    25            y                  50         2                       b
1              b                     26            z                  51         3                       c
2              a                     27            1                  52         4                       d
3              c                     28            2                  53         5                       e
4              d                     29            3                  54         6
5              f                     30            4                  55         7
6              e                     31                               56         8
7              g                     32                               57         9

8              h                     33            !                  58         :
9              i                     34            "                  59         ;
10             j                     35            #                  60         <                       k
11             k                     36            $                  61         =                       l
12             l                     37            %                  62         >                       m
13             m                     38            &                  63         ?                       n
14             n                     39            ’                  64         @                       o
15             o                     40            (                  65         A
16             p                     41            )                  66         B
17             q                     42            *                  67         C
18             r                     43            +                  68         D

19                                   44            ,                  69         E
20             t                     45            _                  70         F                       t
21             u                     46            .                  71         G                       u
22             v                     47            /                  72         H                       v
23             w                     48                               73         I                       w
24             x                     49            1                  74         J                       x
60    ASCII (American Standard Code for Information Interchange)


 75          K                  100         d                  125       }

 76          L                  101         e                  126       ˜

 77          M                  102         f                  127       ∆

 78          N                  103         g                  128       Ç

 79          O                  104         h                  129       ü

 80          P                  105         i                  130       é

 81          Q                  106         j                  131       â

 82          R                  107         k                  132       ä

 83          S                  108         l                  133       à

 84          T                  109         m                  134       å

 85          U                  110         n                  135       ç

 86          V                  111         o                  136       ê

 87          W                  112         p                  137       ë

 88          X                  113         q                  138       è

 89          Y                  114         r                  139       ï

 90          Z                  115         s                  140       î

 91          [                  116         t                  141       ì

 92          \                  117         u                  142       Ä

 93          ]                  118         v                  143       Å

 94          ˆ                  119         w                  144       É

 95          _                  120         x                  145       æ

 96          ‘                  121         y                  146       Æ

 97          a                  122         z                  147       ô

 98          b                  123         {                  148       ö

 99          c                  124         |                  149       ò
                                   ASCII (American Standard Code for Information Interchange)   61

DECIMAL   C H A R AC T E R   DECIMAL       C H A R AC T E R   DECIMAL     C H A R AC T E R           a
150       û                  175           »                  200
151       ù                  176                              201

152       ÿ                  177                              202
153       Ö                  178                              203
154       Ü                  179                              204
155       ¢                  180                              205                                    f
156       £                  181                              206                                    g
157       ¥                  182                              207                                    h
158                          183                              208                                    i
159       ƒ                  184                              209
160       á                  185                              210
161       í                  186                              211
162       ó                  187                              212

163       ú                  188                              213
164       ñ                  189                              214
165       Ñ                  190                              215                                    o
166                          191                              216                                    p
167                          192                              217                                    q
168                          193                              218                                    r
169                          194                              219                                    s
170                          195                              220
171                          196                              221
172                          197                              222

173       ¡                  198                              223
174       «                  199                              224
62        ASCIIbetical Sorting

    DECIMAL       C H A R AC T E R   DECIMAL         C H A R AC T E R   DECIMAL     C H A R AC T E R

    225                              236                                247

    226                              237                                248

    227                              238                                249

    228                              239                                250

    229                              240                                251

    230                              241                                252

    231                              242                                253

    232                              243                                254

    233                              244                                255

    234                              245

    235                              246

w                                                         CASEs provide services for many types of
ASCIIbetical Sorting                                      applications; the SASEs represent or provide
    A sorting strategy that uses the ASCII                services for specific applications or genres.
    character set as the basis for the ordering.
    In ASCII, numbers and special symbols pre-
    cede letters; uppercase letters precede lower-        The following CASEs are commonly used:
    case ones.
                                                             ACSE (Association Control Service
w                                                              Element): This element establishes
ASE (Application Service Element)                              the appropriate relationship between
    In the OSI Reference Model, an ASE (appli-                 two applications (AEs) to enable the
    cation service element) is any of several ele-             applications to cooperate and commu-
    ments that provide the communications and                  nicate on a task. Since all associations
    other services at the application layer. An                or relationships must be established
    application process (AP) or application                    through the ACSE, and since applica-
    entity (AE) requests these services through                tions must establish a relationship to
    predefined interfaces, such as those provided               communicate, the ACSE is needed by
    by APIs (Application Program Interfaces).                  all applications.
       ASEs are grouped into common applica-
                                                             CCRSE (Commitment, Concurrency, and
    tion service elements (CASEs) and specific
                                                               Recovery Service Element): This ele-
    application service elements (SASEs). The
                                                               ment is used to implement distributed
                                                           ASI (Adapter Support Interface)    63

       transactions which may require multi-               do batch data processing on a remote
       ple applications. The CCRSE helps                   machine. With JTM, a node could, for
       ensure that distributed data remains
       consistent by making sure that applica-
                                                           example, start a computation on a
                                                           supercomputer at a remote location
       tions do not interfere with each other              and retrieve the results when the com-    c
       when doing their work and that                      putation was complete.
       actions are performed completely or
                                                        MHS (Message Handling System):
       not at all.
                                                         This element enables applications to
   ROSE (Remote Operations Service Ele-                  exchange messages; for example, when        f
     ment): This element supports interac-               using electronic mail.
     tive cooperation between two appli-                                                             g
                                                        MMS (Manufacturing Message Service):
     cations, such as between a client and
     a server. ROSE provides the services
                                                         This element enables an application on      h
                                                         a control computer to communicate
     needed for the reliable execution of
                                                         with an application on a slave machine
     requested operations and transfer
     of data.
                                                         in a production line or other auto-         j
                                                         mated operation.
   RTSE (Reliable Transfer Service Ele-
                                                        VT (Virtual Terminal): This element
     ment): This element helps ensure that
     PDUs (protocol data units), or packets,
                                                          makes it possible to emulate the behav-    l
                                                          ior of a particular terminal, which
     are transferred reliably between appli-
                                                          enables an application to communicate
     cations. RTSE services can sometimes
     survive an equipment failure because
                                                          with a remote system without consid-       n
                                                          ering the type of hardware sending or
     they use transport-layer services.
                                                          receiving the communications.              o
SASE                                                    The entire set of ASEs required for a par-   p
 The following SASEs are commonly used:              ticular application is known as the applica-
                                                     tion context (AC) for that application.
   DS (Directory Service): This element
                                                     B RO A D E R C A T E G O R Y
     makes it possible to use a global direc-
     tory, which is a distributed database              AC (Application Context)                     s
     with information about all accessible
     network entities in a communications
                                                 ASI (Adapter Support Interface)
                                                     ASI (Adapter Support Interface) is a            u
   FTAM: (File Transfer Access and Man-              standard interface developed by IBM for         v
     agement): This element enables an               enabling Token Ring adapters to talk to any
     application to read, write, or otherwise        of several higher-level protocols. The most     w
     manage files on a remote machine.                recent version of ASI is marketed as LAN
                                                     Support Program.
   JTM (Job Transfer and Manipulation):
     This element enables an application to
64        Asia and Oceania Workshop (AOW)

       Like other adapter interfaces, such                  ASN.1 serves as a common syntax for
    as NDIS (Network Driver Interface Speci-             transferring information between two end
    fication) by Microsoft and ODI (Open                  systems (ESs) that may use different encod-
    Data-Link Interface) by Novell, ASI includes         ing systems at each end.
    at least the following two components:
                                                         P R I M A R Y S O U RC E S
      s   A data-link-layer driver to talk to the           CCITT recommendations X.208 and
          network interface card (NIC)                      X.209; ISO documents 8824 and 8825
      s   A network-layer driver to talk to the          B RO A D E R C A T E G O R Y
          network-level protocols                           Abstract Syntax
w                                                        SEE ALSO
Asia and Oceania Workshop (AOW)                             BER (Basic Encoding Rules)
      AOW (Asia and Oceania Workshop)                Asserted Circuit
w                                                        A circuit that is closed; that is, a circuit with
ASIC (Application-Specific Integrated                     a voltage value. Depending on the logic
Circuit)                                                 being used, an asserted circuit can represent
    Special-purpose chips with logic designed            a 1 (usually) or 0 (rarely).
    specifically for a particular application or      w
    device. ASICs are also known as gate arrays,     Assigned Number
    and they are constructed from standard cir-
                                                         In the Internet community, a numerical
    cuit cells from a library.
                                                         value that serves to distinguish a particular
w                                                        protocol, application, or organization in
ASN.1 (Abstract Syntax Notation                          some context. For example, assigned num-
One)                                                     bers distinguish the different flavors of
    In the OSI Reference Model, ASN.1                    Ethernet protocols used by different imple-
    (Abstract Syntax Notation One) is a nota-            menters. Assigned numbers, which are not
    tion used to describe data structures, such as       addresses, are assigned by the Internet
    managed objects in a network management              Assigned Numbers Authority (IANA).
    system.                                          w
       ASN.1 is machine-independent and is           ASVD (Analog Simultaneous
    used in many networking contexts. For            Voice/Data)
    example, it is used to describe application-
                                                         A proposed modem standard that can be
    layer packets in both the OSI network
                                                         used to transmit multimedia materials—
    management framework and in the Simple
                                                         voice, video, etc.—over ordinary (analog)
    Network Management Protocol (SNMP)
                                                         telephone lines. The ASVD specifications are
    from the Internet TCP/IP protocol suite.
                                                         being finalized by the ITU (International
                                                                               AT Command Set             65

    Telecommunication Union, formerly known
    as the CCITT).
                                                          A D A T A WO RD S E N T BY                           a
                                                          A S Y N C H RO N O U S T RA N S M I S S I O N
       ASVD is offered as an inexpensive (and
    slower) alternative to ISDN (Integrated Ser-
    vices Digital Network). The bandwidth for                                                                  c
    ASVD is considerably more limited than for
    ISDN. The version under consideration sup-
    ports modem speeds of up to 14.4 kbps, but                                                                 e
    somewhat slower speeds for multimedia
    data.                                                                                                      f
    Asynchronous describes a communications
    strategy that uses start and stop bits to indi-                                                            i
    cate the beginning and end of a character,
    rather than using constant timing to trans-
                                                      ATA (ARCnet Trade Association)                           j
    mit a series of characters. In a sense, asyn-         A consortium of vendors and other organi-            k
    chronous transmissions actually synchronize           zations that manages ARCnet specifications.
    for each character. The figure “A data word        w
    sent by asynchronous transmission” shows          AT Command Set                                           m
    the bits used in this communications
                                                          The AT command set was developed by
                                                          Hayes Microcomputer Products to operate
       Asynchronous communications methods
    are generally less efficient but more resistant
                                                          its modems. The AT in the title is an abbre-         o
                                                          viation for attention. This signal precedes
    to disruption than synchronous communica-
                                                          most of the commands used to get a modem
    tions. Asynchronous methods are more effi-
    cient for situations in which traffic comes in
                                                          to do its work. For example, ATDP and                q
                                                          ATDT (for attention dial pulse and attention
    bursts (rather than moving at a regular
    pace). Common examples of asynchronous
                                                          dial tone, respectively) are used to dial a          r
    communications devices are modems and
                                                          number on either a pulse or Touch Tone
                                                              The AT command set quickly became a              t
w                                                         de facto standard. It is now used by most
Asynchronous Modem Eliminator                             modem manufacturers and is supported on              u
(AME)                                                     virtually every modem on the market.                 v
      AME (Asynchronous Modem Eliminator)
                                                          SEE ALSO
66        ATCON

w                                                       needed. The initial implementations will
ATCON                                                   operate at 155.52 megabits per second
    A Novell NetWare program that monitors              (Mbps), then at 622.08 Mbps. Speeds up
    the AppleTalk protocol stack in a multipro-         to 2.488 gigabits per second (Gbps) are
    tocol network. It reports statistics about          planned and have been demonstrated in
    the performance of AppleTalk devices                limited tests.
    and services.                                           The very high bandwidth and the ability
                                                        to transmit multiple media make ATM an
w                                                       attractive, high-speed architecture for both
ATDP (Attention Dial Pulse)                             local-area networks (LANs) and wide-area
    In the Hayes modem command set, a com-              networks (WANs). It is useful for enterprise
    mand to dial a number using a pulse (rotary)        networks, which often connect LANs over
    telephone.                                          wide areas and may need to transport large
                                                        amounts of data over very long distances.
                                                            Long-haul, high-bandwidth capabilities
      AT Command Set
                                                        are particularly attractive for WANs, which
w                                                       have until now been shackled by the rela-
ATDT (Attention Dial Tone)                              tively low bandwidths over long-distance
    In the Hayes modem command set, a com-              lines. FDDI (Fiber Distributed Data Inter-
    mand to dial a number using a Touch Tone            face) is a good architecture for LANs, and
    phone.                                              frame relay has possibilities for WANs, but
                                                        neither of these architectures is suitable for
    SEE ALSO                                            both LANs and WANs. But note that ATM
      AT Command Set                                    is still quite expensive.
                                                            The figure “Context and properties of
ATM (Asynchronous Transfer Mode)                        ATM” summarizes the characteristics of this
    ATM (Asynchronous Transfer Mode) is a
    packet-switched, broadband network archi-         ATM Features
    tecture that is expected to become an estab-
    lished standard by the late 1990s. It forms         ATM has the following features:
    the core of a broadband ISDN (BISDN)                  s   Transmission over fiber-optic lines.
    architecture, which extends the digital trans-            These can be local or long-distance,
    mission capabilities defined by ISDN to                    public or private lines. Long-distance
    allow data, voice, and multimedia transmis-               lines can be leased or dial-up.
    sions on the same lines. It is also known as
    cell relay, to distinguish it from frame relay.       s   Capability for parallel transmissions,
       ATM is a real-time architecture that will              because ATM is a switching architec-
    be able to provide very high bandwidths as                ture. In fact, each node can have a ded-
                                                              icated connection to any other node.
                                                                    ATM (Asynchronous Transfer Mode)             67

C O N T EX T A N D P RO P E R T I E S O F AT M                                                                        a
Context                                                                                                               b
Network Architecture                                                                                                  c
  Switched-Media                                                                                                      d
              ATM (Cell Relay)                                                                                        g
      Frame Relay                                                                                                     h
                                                          ATM                                                         j
  Properties                             Structure
                                         Layers                     Planes           Cells
  Broadband                              Physical
                                         (Two Sublayers)
                                                                                     Constant Size (53 Octets)
                                                                                    (48-Octet Payload)
  Core of BISDN
  Useful for LANs and WANs               ATM Layer                  Control          (5-Octet Header)                 m
  Uses short- or long-haul fiber-optic    (Service Independent)                      Not Byte-Bound/Oriented
   cable                                 AAL                                                                          n
  Initial speeds up to 166.62 Mbps        (Two Sublayers)
   (eventural speeds up to 2.49           (Four Service Classes)                                                      o
                                          A: for Voice, Data
  Can always operate at top speed         B: for Video, etc.                                                          p
   (provided there is enough traffic)
  Can transmit voice, video, data
                                          C: for Connection-
                                            Oriented Mode                                                             q
   (simultaneously, if necessary)         D: for Connectionless
                                            Mode                                                                      r
    s   Operation at maximum speed at all                              s   Error correction and routing in hard-      t
        times, provided there is enough net-                               ware, partly because of the fixed cell      u
        work traffic to give the required                                   sizes.
                                                                       s   Transmission of voice, video, and data
    s   Use of fixed-length (53-byte) packets,                              at the same time. The fixed-length cells    w
        which are known as cells.                                          also make voice transmission more
 68       ATM (Asynchronous Transfer Mode)

          accurate, because there is less timing    interfaces for ATM networks, including the
          variation.                                following:
      s   Easier load balancing, because the           s   SONET connections at 155.52 Mbps
          switching capabilities make it possible          (OC-3, STS-3, or in CCITT terminol-
          to have multiple virtual circuits                ogy, STM-1)
          between sender and receiver.
                                                       s   DS3 connections at 44.736 Mbps
ATM Structure                                          s   100 Mbps connections using 4B/5B
   The ATM architecture is organized into lay-             encoding
   ers, as are other network architectures, and
                                                    S T R U C T U RE O F T H E
   also into planes, which specify domains of
                                                    AT M A RC H I T E C T U RE
   activity. See the figure “Structure of the
   ATM architecture” for a graphic representa-
   tion of the organization of the planes and

Physical Layer

   The ATM physical layer corresponds to the
   OSI Reference Model physical layer. It is
   concerned with the physical medium and
   interfaces, and with the framing protocols
   (if any) for the network.
       The physical layer has two sublayers.
   The lower sublayer, physical medium (PM),
   includes the definition for the medium (opti-
   cal fiber) and the bit-timing capabilities. The
   upper sublayer, transmission convergence
   (TC), is responsible for making sure valid
   cells are being created and transmitted. This
   involves breaking off individual cells from
   the data stream of the higher layer (the ATM
   layer), checking the cell’s header, and encod-
   ing the bit values.
       The user network interface (UNI) speci-
   fied by the ATM forum, an organization
   dedicated to defining and implementing
   ATM, allows for various types of physical
                                                      ATM (Asynchronous Transfer Mode)          69

     s   155 Mbps connections using 8B/10B                  during teleconferences. The protocol is
         encoding                                           AAL 2.
     These interfaces all use optical fiber,             s   Class C is suited for connection-          b
  which is the only medium specified for
  ATM. A work group is investigating the pos-
                                                            oriented data transmissions. The
                                                            protocol is AAL 3 or AAL 5.
  sibility of defining ATM for Category 3
                                                        s   Class D is suited for connectionless
  unshielded twisted-pair (UTP) wire.
                                                            data transmissions. The protocol is        e
                                                            AAL 4 or AAL 5.
ATM Layer                                                                                              f
                                                        AAL 5 supports classes C or D more
  The ATM layer is a service-independent
                                                     efficiently than AAL 3 or AAL 4.                   g
  layer at which cell headers and trailers are
  created, virtual channels and paths are          AAL Sublayers
  defined and given unique identifiers, and
  cells are multiplexed or demultiplexed. The        The AAL has two sublayers:
  ATM layer creates the cells and uses the
                                                        s   CS (convergence sublayer) is the upper     j
  physical layer to transmit them.
                                                            sublayer that provides the interface for   k
                                                            the various services. Users connect to
AAL (ATM Adaptation Layer)
                                                            the CS through service access points       l
                                                            (SAPs). No protocol data units (PDUs)
  The topmost layer, AAL is service-
  dependent. It provides the necessary                      are defined for this level because the
  protocol translation between ATM and                      data passing through is application-       n
  the other communication services (such                    and service-dependent.
  as voice, video, or data) involved in a trans-
                                                        s   SAR (segmentation and reassembly) is
  mission. For example, the AAL translates
                                                            the sublayer that packages variable-       p
  between elements from a pulse-code modu-
  lation (PCM) transmission (which encodes
                                                            size packets into fixed-size cells at the
                                                            transmitting end, and repackages the
  voice data in digital form) and ATM cells.
     The following four types of services,
                                                            cells at the receiving end. The SAR        r
                                                            sublayer is also responsible for finding
  which each use different AAL protocols,
                                                            and dealing with cells that are out of     s
  are defined at the AAL:
                                                            order or lost.                             t
     s   Class A is suited for constant bit
         rate (CBR) data and provides circuit-
                                                        A separate PDU is defined for each
                                                     class of service. Each PDU contains 48
         switching emulation. This is appropri-
         ate for voice data. The protocol is
                                                     octets, which are allocated for the header,       v
                                                     trailer, and data (known as payload in ATM
         AAL 1.
                                                     terminology). Of these, the AAL 1 PDU can
     s   Class B is for variable bit rate (VBR)      carry the most data at a time: a 47-octet         x
         data; for example, video transmissions      payload. AAL 3 and AAL 4 each have a
 70       ATM (Asynchronous Transfer Mode)

  44-octet payload, and AAL 2 has a 45-octet        handle voice data; Class C or D services will
  payload. These PDUs become the data (pay-         handle data from a network, and so forth.
  load) for the ATM cells that are transmitted.        Data comes into the AAL as packets of
                                                    varying sizes, but leaves as fixed-size (48-
ATM Planes                                          octet) SAR PDUs. The details of these PDUs
                                                    depend on the type of service (Class A, B, C,
  Three domains of activity, known as planes,
                                                    or D) being used. The SAR sublayer does the
  are distinguished for ATM:
                                                    necessary chopping and packing.
      s   The control plane, on which calls            The SAR PDUs from the various services
          and connections are established and       are wrapped into ATM cells at the ATM
          maintained.                               layer and multiplexed for transmission onto
                                                    the ATM cell stream. These ATM cells
      s   The user plane, on which users, or
                                                    contain the virtual channel and path iden-
          nodes, exchange data. This is the plane
                                                    tification required for the cell to reach its
          at which ordinary user services are
                                                    destination. The ATM switch uses channel
                                                    and path information to send the cell out
      s   The management plane, on which            through the appropriate port.
          network-management and layer-                The cell stream contains bits and pieces of
          management services are provided.         various types of packets, all in separate cells.
          This plane coordinates the three planes   The cells may be routed, or switched, at var-
          and manages resources for the layers.     ious points on their path, as appropriate for
                                                    maintaining connections at the required
ATM Operation                                       quality of service.
  The figure “ATM transmission elements”                The cell stream is encoded and trans-
  shows the elements used as a transmission         mitted over the physical media connecting
  gets onto an ATM network. The top part of         the ATM network. At the receiving end, the
  the illustration represents the higher (non-      ATM routes the cells to the appropriate ser-
  ATM) service layers; the bottom part repre-       vices at the AAL. The cells are repackaged
  sents the ATM and physical layers in the          into the appropriate packet form by the
  ATM model. The ATM node does the work             AAL service. This service also checks that
  of the AAL and much of the ATM layer.             the entire packet has been received and that
      Data from the various types of services       everything is correct.
  (voice, video, data, and so forth) is handled        At the receiving end, the transmission
  at the AAL layer in an ATM node. The data         sequence is undone, with the services at the
  is converted into ATM cells, regardless of        topmost (for ATM) sublayer unpacking the
  the types of packets that came in. The data       ATM cells to reveal the various types of
  is handled by the appropriate class of ser-       data, which are passed out to the services
  vice. For example, the Class A services will      that handle the data.
                                                            ATM (Asynchronous Transfer Mode)                        71

AT M T RA N S M I S S I O N E L E M E N T S                                                                              a
     Data Stream            Voice Stream              Video Stream                                 ATM Cells             b
   (from network,          (from classic          (from teleconferences,
       router)          telephone service)         image transmissions)                       (from ATM stations)        c
    Packet-Mode             Circuit-Mode                   Video-Mode                                ATM
      Services               Emulation                       Services                              Services
      ATM Cells                ATM Cells                    ATM Cells                              ATM Cells
                                               ATM Node
                                               ATM Switch
                                              (Multiplexer)                                                              m
                                                                        Stream of ATM Cells

                                               Video Cell
                                               Data Cell
                                               Voice Cell
                                                 . . . .                                                                 t
                                               Data Cell                                                                 v
72       ATM (Asynchronous Transfer Mode)

ATM Interfaces                                     S T R U C T U RE O F A N
                                                   AT M C E L L A T T H E U N I
  The ATM architecture distinguishes between
  two interfaces:
     s   The user-network interface (UNI),
         which connects an end-user to the
         network via an ATM switch or other
         device. This interface supplies network
     s   The network-node interface (NNI),
         which connects network nodes to each
         other. This interface makes network
         routing possible.

Cell Structure

  ATM cells are not byte oriented. Even
  though cells are defined as a specific number
  of octets, the fields within such a cell often
  cross byte boundaries.
     ATM cells consist of a five-octet header
  and a 48-octet data, or payload, section. The
  payload section is an SAR PDU, to which a
  five-octet ATM header is added. See the fig-
  ure “Structure of an ATM cell at the UNI.”
                                                         1 being the least significant bit within
     Most of the bits in the header are used for
                                                         that octet.
  virtual path and channel identification. The
  CLP (cell loss priority) bit indicates whether     s   Across octets, bit order goes down-
  the cell can be discarded if network traffic            ward as octets go upward. Thus, the
  volume makes this advisable. If the flag is             lowest order bit in the VPI field is bit 5
  set, the cell is expendable.                           in octet 2. Similarly, the lowest order
     Because header fields can extend over                bit for the VCI field is bit 5 in octet 4;
  multiple octets—for example, the VPI or                the highest order bit in this field is bit
  VCI fields—the ATM specifications include                4 in octet 2, and the bits in octet 3
  the following guidelines for how bits are to           are between the high- and low-order
  be arranged within a field:                             quartets.
     s   Within an octet, bit order goes from         The cell-structure shown in the
         left to right. For example, in octet      figure “Structure of an ATM cell at the
         1, the VPI bits are—from highest to       UNI” applies to cells that travel onto
         lowest—bits 4, 3, 2, and 1, with          the network across the UNI. When cells
                                                   ATM (Asynchronous Transfer Mode)         73

  are moving across the NNI—that is, for           Token Ring adapter. This software may be
  routing purposes—the VPI field is extended        included as a driver on the workstation, or
  to encompass the entire first octet. That is,
  cells at the NNI use 12 bits for VPI and 16
                                                   client machine. Additional software runs a
                                                   LAN emulation server—either on an ATM
  for VCI. There is no generic flow control         switch or on a separate PC.                      c
  field for these cells.                               With ATM LAN emulation, an ATM
                                                   device can be made to look like an Ethernet
ATM Variants                                       or a Token Ring node to a network server.        e
  Because ATM’s progress toward becoming           Below the surface, however, the virtual
  the dominant high-speed architecture has         Ethernet device, for example, is able to oper-   f
  been much slower than anticipated, several       ate at blazing ATM speeds by breaking the        g
  variants on the basic technology have been       Ethernet packets into ATM cells before
  proposed as a means of getting at least some     sending them on. The packets might be sent       h
                                                   across an ATM network to a receiving
  form of ATM into more markets and net-
                                                   device that also supports LAN emulation.
  works. Two of the more interesting variants
  are ATM25 and ATM LAN emulation.                 The packets could then be reassembled at         j
                                                   the receiving end and passed transparently
ATM25                                              to a receiving Ethernet device. Information      k
                                                   in the header area identifies packets as com-
  ATM25 is a 25 Mbps version proposed for          ing from a LAN emulation device. Such
  use in desktop networks—that is, in LANs.
  This version was proposed by the Desktop
                                                   an emulation makes ATM devices indepen-          m
                                                   dent of higher-level protocols (for example,
  ATM25 Alliance, which includes IBM and           TCP/IP or IPX).                                  n
  Apple among its members. This variant
  would run on ordinary UTP (unshielded          ATM Resources
  twisted pair) cables, and would allow 25
                                                   The ATM Forum is a consortium of several
  Mbps transmissions in both directions. The
  ATM Forum is considering the ATM25               hundred vendors, researchers, and other          q
  specifications, and Alliance members are          involved parties. The Forum’s charter is
  currently working on specifications that          to help develop and promote the use of           r
  would enable products from different ven-        ATM-related products and services. Toward
                                                   this end, the forum provides information
  dors to work together, and that would
  enable ATM25 networks to communicate             about ATM, helps develop specifications           t
                                                   for ATM products and use, and generally
  in a transparent manner with other, faster-
  speed ATM networks.                              keeps ATM on the minds of the appropriate        u
                                                   people and groups.                               v
ATM LAN Emulation                                     Forum members are companies that are
                                                   interested in developing or using ATM tech-      w
  This variant uses software to fool a network     nology. These companies are readying prod-
  operating system into thinking that an ATM       ucts for various facets of an ATM network,
  interface card is actually an Ethernet or        such as nodes, switches, PBXs, and routers.      y
74         ATPS (AppleTalk Print Services)

       Various combinations of forum members/                Once attached, the user at the worksta-
    vendors have formed partnerships to create            tion can access any of the server’s services
    and market ATM components. Companies                  (assuming that the user has the necessary
    such as Sprint and AT&T will offer ATM                access rights to those services). The
    services to their customers.                          ATTACH command cannot be used to
       While many aspects of the ATM technol-             connect to the network initially. The LOGIN
    ogy and specifications are still in flux, signif-       command must be used for the first server.
    icant portions have been tested and proven            Then the ATTACH command can be used to
    viable. Vendors have forged ahead and are             attach to additional servers. ATTACH does
    selling ATM products. They are still quite            not execute a login script or redefine the
    expensive, however, partly because the                workstation’s environment. The ATTACH
    absence of finalized specifications has led to          command is not included in NetWare 4.x.
    vendor-specific implementations. This, of
                                                          B RO A D E R C A T E G O R Y
    course, makes interoperability more elusive
    and customers more reluctant.
       ATM variants and emulation schemes             w
    have been proposed in an effort to make           Attachment
    ATM better known. Major ATM vendors                   In electronic mail, an attachment is a file
    have been cutting their prices, which is also         that is sent along with a regular e-mail
    expected to help the established base grow.           message.
    B RO A D E R C A T E G O R I E S
       Network Architecture; Network, Cell-           Attack Scanner
       Switched; Network, Packet-Switched
                                                          An attack scanner is a software package
w                                                         used to probe UNIX networks for security
ATPS (AppleTalk Print Services)                           problems or flaws. The package will essen-
    An NLM (NetWare Loadable Module) that                 tially play the role of an intruder trying to
    provides NetWare nodes with access to                 steal or force access to a network. The use of
    printers and Macintosh nodes with access to           such programs is somewhat controversial.
    NetWare print queues. Settings for this mod-             In April, 1995, a controversial attack
    ule are in the ATPS.CFG file.                          scanner product—SATAN (Security Analysis
                                                          Tool for Auditing Networks) by Wietse
w                                                         Venema and Dan Farmer—was posted to
ATTACH                                                    the Internet. Such a product can be used by
    In Novell’s NetWare 2.x and 3.x, the                  crackers (users trying to break into systems
    ATTACH command tells a file server that a              for malicious purposes) as well as by system
    workstation exists and wants to join the net-         administrators and security people. As a
    work. The server will assign the workstation          result, the Internet community is divided as
    a connection number.                                  to whether such a product should be made
                                                          freely available.
                                                                                     Attribute       75

w                                                         decibel loss per kilometer (expressed as
Attenuation                                               dB/km).
    Attenuation is the loss of signal strength
                                                      w                                                    b
    over distance. It is measured in decibels (dB)    Attribute
    per kilometer (expressed as dB/km) or per                                                              c
                                                          An attribute is a feature or property associ-
    100 feet. In the logarithmic decibel scale, a 3
    dB loss means a 50 percent loss in power, as
                                                          ated with an entity. For example, objects in     d
    computed in the following equation. Specifi-
                                                          network management and entries in an
                                                          X.500 Directory Services database have
    cally, the formula for power loss is:
                                                          attributes.                                      f
                      Power out                              An attribute has a type and a value asso-
       dB = 10 log 10 ----------------------
                       Power in
                                                          ciated with it. The type constrains the form
       In this equation, a 50 percent loss would
                                                          the value can take. For example, an INTE-        h
                                                          GER type may have only a whole number
    actually yield a result of −3 dB. Under cer-                                                           i
                                                          value, or a BOOLEAN may have only a
    tain conditions, the coefficient in the equa-
                                                          value that evaluates to TRUE or FALSE.
    tion will be 20, in which case a result of −6
                                                             Much network management or monitor-
    dB would indicate a 50 percent loss. When
    describing losses, however, the negative sign
                                                          ing activity consists of determining or          k
                                                          changing attribute values. Attribute values
    is dropped, so that a result of −6 dB is
                                                          are read or set by functions that provide the    l
    expressed as a 6 dB loss.
       Attenuation depends on several factors,
                                                          relevant network services.                       m
    including the wire composition and size,
    shielding, and frequency range of the signal.
                                                      File and Directory Attributes                        n
    For copper cable, attenuation increases with          Among the most important attributes are          o
    signal frequency; for optical fiber, attenua-          those associated with files and directories,
    tion is relatively constant over a large fre-         because these ultimately limit what can be       p
    quency range.                                         done on a network. The attributes are gener-
                                                          ally represented as single-bit flag values,
       Fiber-optic cable has the least attenua-
    tion, usually fractions of a decibel per kilo-        with the flag either set or not set.              r
    meter. Unshielded untwisted-pair cable (such             The specific attributes defined vary from
    as the silver, flat-satin cables used in short-        system to system, but attributes are used        s
    distance telephone and modem lines) has the           in every operating system and networking
                                                          environment. Certain attributes assume or
    most attentuation of any cable types used in
    telecommunications. This type of cable is             replace others, and certain attributes over-     u
                                                          ride access rights. See the table “Novell Net-
    not used directly in networks.
                                                          Ware File and Directory Attributes” for
Attenuation Factor
                                                          descriptions of NetWare attributes associ-       w
                                                          ated with files and directories.
    A value that expresses the amount of a                                                                 x
    signal lost over a given distance, such as
76      Attribute


 A (Archive needed)    Set automatically when a file is changed after its most recent backup. (NetWare 2. x,
                       3.x, 4.x)

 C (Copy inhibit)      Set to keep Macintosh files from being copied. Does not apply to DOS files. (NetWare
                       3.x, 4.x)

 Cc (Can’t compress)   Set automatically when a file cannot be compressed because it would not save a signif-
                       icant amount of space. (NetWare 4. x)

 Co (Compressed)       Set automatically to show that a file has been compressed. (NetWare 4. x)

 Di (Delete inhibit)   Set to keep users from deleting a file or directory. (NetWare 3.x, 4.x)

 Dc (Don’t compress)   Set to prevent a file from being compressed. (NetWare 4.x)

 Dm (Don’t migrate)    Set to prevent a file from being migrated to a secondary storage medium, such as an
                       optical disk drive. (NetWare 4. x)

 X (Execute only)      Set to keep a file from being copied, deleted, changed, or backed up. Since this setting
                       cannot be changed, it’s necessary to keep a backup (nonrestricted) copy of the pro-
                       gram before freezing it. Assigning this attribute is not recommended; the same effect
                       can be accomplished with the Ro attribute. (NetWare 2.x, 3.x, 4.x)

 H (Hidden)            Set to keep a file or directory from being displayed in a directory listing. (NetWare 2.x,
                       3.x, 4.x)

 I (Indexed)           Set to make it faster to access a file with many clusters on a hard disk. (NetWare 2.x,
                       3.x, 4.x)

 Ic (Immediate         Set to make sure that a file is compressed immediately. (NetWare 2.x, 3.x, 4.x)

 M (Migrate)           Automatically set to show that a file has been migrated to secondary storage medium.
                       (NetWare 4.x)

 P (Purge)             Set to make sure a file or directory is purged (zeroed) immediately after deletion, so
                       that no data from the file is available. (NetWare 3.x, 4.x)

 R (Rename inhibit)    Set to make sure a file or directory name is not changed. (NetWare 3.x, 4.x)

 Ra (Read audit)       Supported but not used.

 Ro/Rw (Read only/     Set to specify whether a file can be modified. (NetWare 2.x, 3.x, 4.x)
  Read write)

 S (Shareable)         Set to indicate that multiple users or processes can access a file simultaneously.
                       (NetWare 2.x, 3.x, 4.x)
                                                                      AUP (Acceptable Use Policy)              77

    AT T R I B U T E      DESCRIPTION                                                                               a
    Sy (System)           Set to indicate that a file or directory is a NetWare or DOS system file or directory.
                          (NetWare 2.x, 3.x, 4.x)                                                                   b
    T (Transactional)     Set to allow NetWare’s Transactional Tracking System (TTS) to protect a file. (Net-        c
                          Ware 2.x, 3.x, 4.x)
    Wa (Write audit)      Supported but no AU (Access Unit)
    SEE ALSO                                             w                                                          f
       Access Rights                                                                                                g
                                                             An examination of network activity to
AU (Access Unit)                                             make sure that the network monitoring                  h
                                                             and data gathering are working correctly.
    In the 1988 version of the CCITT’s X.400
                                                             Although this is a management activity, it
    Message Handling System (MHS), an AU
    is an application process that provides a
                                                             is done independently of the network man-              j
                                                             agement package in some environments (for
    CCITT-supported service, such as faxing,
                                                             example, in NetWare). An independent                   k
    with access to a Message Transfer System
    (MTS). The MTS can deliver a message to
                                                             audit can check the reliability of the man-
                                                             agement software.
    users or services at any location accessible
    through the MHS.                                     w
        AUs supplement user agents (UAs), which           AUI (Attachment Unit Interface)                           n
    give human users access to an MTS.                       One component of the physical layer, as
                                                             defined in the IEEE 802.x specifications and
    B RO A D E R C A T E G O R Y
       X.400                                                 in the OSI Reference Model. The other two              p
                                                             components are the physical layer signaling
    C O M P A RE                                             (PLS) above the AUI and the physical
       PDAU; UA (User Agent)                                 medium attachment (PMA) below it.                      r
Audio Frequency Range
                                                             SEE ALSO
                                                               Connector, AUI (Attachment Unit
    The range of frequencies that the human ear                Interface                                            t
    can hear, which goes from a frequency of 20
    hertz to about 20 kilohertz (although few
                                                         w                                                          u
                                                          AUP (Acceptable Use Policy)
    people can hear the extremes well). People
                                                             An AUP represents guidelines established for
    can produce sounds within only a small por-
    tion of this range, from about 100 to 3,000              the use of the Internet or of the services from        w
                                                             a particular provider. For example, in the
    hertz, which is the bandwidth of the ordi-
                                                             early days, commercial traffic was not                  x
    nary, acoustically-based telephone system.
                                                             allowed on the Internet, according to the              y
78         Authentication

    NSF’s (National Science Foundation) AUP.           w
    Internet service providers may also stipulate      Authentication System
    AUPs. For example, providers may restrict              An authentication system is a server whose
    or prohibit distribution of newsletters or             job is to check the validity of all identities on
    other postings to large subscriber lists.              the network and of their requests. Most of
                                                           the work is done automatically, without
Authentication                                             requiring any explicit human intervention.
                                                              One example of an authentication system
    In network security and other operations,
                                                           is Kerberos, which was created for Project
    authentication is the process of determining
                                                           Athena at MIT. Kerberos is a distributed
    the identity and legitimacy of a user, node,
                                                           authentication system which verifies that a
    or process. Various authentication strategies
                                                           user is legitimate when the user logs in and
    have been developed. Among the simplest
                                                           every time the user requests a service. Ker-
    are the use of user IDs and passwords.
                                                           beros uses special keys, called tickets, to
       A relatively new authentication scheme,
                                                           encrypt transmissions between Kerberos
    called digital signatures, is very effective and
                                                           and a user.
    almost impossible to fool (unless one has
    access to the private encryption key of one            B RO A D E R C A T E G O R Y
    party). In digital signatures, a user (user A)            Network Security
    uses another user’s (user B’s) public key to
    encrypt the transmission, and uses A’s pri-
                                                       Authority and Format Identifier (AFI)
    vate key to “sign” it. At the receiving end,
    user B uses A’s public key to validate the sig-        SEE
    nature, and user B’s private key to decrypt               AFI (Authority and Format Identifier)
    the transmission.
       The CCITT distinguishes two levels of           Autocall Unit (ACU)
    authentication for directory access in its
    X.509 recommendations:                                 SEE
                                                              ACU (Autocall Unit)
       s   Simple authentication, which uses just
           a password and works only for limited       w
           directory domains.                          AUTOEXEC.BAT
       s   Strong authentication, which uses               Under DOS, AUTOEXEC.BAT is a special
           a public key encryption method                  batch file that is executed automatically
           to ensure the security of a                     when the computer boots or reboots. The
           communication.                                  commands in the file can be used to config-
                                                           ure a working environment. For example,
    B RO A D E R C A T E G O R Y                           commands in an AUTOEXEC.BAT file may
       Network Security                                    load drivers or other files, set a command
                                                           line prompt, set environment variables, load
                                                           a network operating system, and so on.
                                                                      Auto-Partition Algorithm        79

       Various solutions have been developed          w
    to allow some flexibility in booting to an         Automatic Alternate Routing (AAR)                    a
    environment. For example, OS/2 version
    2.x allows each DOS process to have its
                                                          SEE                                              b
                                                            AAR (Automatic Alternate Routing)
    own automatically executed file. For DOS,                                                               c
    various programs have been developed              w
    to allow conditional processing in the            Automatic Call Distributor                           d
    AUTOEXEC.BAT file.                                     A device that automatically switches an          e
                                                          incoming call to the next available line.
    B RO A D E R C A T E G O R Y                                                                           f
       Boot                                           w
                                                      Automatic Number Identification                       g
       AUTOEXEC.NCF; CONFIG.SYS                                                                            h
                                                            ANI (Automatic Number Identification)
    On a NetWare server, AUTOEXEC.NCF is              w
    an executable batch file that is used to con-
                                                      Automatic Repeat Request (ARQ)                       k
    figure the NetWare operating system and to
    load the required modules. The following
                                                            ARQ (Automatic Repeat Request)
    are some of the tasks of AUTOEXEC.NCF:
       s   Store the server name and IPX internal     Automatic Rollback                                   n
           network number.                                In NetWare’s Transaction Tracking System
       s   Load local-area network (LAN) driv-            (TTS), a feature that restores the starting
           ers and the settings for the network           state of a database if a transaction fails       p
           interface cards (NICs).                        before completion.
       s   Bind protocols to the installed drivers.   w
                                                      Automatic Route Selection (ARS)                      r
       s   Load NetWare Loadable Modules
           (NLMs).                                        SEE                                              s
                                                            ARS (Automatic Route Selection)
       s   Set time-zone information on the                                                                t
           network.                                   w
                                                      Auto-Partition Algorithm                             u
       s   Execute certain server commands.
                                                          An algorithm by which a repeater can auto-       v
    C O M P A RE                                          matically disconnect a segment from a net-
       AUTOEXEC.BAT                                       work if that segment is not functioning          w
                                                          properly. This can happen, for example,
                                                          when a broken or unterminated cable causes
80        A/UX

    too many collisions. When the collisions           based on the gauge, or diameter, of the con-
    have subsided, the network segment can be          ducting wire. The lower the gauge, the
    reconnected.                                       thicker the wire and the lower the resistance
                                                       per unit length. The table “Diameter and
                                                       Resistance Values for Selected Wire Gauges”
                                                       shows some gauge values and corresponding
    An implementation of the UNIX operating            diameters.
    system on a Macintosh, enhanced with some
    Macintosh-specific features, such as support
    for the Macintosh Toolbox. A/UX is based           DIAMETER AND RESISTANCE
    on System V Release 2 (SVR2) of AT&T’s             VA L U E S F O R S E L E C T E D WI RE
    UNIX.                                              GAUGES
                                                       AW G                         R E S I S TA N C E
w                                                      VA L U E     DIAMETER        (OHMS/
AUX                                                    (GAUGE)      (MM)            METER)

    In DOS, AUX is the logical name for an             30           0.26            0.346
    auxiliary device. This is usually the serial
                                                       24           0.51            0.080
    communications board, which is more
    commonly known as COM1.                            22           0.64            0.050

w                                                      20           0.81            0.032
                                                       18           1.02            0.020
    In network performance management, the
                                                       16           1.29            0.012
    proportion of time during which a particular
    device, program, or circuit is ready for use.      14           1.63            0.008
    Specifically, the availability of a device is the
                                                       12           2.05            0.005
    ratio of MTBF to (MTBF + MTTR), where
    MTBF and MTTR are mean time before
    failure and mean time to repair, respectively.
    A device is considered available even if it is
    in use.

Avalanche Photodiode (APD)
       APD (Avalanche Photodiode)

AWG (American Wire Gauge)
    AWG (American Wire Gauge) is a classifica-
    tion system for copper wire. The system is
    82      B8ZS (Bipolar with 8 Zero Substitution)

                                                          network can talk to nodes in other networks

    B8ZS (Bipolar with 8 Zero                             by sending packets across the backbone
    Substitution)                                         network.
     A signal-encoding scheme in which a 1 is                The networks attaching to the backbone
     represented alternately as positive and nega-        are known as access networks. Access net-
     tive voltage, and 0 is represented as zero           works may require a gateway or router to
     voltage. B8ZS requires at least one bit of           attach to the backbone network.
     every eight to be a 1.                                  A backbone network can be useful in
                                                          decentralized corporations. For example, a
     SEE ALSO                                             backbone network might be used in a com-
         Encoding, Signal                                 pany in which each department has set up its
                                                          own network and several different architec-
    BAC (Basic Access Control)                            tures are used. Since the backbone network
                                                          leaves the access networks intact, those net-
     In the CCITT X.500 directory services
                                                          works can continue operating as if they were
     model, the more comprehensive of two sets
                                                          not on the larger network. However, the
     of access-control guidelines. The less com-
                                                          backbone gives each of the networks access
     prehensive set is called SAC (Simplified
                                                          to the resources and data of the other access
     Access Control).
     SEE ALSO                                                One obstacle to a successful backbone
         X.500                                            network is the high bandwidth that may be
                                                          required to handle potentially heavy traffic.
w                                                         Because of this consideration, fiber-optic
                                                          cable is the most sensible cabling for back-
     In a hierarchically arranged distributed sys-        bone networks.
     tem, the backbone is the top-level, or cen-
     tral, connection path shared by the nodes or
                                                      Back End
     networks connected to it.
        The backbone manages the bulk of the              In a client/server architecture, the portion of
     traffic, and it may connect several different         an application that runs on the server and
     locations, buildings, and even smaller net-          does the actual work for the application.
     works. The backbone often uses a higher-             The front end runs on the client machine
     speed protocol than the individual local-area        and provides an interface through which the
     network (LAN) segments.                              user can send commands to the back end.

w                                                     w
    Backbone Network                                  Background Process
     A backbone network is one with a central             A process or program that executes inciden-
     cabling scheme (the backbone) to which               tally, while another process or program is
     other networks are attached. Nodes in one            operating in the foreground. The foreground
                                                                                      Backup      83

    process gets the main attention of the CPU       w
    (central processing unit), and the back-         Backscattering                                      a
    ground process takes CPU cycles when the
    foreground process is temporarily idle.
                                                         In a fiber-optic transmission, light that is     b
                                                         reflected back in the direction from which
                                                         the light came.                                 c
Backing Out                                          w                                                   d
    In NetWare’s TTS (Transaction Tracking
    System), the process of abandoning an
                                                         A backup is an archival copy that is stored
    uncompleted database transaction, leaving            on an external medium. For example, a           f
    the database unchanged. TTS takes this
    action to ensure that the database is not cor-
                                                         backup might contain the contents of a hard
                                                         disk or a directory.
    rupted by information from an incomplete                The creation of regular backups is essen-    h
    transaction.                                         tial in a networking environment. An effec-
                                                         tive backup system ensures that data stored     i
      TTS (Transaction Tracking System)
                                                         on the network can be recreated in the event    j
                                                         of a crash or another system failure.
w                                                           Networking packages differ in the type of    k
Backplane                                                backup supported, in the media to which
    A backplane is a circuit board with slots into       material can be backed up, and in the ease
    which other boards can be plugged, as illus-         with which parts of the archived material       m
    trated in the figure “A backplane.” The               can be restored. Backups are generally made
    motherboard in a PC is a backplane.                  to tape or to erasable optical (EO) media.      n
       A segmented backplane is a backplane              No serious network should be backed up
                                                         to floppy disks.
    with two or more buses, each with its own
    slots for additional boards.                            Various types of backups are distin-         p
                                                         guished, including full, differential, and
w                                                        incremental. In full backups, a copy is made
                                                         of all the data.                                r
    The metal bracket at one end of a circuit               In differential and incremental backups,
    board, usually at the back when the board is         only the data that has been added or            s
    plugged into an expansion slot. The back-            changed since the previous backup is
    plate, also known as an end bracket or               included. Differential and incremental back-
    mounting bracket, typically has cutouts for          ups assume a full backup has been done and      u
    connectors and switches. PCs usually come            they merely add to this material. Such back-
    with blank backplates over each expansion            ups use the Archive flag (attribute), which is   v
    slot, which are removed when you plug a
    board into the slot.
                                                         supported by DOS and most networking            w
                                                         environments. This flag is associated with a
                                                         file and is set whenever the file is changed      x
                                                         after the file is backed up.
84     Backup


                                  Keyboard Connector

                                                                             Power Connectors

                                                                             Dip Switches


                                                486   DX


 for Optional                                                                Jumpers

Memory Chip

     The backed up material should generally          for this is that most backup programs will
 be stored in a different physical location           not back up a file that is open. Truly, the
 from the original material, and should be            work of a system administrator is never
 protected from disasters such as fire, flood,          done.
 magnets, theft, and so on.                              When you restore the data, you restore
     Backup operations should be done at a            the last full backup first, then restore each
 time when the network is not being used for          incremental backup made since the last full
 its ordinary activity, which generally means         backup.
 outside regular working hours. One reason
                                                                                           Balun     85

    SEE ALSO                                            w
        Archive                                         Bad Block Table                                      a
    RELATED AR TICLES                                       In storage management, a table in which all      b
                                                            known defective areas of a hard disk are
        Data Protection; Disk Duplexing; Disk
        Mirroring                                           listed to ensure that nothing will be written    c
    B AC K U P T I P S
                                                            to these areas. The process of protecting
                                                            data in this manner is known as bad-block
    s   Keep multiple copies of backups; redundancy
                                                            revectoring, or Hot Fix in Novell’s NetWare.     e
        should be a part of your backup plan.           w                                                    f
    s   Test your backups to make sure that they are
                                                               A balun is a hardware device used to
        what you think they are.
                                                            adjust impedances in order to connect differ-    h
    s   Store your backups in a secure, off-site            ent types of cable. The name comes from
        location.                                           balanced/unbalanced, because the device is       i
    s   Replace your backup media on a regular basis.
                                                            often used to connect twisted pair (bal-         j
                                                            anced) to coaxial (unbalanced) cable.
    s   Consider making incremental backups of                 Baluns may have different connectors at       k
                                                            each end to make them compatible with the
        critical data at more frequent intervals.
                                                            cable types being connected. For example, a
                                                            balun might have a BNC connector at one          m
                                                            end and an RJ-45 connector at the other.
Backward Error Correction (BEC)
                                                               A balun makes it possible to use twisted-     n
        BEC (Backward Error Correction)
                                                            pair wiring that may already be installed in
                                                            parts of a building or office in conjunction
                                                            with coaxial cable that is coming from else-     p
                                                            where or that has been installed more
Bad-Block Revectoring
                                                            recently. The balun controls the electrical
    In data protection, the process by which
    material written to a defective area of the
                                                            signal’s passage from one cable type to the      r
                                                            other, but does not change the signal in any
    hard disk is retrieved and rewritten to a dif-
                                                            other way. Similarly, a balun enables you to     s
    ferent, nondefective area of storage. The
                                                            connect a network interface card designed
    defective area is identified as such in a bad
                                                            for use with coaxial cables to a hub that uses
    block table, so that future writes will not be
    made to the area. Bad-block revectoring is
                                                            twisted-pair cabling.                            u
                                                               Baluns vary with respect to the cable
    known as a Hot Fix in Novell’s NetWare.
                                                            gauge (thickness) supported and to the max-      v
                                                            imum cable distance over which the signal is     w
                                                            supported. This distance may be as high as
                                                            360 to 460 meters (1,200 to 1,500 feet).         x
                                                            Coaxial boosters may be used to increase
86          Bandwidth

    signal strength in the coaxial cable, and thus                  For example, the bandwidth of the human
    increase the distance over which the signal                     voice is roughly 2,700 Hz (3,000 - 300).
    will be supported by the balun. However,                           A larger bandwidth means greater poten-
    such boosters can cost up to ten times as                       tial data-transmission capability. For digital
    mach as a balun, and will only double the                       signals, a higher bit rate represents a larger
    supported distance.                                             bandwidth. However, the higher the fre-
                                                                    quency, the shorter the wavelength. A higher
    B RO A D E R C A T E G O R I E S
                                                                    bandwidth (that is, a higher signal fre-
        Connector; Intranetwork Link
                                                                    quency) means faster transmission, which
w                                                                   means a shorter signal. With a short signal,
Bandwidth                                                           there is a smaller margin for error in inter-
    Bandwidth refers to the amount of data a                        preting the signal. This means that the
    cable can carry; measured in bits per second                    effects of attenuation and other signal dis-
    (bps) for digital signals, or in hertz (Hz) for                 tortion must be kept to a minimum.
    analog signals such as sound waves. An ana-                        A signal traveling along a cable degrades
    log bandwidth is computed by subtracting                        with distance. It is possible to connect the
    the lower frequency from the higher one.                        cable to special components that can clean
                                                                    up and rejuvenate a signal. High-frequency
                                                                    electrical signals must be cleaned up

    W H AT TO L O O K F O R I N A B A L U N

    Baluns may include a stretch of cable (at extra cost, of course). Here are some things to consider when you’re
    shoppping for a balun:

    s   Baluns work most reliably when the cable has low capacitance (20 picofarads/foot or less) and when the
        cable impedance is not too high.

    s   Baluns are available in different qualities, based on the type and gauge (thickness) of cable at either end.
        Make sure the balun you select supports the cable properties and distances you need and then some. To be
        on the safe side, don’t use a balun (or any other kind of connector, for that matter) at the maximum rated

    s   Some network interface card manufacturers recommend specific baluns for their boards. Similarly, some
        manufacturers suggest that you do not use baluns with their hubs or cards. Check with the manufacturer to
        determine whether either is the case with the network interface card or hub you plan to use.

    s   When using a balun on a network, you’ll almost certainly want a balun designed for data transmission,
        because this type is made for direct (rather than reversed) pin-to-pin connections.

    s   Baluns pass signals on, so the balun’s reliability depends on the signal’s quality. For this reason, it’s not a good
        idea to use a balun with passive hubs, which don’t clean and strengthen the signal before passing it on.
                                                                               Bang Path      87

  frequently, which means single cable seg-
  ments must be short.
                                                Digital Transmission Bandwidths
                                                    For digital transmissions, bandwidths range
     Some commonly used frequency bands
                                                    considerably. Here are some examples of
  for analog transmissions are shown in the
  table “Bandwidths on the Electromagnetic          bandwidth values for digital transmissions:      c
  Spectrum.”                                          s   Some digital telephone lines: less than    d
                                                          100 kbps
Radio Spectrum Bandwidths                                                                            e
                                                      s   ARCnet networks: 2.5 Mbps
  Very low frequency (VLF) through super
  high frequency (SHF) are considered the             s   ARCnet Plus networks: 20 Mbps
  radio spectrum. The bandwidths are used as          s   Ethernet networks: 10 Mbps
                                                      s   Fast Ethernet networks: 100 Mbps           h
    s   AM radio broadcasts in the medium
        frequency (MF) range (535 to 1,605            s   Token Ring networks: 1, 4, or 16           i
        kHz).                                             Mbps
        FM radio and VHF television broad-                Fast Token Ring networks: 100 Mbps

        cast in the very high frequency (VHF)
                                                      s   Fiber-optic (FDDI) networks: About
        range (88 to 108 MHz for FM; the
                                                          100 Mbps, but can theoretically be         l
        split ranges from 54 to 88 MHz and
        from 174 to 216 MHz for VHF
                                                          several orders of magnitude higher         m
        television).                                  s   ATM networks: about 655 Mbps, with
                                                          speeds as high as 2.488 gigabits per
    s   Cable stations broadcast over several
        bands (frequency ranges) in the VHF
                                                          second (Gbps) in the future                o
        and ultra high frequency (UHF) ranges   w                                                    p
        (108 to 174 MHz in the VHF range;       Bang Path
        216 to 470 MHz in the VHF and UHF           On the Internet, a bang path is a series of
        ranges).                                    names that specifies a path between two           r
                                                    nodes. A bang path is used in uucp (UNIX-
    s   UHF television broadcasts in the UHF
        range (470 to 890 MHz).                     to-UNIX copy program) and sometimes for          s
    s   Radar operates at 10 different bands
                                                    e-mail (electronic mail) or communications
                                                    on BITNET. The path consists of domain or
        over a huge frequency range (230            machine names separated by exclamation           u
        MHz to 3 THz).                              points (!), known as bangs in some comput-
                                                    ing circles. For example, in a bang path such
                                                    as hither!thither!yon, hither might be a gate-   w
                                                    way, thither a computer, and yon a user.
                                                       Bang paths go back to the days before         x
                                                    automatic routing, because explicit paths        y
88         Bang Path

 NAME                             (FREQUENCY                   WAVE L E N G T H        COMMENTS

 Ultra-low frequency (ULF)        .001 Hz (hertz)–1 Hz         300 Gm (gigameter, or   Subsonic
                                                               billions of meters)—
                                                               300 Mm (megameter, or
                                                               millions of meters)

 Extra low frequency (ELF)        30 Hz–300 Hz                 10 Mm–1 Mm              Audible spectrum

 Voice frequency (VF)             300 Hz–3 kHz (kilohertz)     1 Mm–100 km

 Very low frequency (VLF)         3 kHz–30 kHz                 100 km–10 km
                                  20 kHz–100 kHz               150 km–30 km            Ultrasonic

 Low frequency (LF)               30 kHz–300 kHz               10 km–1 km              Long wave

 Medium frequency (MF)            300 kHz–3 MHz                1 km–100 m              Medium wave

 High frequency (HF)              3 MHz–30 MHz                 100 m–10 m

 Very high frequency (VHF)        30 MHz–300 MHz               10 m–1 m

 Ultra-high frequency (UHF)       300 MHz–3 GHz                1 m–10 cm               Ultra-shortwave

 Super high frequency (SHF)       3 GHz– 30 GHz                10 cm–1 cm

 Extremely high frequency (EHF)   30 GHz–300 GHz               1 cm–1 mm               Ultramicrowave
                                  300GHz–300THz                1 mm–1 micron

 Infrared(IR)                     300 GHz–430 THz              1 mm–0.7 micron

 Visible                          430 THz– 750 THz             0.7 micron–0.4 micron   Visible spectrum

 Ultraviolet (UV)                 750 THz–30 PHz               400 nm - 10 nm          Ultraviolet
                                  (petahertz, or quadril-
                                  lions of hertz; a quadril-
                                  lion is 1015, or roughly

 X-ray                            30 PHz–30 EHz                10 nm–0.01 nm           X-ray
                                  (exahertz, or quintillions
                                  of hertz; a quintillion is
                                  1018, or roughly 260)
                                                                   Basic Access Control (BAC)        89

    were needed when sending to or communi-           w
    cating with another location.                     Baseband                                              a
                                                          In networking, a baseband connection is one       b
Banner Page                                               that uses digital signals, which are sent over
                                                          wires without modulation; that is, binary         c
    A banner page is output by a printer in a
    network environment to separate print jobs.
                                                          values are sent directly as pulses of different
                                                          voltage levels rather than being superim-
    A banner page is also known as a job sepa-
    rator page. Printing of this page is controlled
                                                          posed on a carrier signal (as happens with        e
                                                          modulated transmissions). Baseband net-
    by the network operating system.
                                                          works can be created using twisted-pair,          f
       A banner page might indicate the name of
    the user who printed the file and other infor-
                                                          coaxial, or fiber-optic cable.                     g
                                                              Even though only a single digital stream
    mation. You can eliminate banner pages in
                                                          is transmitted over a baseband connection,        h
    NetWare and in most other network operat-
    ing systems.
                                                          it is possible to transmit multiple signals.
                                                          This is done by multiplexing (combining
w                                                         several signals in a transmission by inter-       j
Base Address                                              leaving the signals using, for example,
    In memory allocation, a base address defines           time slices).                                     k
    the starting or reference location for a block            This digital signaling is in contrast to      l
    of contiguous memory. The memory may be               broadband, in which analog signals are sent
    general-purpose, or it may serve as cache or          over multiple channels at the same time.          m
                                                          Each channel is allocated a different fre-
    port memory. Here are some examples of
                                                          quency range.
    different types of base addresses:
      s   A base I/O (input/output) address is
          the starting location for the memory
                                                      Baseline                                              p
                                                          In performance analysis, a reference level or
          area allocated for an I/O port. The
          processor uses this address to find the          the process of determining this level. For
          correct port when the processor needs           example, in a networking context, a baseline      r
          to communicate with a device.                   measures performance under what is consid-
                                                          ered a normal load. Commonly used base-           s
      s   A base memory address is the starting
                                                          line measures include transmission rate,
          location for a block of memory, such
                                                          utilization level, and number of lost or erro-
          as a buffer area.
                                                          neous packets.                                    u
      s   A base video address is the starting
          location for video memory.
                                                      w                                                     v
                                                      Basic Access Control (BAC)
                                                            BAC (Basic Access Control)                      x
90        Basic Information Unit (BIU)

w                                                w
Basic Information Unit (BIU)                     Baud Rate
    SEE                                              The baud rate is the measure of the number
      BIU (Basic Information Unit)                   of times an electrical signal can be switched
                                                     from one state to another within a second.
w                                                    The faster a switch can occur, the higher the
Basic Link Unit (BLU)                                baud rate.
    SEE                                                 The relationship between baud and bit
      BLU (Basic Link Unit)                          transfer rates depends on the number of bit
                                                     values that are encoded in a single signal.
w                                                    When each signal represents one bit, the bit
Basic Mode
                                                     and baud rates are equal; when a signal
    In an FDDI II network, a mode of operation       encodes multiple bits, the bit rate is a multi-
    in which data can be transmitted using           ple of the baud rate.
    packet-switching. This is in contrast to            The term baud comes from Baudot, the
    hybrid mode, in which both data and voice        name of a French telegraph operator who
    can be transmitted.                              developed a five-bit encoding system in the
                                                     late 19th century. This Baudot code is still
                                                     used, officially known as International Tele-
      FDDI (Fiber Distributed Data Interface)
                                                     graph Alphabet #1.
w                                                       Since it is a violation of the bylaws for
Basic Rate Access (BRA)                              workers in computers and communications
                                                     to pass up an opportunity to create an acro-
                                                     nym, the term also doubles as the acronym
      BRA (Basic Rate Access)
                                                     for bits at unit density.
Basic Telecommunications Access                      C O M P A RE

Method (BTAM)                                          Bit Rate

    SEE                                          w
      BTAM (Basic Telecommunications Access      BBS (Bulletin Board System)
      Method)                                        A BBS is one or more computers set up with
                                                     modems so that users can access those com-
                                                     puters from remote locations. Users dialing
Basic Transmission Unit (BTU)
                                                     into the BBS can send messages, get techni-
    SEE                                              cal support from a vendor, upload or down-
      BTU (Basic Transmission Unit)                  load files, and so on.
                                                        Many BBSs are set up by vendors to
                                                     provide users with a forum for communica-
                                                     tion and with delayed access to technical
                                                                         Bellman-Ford Algorithm     91

    support. Some BBSs are set up to provide              operation) has occurred in the node sending
    services to a specialized market, generally           the beacon frame or in this node’s nearest
    for a fee. (Fee-based BBSs are often given
    more aggrandized names, such as Informa-
                                                          addressable upstream neighbor (NAUN).            b
    tion Services.)
                                                      BCP (Byte-Control Protocols)
w                                                         Protocols that are character- (rather than       d
BCC (Block Check Character)
    In longitudinal redundancy checks (LRCs), a
                                                          bit) oriented.                                   e
    character inserted at the end of a block to
                                                      BEC (Backward Error Correction)                      f
    provide error-detection capabilities. Each of
    the character’s bits is a parity bit for a col-       Error correction in which the recipient          g
                                                          detects an error and requests a retransmis-
    umn of bits in the block.
                                                          sion. The amount of material that needs to
    SEE ALSO                                              be retransmitted depends on the type of con-     i
      CRC (Cyclic Redundancy Check)                       nection, how quickly the error was detected,
                                                          and the protocols being used.                    j
BCD (Binary Coded Decimal)                                C O M P A RE                                     k
    An encoding scheme in which each digit is
    encoded as a four-bit sequence.
                                                            FEC (Forward Error Correction)                 l
B Channel                                                 A bel is a unit for measuring the relative       n
    In an ISDN system, the bearer channel that
    carries voice or data at 64 kilobits per sec-
                                                          intensity of two levels for an acoustic, elec-
                                                          trical, or optical signal. The bel value is
    ond in either direction. This is in contrast to       actually proportional to the logarithm (to       p
    the D channel, which is used for control sig-         base 10) of this ratio.
    nals and data about the call. Several B chan-            For example, if one voltage is 10 times as
    nels can be multiplexed into higher-rate H            strong as another, the higher voltage is one     r
    channels.                                             bel higher than the lower one; similarly, if
                                                          one sound is 100 times as loud as another,       s
                                                          the louder sound is two bels louder. The
      BRI (Basic Rate Interface); PRI (Primary
                                                          decibel, a tenth of a bel, is used more com-
      Rate Interface)
                                                          monly when computing such values.                u
BCN (Beacon)                                          w
                                                      Bellman-Ford Algorithm
    A frame used in a token ring network to
                                                          An algorithm for finding routes through an        w
    indicate that a hard error (one that is serious
    enough to threaten the network’s continued
                                                          internetwork. The algorithm uses distance        x
                                                          vectors, as opposed to link states. The
92        BER (Basic Encoding Rules)

    Bellman-Ford algorithm is also known                     The Value, or contents, field represents
    as the old ARPAnet algorithm.                         the information associated with the ASN.1
                                                          object as a byte string. For primitive types,
                                                          this is a single value; for constructed types,
                                                          there may be several values, possibly of dif-
w                                                         ferent types, involved.
BER (Basic Encoding Rules)
    In the ISO’s Abstract Syntax Notation One         BER Encoding
    (ASN.1), the BER are the rules for encoding
                                                          The encoding may be any of the following:
    data elements. Using the BER, it is possible
    to specify any ASN.1 element as a byte                  s   Primitive/fixed length, which consists
    string. This string includes three compo-                   only of a primitive object and which is
    nents, and the encoding may take any of                     always a fixed length. For example, an
    three forms, depending on the information                   integer variable is of this type.
    being encoded.                                          s   Constructed/fixed length, which con-
Components of BER                                               sists of a group of objects and values,
                                                                with a fixed total length. For example,
    The components of BER are the Type,                         this might be a record with only pre-
    Length, and Value fields.                                    defined components, all of which have
       The Type, or identifier, field indicates the               a fixed and known length.
    class of object, as well as the string’s form.
                                                            s   Constructed/variable length, which
    Examples of ASN.1 types include BOOL-
                                                                consists of a group of objects whose
                                                                total size may vary from case to case,
                                                                so that a special value is needed to
    Of these, the first two are primitive, the next
                                                                indicate the end of the value.
    three may be primitive or constructed types,
    and the SEQUENCE OF type is always con-                  The BER can provide an encoding for any
    structed. (A primitive object consists of a           valid ASN.1 object. One difficulty is that the
    single element of a particular type of                rules can sometimes provide more than one.
    information, such as a number or logical              In this case, the rules may be too general,
    value; a constructed type is made up of other         because all the “synonymous” rules eat up
    simpler elements, such as primitive objects           overhead.
    or other constructed types.)
       The Length field indicates the number of
                                                      BER Variants
    bytes used to encode the value. Values actu-
    ally may have a definite or an indefinite               Several variants of the BER have been
    length. For the latter case, a special value is       proposed and are being developed. In gen-
    included in the last byte.                            eral, these are designed to provide faster,
                                                          simpler, and/or more generic encodings. The
                                                                                      Big-Endian     93

    following are some of the alternatives that        w
    have been proposed:                                Berkeley Internet Name Domain                        a
       s   CER (canonical encoding rules), which                                                            b
           represent a subset of the BER. With the
           canonical rules, it should be possible
                                                             BIND (Berkeley Internet Name Domain)
           to eliminate any redundant paths,
           which can slow down performance
                                                       BERT (Bit Error Rate Tester)                         e
                                                           A hardware device for checking a transmis-
       s   DER (distinguished encoding rules),             sion’s bit error rate (BER), or the proportion   f
           which are also a subset of BER.                 of erroneous bits. The BERT sends a pre-         g
                                                           defined signal and compares it with the
           LWER (lightweight encoding rules),

           which make faster encoding possible,
                                                           received signal. BERTs are moderately            h
                                                           expensive devices that are used most com-
           but may result in larger transmissions.
                                                           monly for troubleshooting wiring.
       s   PER (packed encoding rules), which
                                                           C O M P A RE
           are used to compress the information
           about an object.                                   BLERT (Block Error Rate Tester)               k
    P R I M A R Y S O U RC E S                         w
                                                       BIA (Burned-In Address)
       CCITT recommendation X.209; ISO
       document 8825                                       A hardware address for a network interface
    B RO A D E R C A T E G O R Y
                                                           card. Such an address is assigned by the         n
                                                           manufacturer and is unique for each card.
       ASN.1                                                                                                o
w                                                      BIB (Bus Interface Board)                            p
BER (Bit Error Rate)
    Number of erroneous bits per million (or
                                                           An expansion board. In particular, a net-        q
                                                           work interface card (NIC), which serves as
    billion or trillion) bits in a transmission or a       an interface between the node (computer)         r
    transfer (as from a CD to memory). The                 and the network medium.
    BER depends on the type and length of                                                                   s
    transmission or on the media involved in a         w
    transfer.                                          Big-Endian                                           t
    C O M P A RE
                                                           In data transmission and storage, the order      u
                                                           in which bytes in a word are processed
       BLER (Block Error Rate)
                                                           (stored or transmitted). The term comes          v
                                                           from Jonathan Swift’s Gulliver’s Travels,        w
                                                           in which a war is fought over which end of
                                                           an egg should be cracked for eating. This        x
94        BIND (Berkeley Internet Name Domain)

    ordering property is also known as the pro-    the bindery, this information is represented
    cessor’s byte-sex.                             as a flat database.
       In big-endian implementations, the high-       The bindery has three types of
    order byte is stored at the lower address.     components:
    Processors in mainframes (such as the IBM
                                                     Objects: Users, devices, workgroups,
    370 family), some minicomputers (such as
                                                       print queues, print servers, and so on.
    the PDP-10), many RISC machines, and also
                                                       Most physical and logical entities are
    the 68000 family of processors use big-
                                                       regarded as objects.
    endian representations. The IEEE 802.5
    (token ring) and the ANSI X3T9.5 FDDI            Properties: Attributes, specifically, as
    standards use big-endian representations. In       assigned to bindery objects, such as
    contrast, the 802.3 (Ethernet) and 802.4           full name, login restrictions, or group
    (token bus) standards use little-endian            membership information.
                                                     Property data sets: The values that will
       The term is used less commonly to refer
                                                       be stored in an object’s property list.
    to the order in which bits are stored in a
    byte.                                              The bindery has been replaced in Net-
                                                   Ware 4.x by the NetWare Directory Services
    C O M P A RE
                                                   (NDS), in which information is represented
      Little-Endian; Middle-Endian
                                                   hierarchically in tree format.
w                                                      However, version 4.x includes bindery-
BIND (Berkeley Internet Name                       emulation capabilities, which makes it possi-
Domain)                                            ble to integrate bindery-based objects into a
    In the Internet community, a domain name       network based on NDS. In NetWare 4.1, the
    system (DNS) server developed at the Uni-      Bindery services utility creates a bindery
    versity of California, Berkeley, and used on   context within which the bindery objects
    many Internet machines.                        appear as a flat database—as required by
                                                   earlier versions of NetWare. This perspective
w                                                  is valid in only a limited context, which
Bindery                                            makes it possible to integrate the bindery
    In Novell’s NetWare products, the bindery is   information into the NDS while still provid-
    a database maintained by the network oper-     ing a pre-4.x server with access to the bind-
    ating system (NOS) on each server. The         ery’s contents.
    bindery is located in the SYS:SYSTEM direc-        Another 4.1 utility, NetSync, makes it
    tory and contains information about all the    possible to manage up to 12 NetWare 3.x
    users, workstations, servers, and other        servers within a NetWare 4.1 network. This
    objects recognized by the server.              makes all 12 servers look like a single server
       The bindery information determines the      to users—a user would need only one login
    activities possible for the user or node. In   to access as many of the NetWare 3.x servers
                                                                            BIOS Extensions      95

    as desired. (As always, such access assumes     w
    that the user has the necessary privileges.)    Binding and Unbinding                               a
        With NetSync, it also becomes easier to
    update resources on different machines.
                                                        In a local-area network (LAN), binding is       b
                                                        the process of associating a communication
    B RO A D E R C A T E G O R Y
                                                        protocol, such as TCP/IP, IPX/SPX, or           c
                                                        AppleTalk, and a network interface card
                                                        (NIC). Unbinding is the process of dissociat-
w                                                       ing the protocol from the NIC.                  e
Bindery Emulation                                          The LAN driver for a card must have at
    In Novell NetWare 4.x, bindery emulation is         least one communication protocol associ-        f
    a NetWare Directory Service that makes the          ated with it. The LAN driver will be able to    g
    Directory database emulate a flat database.          process only those packets that use the asso-
       In NetWare 2.x and 3.x, information              ciated protocol.                                h
    about all network objects is stored in a
                                                    w                                                   i
    flat database, called the bindery. A flat data-   BIOS (Basic Input/Output System)
    base is one in which all objects in the                                                             j
                                                        The BIOS is a collection of services on a
    database exist as entities of equal standing;
    an object cannot contain another object.
                                                        ROM (read-only memory) chip. The BIOS           k
                                                        services enable hardware and software,
    In NetWare 4.x, network objects and their
                                                        operating systems and applications, and also
    related information are contained in a hier-
    archical database called the Directory. A
                                                        applications and users to communicate with      m
                                                        each other. The BIOS services are loaded
    hierarchical database can contain several
    levels of objects, which means that objects
                                                        automatically into specific addresses and        n
    can contain other objects.
                                                        should always be accessible.
                                                           BIOS services are updated and expanded
       Bindery emulation allows programs that
    were written to run under the NetWare
                                                        to handle newer devices and greater             p
                                                        demands. To get a newer BIOS, you simply
    bindery to find the network object informa-
                                                        need to replace the ROM chip in your com-
    tion they need in NetWare 4.x’s Directory by
    making the information in the Directory
                                                        puter with an appropriate upgrade chip.         r
    appear as a flat structure.
       Such bindery emulation is provided
                                                    BIOS Extensions
    by the Bindery services utility, which makes        A collection of services that supplement        t
    the bindery’s contents look appropriate
    for whatever server is querying it (i.e, 3.x
                                                        those provided by the standard BIOS (Basic      u
                                                        Input/Output System). Like the standard
    or 4.x).                                            BIOS, BIOS extensions are implemented on        v
                                                        a ROM (read-only memory) chip, located
    B RO A D E R C A T E G O R Y
                                                        on the motherboard or on an expansion
                                                        board.                                          x
96        Bipolar with 8 Zero Substitution (B8ZS)

w                                                      s   Distribution services are those in
Bipolar with 8 Zero Substitution                           which information (in the form of
(B8ZS)                                                     video, documents, or data) can be
    SEE                                                    broadcast to whomever has the
      B8ZS (Bipolar with 8 Zero Substitution)              resources and rights to receive the
                                                           broadcast. Distribution services are
BISDN (Broadband ISDN)                                     divided into those for which the user
                                                           has no control over the presentation
    BISDN is an extension of the ISDN (Inte-               (other than to turn it on or off) and
    grated Services Digital Network) to allow              those where the user can control which
    multiple types of information to be transmit-          elements are received. Examples of the
    ted. BISDN can handle voice, video, and                former include TV programming and
    graphics, as well as data.                             electronic newspapers; examples of the
       Whereas ISDN networks generally use                 latter include retrieval of selected news
    some form of time division multiplexing                items and certain online courses.
    (TDM) for actual transmissions, BISDN net-
    works generally use ATM (asynchronous           P R I M A R Y S O U RC E S
    transfer mode) as their transmission technol-      BISDN is discussed in more than a few of
    ogy. ATM is often erroneously regarded as          the documents in the ITU-T I.xxx docu-
    being equivalent to BISDN.                         ment series. For example, I.113 provides
                                                       a vocabulary for BISDN, and I.121 pro-
BISDN Services                                         vides a list of the documents that discuss
                                                       BISDN or ATM or both. These include
    Figure “BISDN Services” summarizes the
                                                       I.150 (ATM for BISDN), I.211 (BISDN
    kinds of capabilities that have been defined
                                                       services), I.311 (General BISDN network-
    for BISDN networks. The services are
                                                       ing aspects), I.327 (BISDN functional
    grouped into two main groups, each with
                                                       architecture), I.361, I.362, and I.363
    multiple service classes:
                                                       (ATM layers), I.413 and I.432 (BISDN
      s   Interactive services are those in which      User-network interface), and I.610 (Oper-
          the user can initiate the service and        ation and maintenance for BISDN). In
          influence its direction. Three classes        some cases, these recommendations must
          are distinguished, and each class            be read in relation to their ISDN counter-
          includes several examples. For exam-         parts, whose numbers are generally lower
          ple, conversational services include         than the corresponding BISDN docu-
          video-conferencing and video-                ment. For example, I.210 discusses ISDN
          telephony (for shopping, learning,           services.
          etc). Online research is included
                                                    C O M P A RE
          among interactive services.
                                                       ISDN (Integrated Services Digital
                                                                  Bit Error Rate Tester (BERT)   97

    B I S D N S E RV I C E S                                                                          a
                      Interactive Services                        Distribution Services               b
             Conversational Service                                 ..without individual user         d
               Video (e.g., videoconferencing, video
                                                                    presentation control
                                                                     Video (e.g., cable and
               Data (information, files, teleaction
                                                                    extended cable TV, high-
                                                                    definition TV, pay-TV)            f
               telemetry, alarms, etc.)                              Text, graphics, and ima-
                                                                    ges (e.g., document
               Documents (high-speed fax, images)
              Messaging Service
                Video mail                                                                            j
                Document mail                                       ..with individual user
                                                                    presentation control              k
                                                                      Text, graphics,
                                                                    sound, and images                 l
                                                                    (e.g., remote education,
              Retrieval Service                                     tele-advertising,                 m
                Video, document, and data retrieval                                                   n
w                                                      w                                              r
Bit                                                    Bit Error Rate (BER)
    A binary digit; the smallest unit of infor-
    mation. A bit can have a value of 0 or 1                 BER (Bit Error Rate)                     t
    in a digital system. All but the low-level
    protocols move information in larger               w                                              u
                                                       Bit Error Rate Tester (BERT)
    chunks, such as bytes, which consists of                                                          v
    multiple bits.                                         SEE
                                                             BERT (Bit Error Rate Tester)             w
98        Bit Interval

w                                                          most Internet installations. An address such
Bit Interval                                               as user@computer.bitnet will suffice, because
    Bit interval, also known as bit time, refers to        most Internet mail programs recognize
    the amount of time a digital signal is left at a       bitnet as a pseudo domain name.
    particular voltage level to indicate a value.             In Canada, BITNET is known as
    Usually, the level will indicate the value of a        NetNorth, and in Europe it is known as
    single bit, but it is possible to encode more          EARN (for European Academic Research
    than a single bit in a voltage level, thereby          Network).
    transmitting more than one bit in a single bit
    interval.                                          Bit Rate
       In general, the longer the bit interval, the
                                                           Bit rate is a measure of throughput, or rate
    slower the transmission rate. For example,
                                                           of data transfer. It represents the number of
    when encoding a single bit at a time, a bit
                                                           bits that are transmitted within a second in a
    interval of .01 second means a transmission
                                                           digital communication, measured in bits per
    rate of only 100 bits per second (bps).
                                                           second (bps). The faster the bit rate, the
    RELATED AR TICLES                                      shorter the bit interval (the interval to signal
       Bit Rate; Encoding, Signal                          a bit value). For example, at a bit rate of
                                                           5,000 bps, each bit interval can be at most
                                                           .0002 second when a single bit is transmit-
BITNET (Because It’s Time Network)
                                                           ted in each bit interval.
    BITNET is a computer network that con-                    Bit rate is often used interchangeably with
    nects many educational institutions in North           baud rate, but these two measurements are
    America and Europe. BITNET was set up                  not exactly the same. Baud rate refers to the
    through EDUCOM, a nonprofit educational                 number of electrical signal transitions made
    consortium. It is designed to provide com-             in a second. If a single bit is encoded in each
    munication facilities and easy access to               signal, the bit rate and baud rate will be
    files—even from remote locations—provided               equal. However, if multiple bits are encoded
    that the user has the appropriate access priv-         in a single signal, the bit rate will be higher
    ileges. Today, BITNET connects more than               than the baud rate.
    1,000 locations.
       Partly because the early nodes were pre-        w
    dominantly IBM mainframes, BITNET still            Bit Stuffing
    uses the RSCS (Remote Spooling Communi-                In data transmission, a technique for ensur-
    cations Subsystem) and NJE (Network Job                ing that specific bit patterns do not appear
    Entry) protocol suites. Because of this, a             as part of the data in a transmission. For
    gateway is needed to communicate with                  example, if six consecutive 1 values are
    other networks, such as the Internet.                  encountered in the transmitted data, a 0 bit
       Once a gateway between the Internet and             would be inserted after the fifth consecutive
    BITNET is known, it is relatively easy to              1 bit. The receiver removes any inserted bits
    send a message to a user on BITNET from                when processing the transmission.
                                                                                       Block       99

w                                                       lightning strikes, and other natural and
BIU (Basic Information Unit)                            man-made disasters.
       In SNA network communications, a                 SEE ALSO                                        b
    packet of information created when the
    transmission control layer adds a request/
                                                          Power Disturbances                            c
    response header (RH) to a request/response
    unit (RU). This unit is passed to the path
                                                    BLER (Block Error Rate)
    control layer.                                      In communications, an error rate based on       e
                                                        the proportion of blocks with errors. Com-
                                                        pare it with BER (bit error rate), which is
       SNA (Systems Network Architecture)
                                                        based on the number of erroneous bits           g
w                                                       per million (or billion or trillion) bits in
BIU (Bus Interface Unit)                                a transmission.                                 h
    An adapter card. In particular, a network
    interface card (NIC), which acts as an          BLERT (Block Error Rate Tester)
    interface between a node (computer) and                                                             j
                                                        A hardware device for determining a trans-
    the network.
                                                        mission’s block error rate (BER), which is      k
BIX (BYTE Information Exchange)
                                                        the proportion of blocks with erroneous
                                                        bits. This device is also known as a BKERT.
    BYTE Magazine’s commercial online infor-
    mation service. BIX provides the usual
    gamut of mail, news, and entertainment ser-
                                                    Block                                               n
    vices, as well as Internet access—including
                                                        A block is an area of memory or storage
                                                        with a fixed size. A network operating sys-
    e-mail (electronic mail), ftp (file transfer
    protocol), and telnet services. In addition
                                                        tem block can be anywhere from 4 to 64          p
                                                        kilobytes (KB). DOS blocks are typically a
    to a base monthly fee (which depends on the
                                                        multiple of 2 KB. NetWare blocks are typi-
    amount of access requested), costs for Inter-
    net use include access and storage charges.
                                                        cally 4 KB. However, the actual block size      r
                                                        depends on the size of the volume on which
    FO R I N F O R M A T I O N                          storage is being allocated.                     s
       Call 800-695-4775; 617-354-4137. You                In some environments, such as in
       can use telnet to access BIX over the            NetWare, a block represents the smallest
       Internet. To do this, telnet to                  chunk of storage that can be allocated at a     u
       x25.bix.com.                                     time. (In NetWare, you can accept the sug-
                                                        gested block size, which is based on the size   v
Blackout                                                of the volume, or you can specify the block
                                                        size you want to use.)
    A total loss of electrical power. Blackouts                                                         x
    can be caused by cut or broken power lines,
100       Block Error Rate (BLER)

      Two types of blocks are distinguished:            similar, but not identical, Ethernet variant
                                                        defined in the IEEE 802.3 standard.
      Disk-allocation block: Used to store
        network data, at least temporarily.             SEE ALSO
      Directory-entry block: Used to store
        directory information.                      w
       NetWare 4.x supports block suballoca-
    tion, in which a block can be broken into           In gopher environments on the Internet, a
    512-byte chunks. These chunks can be used           bookmark is used to mark a specific menu
    to store the ends of several files. For exam-        or directory on a gopher server. Once the
    ple, with a 4 kilobyte (KB) block size, three       bookmark has been created and placed at
    5 KB files would fit into four blocks. Each of        the desired location, it’s possible to get
    the files would use one block and two 512-           almost immediate access to that location,
    byte chunks in the fourth block. In contrast,       rather than having to work your way
    these files would require six blocks (two per        through layers of menus.
    file) in NetWare 3.x.                                SEE ALSO

w                                                         Gopher
Block Error Rate (BLER)
    SEE                                             Boot
      BLER (Block Error Rate)                           The process by which a computer is started
                                                        up and its operating system kernel is loaded
Block Error Rate Tester (BLERT)                         into RAM (random-access memory) is called
                                                        the boot, or bootstrap, process. Although
                                                        the details may differ when booting to dif-
      BLERT (Block Error Rate Tester)                   ferent disk operating systems or network
w                                                       operating systems, the basic steps are the
BLU (Basic Link Unit)                                   same:
    In IBM’s SNA (Systems Network Architec-               s   Execute a hardware self-test.
    ture) networks, a block, or packet, of infor-
                                                          s   Look in a predefined place for the boot
    mation at the data-link layer.
                                                              sector and load this code.
                                                          s   Execute the boot sector program to
      SNA (Systems Network Architecture)
                                                              load other programs.
w                                                         s   Execute these programs to load still
Blue Book Ethernet
                                                              other programs or to configure the
    Ethernet version 2.0. This term is sometimes              operating environment.
    used to distinguish Ethernet 2.0 from the
                                                                               BRI (Basic Rate Interface)        101

       s   Repeat the previous step as often as              w
           dictated by the programs being loaded               BRA (Basic Rate Access)                                  a
           and by their initialization code.                     Access to an ISDN Basic Rate Interface                 b
                                                                 (BRI), an interface with two 64 kilobits per
BOOTCONF.SYS                                                     second (kbps) B channels (for voice and                c
    In Novell NetWare, a configuration file that
                                                                 data) and one 16 kbps D channel (for call
                                                                 and customer information). Compare it with
    specifies how a diskless workstation can
    boot the operating system in order to access
                                                                 PRA, which is access to a PRI (Primary Rate            e
    the network.                                                                                                        f
                                                               Braid Shield                                             g
Boot ROM
    A ROM (read-only memory) chip used in
                                                                 In coaxial cable, a braid or mesh conductor,           h
                                                                 made of copper or aluminum, that sur-
    diskless workstations to enable these
                                                                 rounds the insulation and foil shield. The             i
    machines to boot and connect to a network.
                                                                 braid helps protect the carrier wire from              j
w                                                                electromagnetic and radio frequency
Bounce                                                           interference.                                          k
    A term for the action of returning an unde-                  SEE ALSO                                               l
    liverable e-mail message. In such a case, the                   Cable, Coaxial
    postmaster on the system returns the mes-                                                                           m
    sage, along with a bounce message, to the
    sender.                                                    BRI (Basic Rate Interface)                               n
                                                                 A BRI is an interface between a user and an            o
                                                                 ISDN (Integrated Services Digital Network)
    T H E D O S B O OT S T R A P P R O C E S S

    1. A program (the ROM-BIOS) in ROM (read-only -memory) executes. This program checks the hardware
       components by doing a POST (power-on self-test).                                                                 s
    2. The ROM-BIOS program loads and executes a program from the boot sector on a floppy or hard disk.                  t
    3. This boot sector program loads hidden files, which, in turn, load the basic device drivers for DOS (key-          u
       board, disk, and display) and execute the DOS initialization code. Part of this initialization loads the DOS
       kernel.                                                                                                          v
    4. The DOS kernel builds various tables it will need, initializes device drivers, and executes instructions found   w
       in CONFIG.SYS, if this file exists.
    5. The DOS kernel loads COMMAND.COM, the DOS command processor.
102       Bridge

    switch. The BRI specifies two 64 kilobit per        router) is a device that has the capabilities of
    second (kbps) B channels (for voice and            both a bridge and a router.
    data) and one 16 kbps D channel (for cus-             A bridge’s capability to segment, or
    tomer and call information).                       divide, networks is one difference between a
       This channel combination is sometimes           bridge and a repeater. A repeater is a device
    denoted as 2B+D. It can be compared with           that moves all packets from one network
    PRI (Primary Rate Interface).                      segment to another by regenerating, retim-
       Access to a BRI is provided by a BRA            ing, and amplifying the electrical signals.
    (basic rate access).                               The main purpose of a repeater is to extend
                                                       the length of the network transmission
                                                       medium beyond the normal maximum cable
    The term bridge generally refers to a hard-
    ware device that can pass packets from one       Protocol Independence of Bridges
    network to another. Bridges operate at the
                                                       A bridge is independent of, and therefore
    OSI Reference Model’s second lowest layer,
                                                       can handle packets from, higher level proto-
    the data-link layer. A bridge makes the net-
                                                       cols. This means that different higher level
    works look like a single network to higher
                                                       protocols can use the same bridge to send
    level protocols or programs.
                                                       messages to other networks.
       A bridge serves both as a medium (the
                                                          To protocols at higher OSI layers (most
    bridge part) and as a filter. It allows packets
                                                       immediately, the network layer), the pres-
    from a node on one network to be sent to a
                                                       ence of a bridge is transparent. This means
    node on another network. At the same time,
                                                       that two networks connected by a bridge are
    the bridge discards any packets intended
                                                       treated as part of the same logical network
    for the originating network (rather than
                                                       by protocols such as Novell’s IPX/SPX,
    passing these to the other network).
                                                       IBM’s NetBIOS, or the widely used TCP/IP.
Bridges versus                                         This transparency makes it possible to
Routers, Brouters, and Repeaters                       access a logical network that is much larger
                                                       than the largest physical network allowed.
    The terms bridge and router are often used
    interchangeably. In fact, in older documen-      Packet Transmission
    tation, Novell referred to its routers as
                                                       Because it operates at the data-link layer, a
    bridges. A router is a device that can send
                                                       bridge just checks the address information in
    packets to network segments on the way to
                                                       a packet to determine whether to pass the
    their destination. Unlike bridges, routers
                                                       packet on. Beyond that checking, a bridge
    operate at the network layer of the OSI Ref-
                                                       makes no changes to a packet.
    erence Model. However, bridges and routers
                                                          A bridge sees each packet that is trans-
    have come to take on some of each others’
                                                       mitted on each of the networks the bridge
    properties. In fact, a brouter (from bridging
                                                       connects. If a packet from network A is
                                                                                        Bridge      103

  addressed to a local node (that is, to one in         LLC Layer versus MAC Layer Bridges
  network A), the bridge discards the packet
                                                          MAC-layer bridges operate at the media-
  since the packet will be delivered internally
  through the network. On the other hand, if
                                                          access control (MAC) sublayer, the lower         b
  a packet from network A is addressed to a
                                                          sublayer into which the IEEE divides the
                                                          data-link layer of the OSI Reference Model.
  remote node (on network B), the bridge
  passes the packet over to network B. The
                                                          These bridges can connect only networks          d
                                                          using the same architecture (Ethernet to
  figure “A simple local bridge” shows how
                                                          Ethernet, Token Ring to Token Ring, and so
  a bridge can connect two networks.
     The bridge greatly reduces traffic on both
                                                          on), because the bridge expects to handle a      f
                                                          particular packet format, such as Ethernet
  networks by protecting each network from
                                                          or ARCnet.                                       g
  the other’s local messages. This makes each
                                                             LLC-layer bridges operate at the upper
  of the smaller networks faster, more reliable,
                                                          sublayer of the data-link layer, the logical
  and more secure, while retaining transparent
  communication with the other network (or
                                                          link-level control (LLC) sublayer. These         i
                                                          types of bridges can connect different archi-
                                                          tectures (such as Ethernet to Token Ring),       j
     When routing packets, a bridge uses only
  node addresses; it does not take network
                                                          because these architectures use the same         k
                                                          LLC sublayer format, even if they use differ-
  addresses into account. A node address is a
  physical address, associated with a network
                                                          ent formats at the MAC sublayer.                 l
                                                             Most older bridges are of the MAC-layer
  interface card (NIC), rather than with a par-
                                                          type and can connect only same-architecture
  ticular network.
                                                          networks; most newer products are of the         n
Types of Bridges                                          LLC-layer type and can connect dissimilar
                                                          architectures.                                   o
  Bridges can be categorized by several differ-
  ent features. The table “Bridge groupings”            Transparent Routing versus Source Routing
  summarizes the various categories.
                                                          The manner in which a bridge routes pack-
  B R I D G E G RO U P I N G S
                                                          ets depends largely on the architectures         r
                                                          involved. Bridges connecting Ethernet net-
  Level              LLC (logical-link-control) layer     works use transparent routing, a packet-
                     versus MAC (media-access-
                     control) layer
                                                          routing method in which the bridge deter-        t
                                                          mines a route. Transparent bridges deter-
  Operation          Transparent versus source            mine “on the fly” where a packet belongs.         u
                                                          Such bridges learn and store the location of
                                                          each node, and then route packets accord-
  Location           Internal (card) versus external
                                                          ingly. A transparent bridge can carry out its    w
                                                          routing without explicit instruction or atten-
  Bridged distance   Local versus remote                  tion from the user. The bridge determines        x
104    Bridge


 the locations of a node by looking it up in a   information is inserted by the sender and
 table the bridge has built.                     can be determined by sending a discovery
    In contrast, most bridges connecting         packet. This packet uses the spanning tree
 Token Ring networks use source routing.         algorithm to find the most efficient route to
 This is a deterministic routing method in       the destination and reports this route to the
 which the source node must provide the          sender.
 route as well as the destination for the           Source routing bridges determine an
 packet. The source node learns the available    explicit path to the destination node and
 routes through route discovery. The routing     include this routing information in the
                                                                                     Bridge    105

  packet. Surprisingly, the requirements for          serial connection to a modem). External
  source routing capabilities are considerably        bridges need their own power supply and
  more complex than for transparent bridges.
  Accordingly, source routing capabilities are
                                                      they usually include a connector for access-
                                                      ing WANs.
  generally available as options for a bridge.                                                          c
  Although source routing requires more work        Local versus Remote Bridges
  to find the path initially, it is more efficient
                                                      A bridge may be local or remote. A local
  once the path has been established because
                                                      bridge connects two networks in the same
  there is no longer any reason for the bridge
  to find a path.
                                                      geographical location, such as networks on        f
                                                      either side of the hall or on either side of an
     According to the IEEE 802.3 specifica-
                                                      office floor. Usually, these types of bridges       g
  tions, all bridges should be capable of using
  transparent routing. Some can also do
                                                      are added to break a large, busy network
                                                      into two smaller networks. This reduces net-
  source routing. A bridge can distinguish
  between the two approaches by checking the
                                                      work traffic on each of the newly formed           i
  packet being sent. Depending on the value of
                                                         By using the spanning tree algorithm           j
  a particular bit in the source address field, a
  packet may include source-routing
                                                      specified in the IEEE 802.1 standard, local        k
                                                      bridges can ensure that only a single path is
                                                      used to send a packet between a source and        l
                                                      a destination. If this path is not usable, the
Internal versus External Bridges
                                                      algorithm can find an alternate path.
  A bridge may be internal or external. An
  internal bridge is on a card plugged into an
                                                         A remote bridge connects two networks          n
                                                      separated by considerable geographical dis-
  expansion slot in a server. The server is part      tance, large enough to require a telecommu-       o
  of both networks. An internal bridge gets its
  power from the PC’s bus. Internal bridges
                                                      nications link. Remote bridges must be used       p
                                                      in pairs, with one at each end of the link, as
  generally include multiple types of connec-         shown in the figure “A simple configuration         q
  tors. A special type of internal bridge is used     involving remote bridges.”
  to connect to wide-area networks (WANs).               A remote bridge connects to a local-area
  This type of bridge will have connectors for
  modem or telephone connections, such as
                                                      network at one end and to a switching net-        s
                                                      work, such as one with an X.25 interface, at
  D-shell or RJ-type connectors.                      the other end. Each remote bridge is con-         t
     An external bridge is a stand-alone com-
  ponent to which each network is connected
                                                      nected to a network at one port and to a net-
                                                      work cloud at another port. (A cloud is a
  by cable. The external bridge is part of both       working concept that is used to indicate a        v
  networks. An external bridge generally has          connection that is taken for granted, for pur-
  multiple connectors; for example, BNC for           poses of the discussion and whose details are
  coaxial cable (as in Ethernet or ARCnet net-
  works); modular (RJ-xx) for twisted-pair
                                                      not specified.)                                    x
  cable, and possibly DB-9 or DB-25 (for                                                                y
106     Bridge

    The interfaces are likely to be different             RS-232) at the other. The cloud represents
 at these two ports. For example, a remote                the point-to-point link between the two
 bridge may connect to an Ethernet network                remote bridges.
 at one port and to a serial interface (such as

                                                                                     Bridge    107

      Remote bridges also need a protocol to           learning bridges, since static bridges do not
  communicate with each other. For example,            meet IEEE 802.1 specifications.
  if the remote bridges communicate over an
  ISDN or an X.25 line, the bridge at each end       Multiple Bridges and the                           b
  needs to be able to communicate using the          Spanning Tree Algorithm                            c
  switched network (ISDN or X.25) protocol.
      The throughput in a remote bridge is
                                                       Multiple bridges may be used to connect          d
                                                       several networks. Any one bridge connects
  likely to be limited by the long-distance            only two networks directly, but may connect      e
  connection. At the local end, the bridge will        more than two networks indirectly. The
  generally have the same nominal speed as             bridge is attached to each network by a port.
  the network (10 Mbps for Ethernet, 4 or 16               If there are multiple bridges, the bridges   g
  Mbps for Token Ring, and so on). At the              communicate with each other and establish
  remote end, the throughput will depend on            a layout in order to find a spanning tree for     h
  the type of connection. At this end, possible
  speeds may run from a few kilobits per sec-
                                                       all the networks. A spanning tree is one that    i
                                                       includes paths to all nodes that can be
  ond to several megabits per second.                  reached on the network but includes no           j
                                                       more paths than are necessary to completely
Learning Bridges versus Static Bridges
                                                       interconnect the nodes and networks
  A learning bridge is one that automatically          involved. Most important, a spanning tree        l
  builds a table of node addresses, based on           does not include any loops (closed paths)
  the NICs the bridge finds on the network.             which could trap a packet, thereby effec-        m
  The bridge builds the table by using the
  information broadcast when a new node
                                                       tively shutting down the network.
                                                           Because larger network clusters make
  logs on and by checking on the source and            multiple paths possible, there is the danger     o
  destination addresses as packets pass                that the same message will get broadcast all
  through the bridge.                                  over the networks through multiple paths.
     The performance of a learning bridge              This will produce a great deal of extraneous     q
  improves over time as the bridge com-                network traffic and can, in fact, bring down
  pletes its table of node locations. Until it         the network. A closed path, or loop, among       r
  knows the location of a node, the bridge
  assumes the node is on the remote network
                                                       the networks could be damaging because
                                                        it could start an unending packet-passing
  and so passes on the packets. The bridge is          process. The spanning-tree algorithm, speci-     t
  constantly updating its table—adding new             fied in IEEE 802.1, is applied to provide a
  addresses and dropping addresses that have           path between every pair of accessible nodes      u
  not been mentioned within a period of time.          on the network and ensure that there are no      v
     In contrast, a static bridge is one that can-     loops in the paths to be used by the bridge.
  not build its own address table. Instead, the            Although the spanning tree algorithm         w
  addresses must be entered by hand. Fortu-            ensures that the same packet won’t take
  nately, static bridges have all but disap-           multiple paths to the same destination, the
  peared. Just about all modern bridges are            algorithm doesn’t rule out the possibility of    y
108      Bridge


 When you’re investigating bridges, you’ll want to get details about bridge features and capabilities.Vendors
 should be able to provide both marketing and technical information about their products. Make sure
 to get the technical information. The vendors’ materials should provide information about at least the fol-

 s   Whether the bridge is local or remote.

 s   Whether the bridge is internal or external.

 s   Media and architecture supported for the local network; for example, twisted-pair Ethernet, 16 Mbps
     Token Ring, or FDDI. It’s a good idea to ask explicitly about your particular configuration and to get the
     answer in writing.

 s   If applicable, what interface the bridge supports for a remote connection. For example, it may support
     RS-232, RS-422, V.35, T1, or DSx.

 s   Number of ports.

 s   Transmission speeds, both local and long distance, if applicable. The smaller of these values is the critical
     one. Number of packets passed is generally a more useful figure than the actual bit-transfer rate.

 s   Whether the bridge supports load balancing.

 s   Whether the bridge can collect network performance data, such as number of packets received, for-
     warded, and rejected, number of collisions, and errors during a transmission. Such network management
     services may require additional software (which may cost several thousand dollars).

 s   Price, which can range from a few hundred dollars to over $10,000.

 When you’re selecting a remote bridge, you need to worry about compatibility with the network and also
 with the long-distance services that will be used. Keep in mind that you may need to budget for two remote
 bridges if you’re responsible for the networks at both ends of the connection.

 For more specific and more advanced questions, such as about a bridge’s compatibility with a particular net-
 work configuration, you may need to talk to the bridge vendor’s technical support staff. In many cases, the
 network vendor (Novell, Banyan, and so on) will have a database of hardware that has been explicitly tested
 with the vendor’s networking products. Be forewarned that these vendors may want to charge you for
 revealing this information.
                                                                  Broadband Transmission        109

    multiple paths being used to transmit differ-   transmission is modulated into frequency
    ent packets between the same source and         bands, or channels, and is transmitted in
    destination. Higher-end bridges include the
    ability to do load balancing by distributing
                                                    these channels.
                                                       Guard bands, which are small bands of
    traffic over more than one path between a        unused frequencies, are allocated between         c
    source and destination.                         data channels. These provide a buffer
       Recently, wireless bridges have become       against interference due to signals from one
    available for limited-distance remote con-      data channel drifting or leaking over into a      e
    nections. Remote bridges that use radio         neighboring one. The figure “A broadband
    waves can be up to 25 or 30 miles apart—        transmission” shows how data channels and         f
    provided the terrain and weather allow it,      guard bands are used.                             g
    and provided the two bridges have direc-
    tional antennas available. Remote bridges       A B RO A D B A N D T RA N S M I S S I O N         h
    using lasers can be up to about 3,500 feet
    apart. Since focused signals must be sent in
    both cases, such bridges must be within each                                                      j
    other’s line of sight.                                                  1010

       Wireless remote bridges are susceptible to
    two kinds of interference:                                      01010111110100
                                                                   111000110101001                    l
       Inward interference, which can occur                         00010101000101

         when another device is operating in the
         same bandwidth and the two signals                                                           n
         interact with each other.
       Outward interference, in which the device
         under consideration is causing interfer-                                                     p
         ence in a different device.
    B RO A D E R C A T E G O R Y
       Internetwork Link
       Brouter; Gateway; Repeater; Router;                                                            t
       Switch                                          For example, cable TV (CATV) uses
                                                    broadband transmission, with each channel
Broadband Transmission                              getting a 6 megahertz (MHz) bandwidth.            v
                                                    Broadband transmissions use coaxial or
    A broadband transmission is an analog
                                                    fiber-optic cable and they can transmit voice,     w
    communication strategy in which multi-
    ple communication channels are used
                                                    data, or video.                                   x
    simultaneously. The data in a broadband
110       Broadcast

       When digital data is being transmitted, a     w
    modem or other device demodulates the sig-       Brouter
    nals back into digital form at the receiving         A brouter (also known as a bridging router
    end. A modem used for broadband trans-               or, less commonly, as a routing bridge) is a
    missions needs two bands of at least 18              device that combines the features of a bridge
    MHz bandwidth each: one band for sending             and a router. A brouter can work at either
    and the other for receiving.                         the data-link layer or the network layer.
                                                            Working as a bridge, a brouter is protocol
Broadcast                                                independent and can be used to filter local-
                                                         area network traffic. Working as a router, a
    In a network transmission, sending a mes-
                                                         brouter is capable of routing packets across
    sage to all connected nodes. This is in con-
    trast to a transmission that is targeted at a
    single node. Most packet formats have a              B RO A D E R C A T E G O R I E S
    special address value to indicate a packet              Bridge; Internetwork Link; Router
    that is being broadcast. Compare broadcast
    with multicast.
w                                                        A short-term decrease in voltage level, spe-
Broadcast Storm                                          cifically when the voltage is more than 20
    In network traffic, a condition in which              percent below the nominal RMS voltage.
    packets are broadcast, received, and then            Brownouts can occur when a piece of heavy
    broadcast again by one or more of the                machinery is turned on and temporarily
    recipients. The effect of a broadcast storm is       drains the available power, or when every-
    to congest a network with redundant traffic.          one feels the need to run their air condition-
    Broadcast storms can arise, for example, in          ers at the same time.
    bridged networks that contain loops (closed
                                                         SEE ALSO
                                                            Power Disturbance
Broadcast Transmission                               w
    In an AppleTalk network that uses the
                                                         A browser is a hypertext file reader. That is,
    LocalTalk architecture and its LocalTalk
                                                         a browser is a program that can display
    Link Access Protocol (LLAP), a transmission
                                                         material containing links to other material
    sent to each node in the network. Compare
                                                         (perhaps located in other files), and can pro-
    broadcast transmission with directed
                                                         vide quick and easy access to the contents
                                                         associated with such links.
                                       BTAM (Basic Telecommunications Access Method)      111

  Browsers may be text, graphics, or multi-             Forms-capable browsers allow users to
media based:                                         fill in information on forms or question-
  s   A text-based, or line-oriented, browser
                                                     naires. Most graphics-based browsers are
      is unable to display anything but rudi-
      mentary graphics, and is generally line-
                                                        Browsers have long been used in pro-       c
                                                     gramming environments—for example, in
      oriented, but can still switch to any
                                                     the SmallTalk environment created at Xerox
      material that is formatted in a suitable
      manner for the browser. WWW and
                                                     PARC in the 1970s and 1980s. These read-      e
                                                     ers have really come into widespread use
      Lynx are examples of such browsers.
                                                     with the growth of the World Wide Web         f
      Both are accessible on the Internet.
                                                     (WWW) on the Internet.                        g
  s   Graphics browsers can handle both
      text and graphics, require a mouse,
                                                     SEE ALSO
                                                       HotJava; Mosaic; WWW
      and generally have a much nicer dis-
      play than line-oriented browsers. Cello    w
      and Mosaic are examples of graphics-       BSD Socket Layer                                  j
      based browsers.                                In BSD UNIX, the layer that represents
  s   Multimedia browsers can display                the API (Application Program Interface)
      sound and video, in addition to the            between user applications and the network-    l
      capabilities of graphics browsers.             ing subsystem in the operating system
      Mosaic is also a multimedia browser.           kernel.                                       m
      Variants of the mosaic browser are
                                                 w                                                 n
      available for several computing envi-      BSD UNIX (Berkeley Software
      ronments. For example, XMosaic is a        Distribution UNIX)
      browser for the X Window System.
      HotJava, a recently announced
                                                     A UNIX version implemented at the Univer-     p
      browser from Sun Microsystems is
                                                     sity of California, Berkeley. BSD UNIX
                                                     introduced several enhancements to AT&T’s
      generally regarded as taking browser
      technology to a new level. HotJava can
                                                     original implementation, including virtual    r
                                                     memory, networking, and interprocess com-
      handle multimedia material, includes
                                                     munication support.                           s
      security capabilities, and offers an
      object-oriented programming language       w                                                 t
      for creating platform-independent          BTAM (Basic Telecommunications
      applications easily. Because HotJava       Access Method)
      differs so drastically from existing           An early access method for communications     v
      browsers, it remains to be seen how
      quickly— or whether—HotJava
                                                     between IBM mainframes and terminals.
                                                     BTAM is still used, but is largely obsolete
      becomes widely used.                           because it does not support IBM’s SNA (Sys-   x
                                                     tems Network Architecture). ACF/VTAM
112       Btrieve

    has replaced BTAM as the method of choice             system failure. In addition to the stan-
    for remote communications with IBM                    dard ones such as record locking, data
    mainframes.                                           protection measures include logging,
                                                          which records any changes made to
                                                          designated files so that the changes can
                                                          be undone later, if necessary. The roll
    In Novell’s NetWare 3.0 and later, Btrieve is         forward modules mentioned in the
    a key-indexed record management program               next section provide the mechanism
    that allows you to access, update, create,            for such corrections. Data protection
    delete, or save records from a database.              measures also include shadow paging,
    Btrieve is a program (actually several                in which page images are saved before
    programs) that can run in either of two ver-          making any changes on the page.
    sions: client- or server-based.                       Btrieve can back up files even while
       In addition to record-management capa-             they’re in use by using continuous
    bilities, Btrieve includes the following:             operation.
      s    Communications facilities, for             s   Support for NetWare Directory Ser-
          both local and remote communica-                vices (NDS), which are new with Net-
          tions between a program and a record            Ware 4.x. This support is available
          base. The Btrieve Message Routers               only beginning with version 6.1 of
          (that is, BROUTER.NLM and                       Btrieve.
          BDROUTER.NLM) handle outgoing
          requests; BSPXCOM handles incom-            s   Security measures such as the ability to
          ing requests from a remote source (a            encrypt and decrypt data and also the
          workstation or another server).                 ability to assign ownership to files.

      s   Requesters (DOS, OS/2, and so on),          s   Memory management and caching
          which provide Btrieve access for appli-         capabilities to help speed up access and
          cations running on workstations. The            other operations.
          requesters are: BREQUEST.EXE (for            Btrieve creates and maintains a key-
          DOS), BTRCALLS.DLL (for OS/2),            indexed record base (or database). A
          and WBTRCALL.DLL (for Windows).           key-indexed database is one in which keys,
      s   Utilities for setting up, monitoring,     or record fields, are used as the basis for cre-
          and maintaining the record base,          ating an index, which is information that
          among other things. These utilities are   guides access to a database.
          mentioned briefly in the next section.        A Btrieve record base uses a specially
                                                    defined data format, which is also supported
      s   Special data-protection measures for      by database programs and other applica-
          dealing with the record base in case of   tions from third-party vendors.
                                                                                            Btrieve   113

                                                              cate with a Btrieve record base. This pro-
Btrieve-Related Modules
                                                              gram relays requests from the user or from
  The Btrieve programs are provided in Net-
  Ware Loadable Modules (NLMs). The most
                                                              an application to the Record Manager on
                                                              the appropriate server.
  fundamental of these are BTRIEVE.NLM                           Other NLMs handle more specialized           c
  and BSPXCOM.NLM.                                            duties. For example, BROUTER.NLM and
     BTRIEVE contains the Record Manager                      BDROUTER.NLM handle Btrieve-related
  program that does the work on the server.                   requests from a server to a remote server.      e
  This program performs disk I/O (input/out-                  The figure “Relationships of Btrieve ele-
  put) for Btrieve files on the server. This pro-              ments” shows how the various Btrieve            f
  gram must be loaded on any server that has
  Btrieve files.
                                                              elements fit together.                           g
                                                                 Several Btrieve utilities provide the more
     BSPXCOM handles requests to the server                   nitty-gritty services needed to handle the      h
  from any workstation or another remote                      record bases:
  source. BSPXCOM must be loaded on any                                                                       i
                                                                s   BTRMON.NLM monitors Btrieve
  server that needs to communicate with a
  Btrieve requester program on a workstation.                       activity on the server.                   j
     Such a Btrieve requester must be loaded                    s   BSETUP.NLM and BREBUILD.NLM               k
  on any workstation that needs to communi-                         are used to change configurations and
114      BTU (Basic Transmission Unit)

         to update Btrieve data files from ver-            B RO A D E R C A T E G O R Y
         sion 5.x to 6.x, respectively.                      NetWare
     s   BUTIL.NLM imports and exports                w
         Btrieve data, and transfers data             BTU (Basic Transmission Unit)
         between Btrieve files.                            In IBM’s SNA communications, an aggre-
     s   BDIRECT.NLM provides support for                 gate block of one or more path information
         the NDS in NetWare 4.x. This NLM is              units (PIUs) that all have the same destina-
         available only in Btrieve versions 6.1           tion. Several PIUs can be combined into a
         and later.                                       single packet, even if they are not all part of
                                                          the same message. BTUs are created at the
     s   BROLLFWD.EXE (for DOS),                          path-control layer.
         PBROLL.EXE (for OS/2), and
         WBROLL.EXE (for Windows) are                     SEE ALSO

         the roll forward utilities. These are               SNA (Systems Network Architecture)
         used to restore a Btrieve file in case        w
         of some type of system failure.              Buffer, Fiber-Optic Cable
Server- and Client-Based Btrieve                          In fiber-optic cabling, a layer immediately
                                                          surrounding the cladding (which surrounds
  The server-based version runs the Btrieve               the fiber core). The tighter this buffer is
  Record Manager on the server and a special              wrapped around the cladding, the less
  (operating system dependent) requester pro-             opportunity the cladding and core have to
  gram on the workstation. The Record Man-                move around in the cable.
  ager handles the I/O for the database; the
  requester handles the I/O between worksta-              SEE ALSO

  tion and server.                                           Cable, Fiber-Optic
     The client-based version does all its pro-       w
  cessing on the workstation, and makes I/O           Buffer, Memory
  calls (calls involving the record base)
                                                          In memory or storage applications, a buffer
  through the workstation’s operating system.
                                                          is a temporary storage location that is gener-
  The client-based version is available only to
                                                          ally used to hold intermediate values, or
  developers who want to create applications
                                                          other types of data, until they can be pro-
  that can use Btrieve data files.
                                                          cessed. The storage may be allocated in ordi-
     If the calls are for the server’s record base,
                                                          nary RAM (random-access memory), on a
  the Btrieve requester redirects the calls to the
                                                          hard disk, or in special memory registers
  server. The figure “A client and server using
                                                          (such as on a UART chip, which is used for
  Btrieve” shows this situation. Note that the
                                                          serial communications).
  Btrieve requester is provided as part of a
                                                              A print buffer is one common example. A
  server-based Btrieve implementation.
                                                          spooler program saves a file to be printed in
                                                          the print buffer, and deals with the file as
                                                                    Burned-In Address (BIA)   115

A C L I E N T A N D S E R VE R U S I N G B T R I EVE                                                 a
CPU (central processing unit) availability         w
allows. Buffers provide faster access to               Buffered Repeater                             p
stored data.
   Three types of buffer allocations are
                                                        In a network cabling scheme, a device that   q
                                                        can clean and boost signals before sending
distinguished:                                          them on. A buffered repeater can hold a      r
                                                        message temporarily for example, when
   File-cache buffer: Used to store disk-
                                                        there is already a transmission on the
      allocation blocks temporarily.
   Directory-cache buffer: Used to store the
                                                        network.                                     t
     DET (directory-entry table) blocks.                SEE ALSO                                     u
   Packet-receive buffer: Used to hold                                                               v
     incoming packets until they can be            w
     processed.                                        Burned-In Address (BIA)                       w
                                                        SEE                                          x
                                                          BIA (Burned-In Address)
116       Burstiness

w                                                       values they can carry at a time, in their
Burstiness                                              speed, and in their control mechanisms:
    In the CCITT recommendations for B-ISDN,              Bit values: In the PC world, 8-, 16-, and
    a measure of the distribution of data over               32-bit data buses are common. On
    time. The definition for the term has not yet             workstations and larger machines, 64-
    been finalized. One definition being consid-               and 80-bit buses are common.
    ered is the ratio between maximum, or peak,
    and mean (average) bit rate.                          Speed: The speed of a bus depends on the
                                                            system clock. Bus speed is generally
w                                                           measured in megahertz (MHz). The
Burst Mode                                                  IBM-PC bus has gone from a 4.77
    A high-speed transmission mode in which                 MHz clock speed in the original PC to
    the transmitter takes control of the commu-             66 MHz in today’s high-end machines.
    nications channel temporarily, until its                Other chips can support clock speeds
    transmission is complete. This mode is used             of over 100 MHz.
    in internal communications, such as between
                                                          Control: Buses may be controlled through
    hard disk and bus, and also in communica-
                                                            interrupts or through polling.
    tions between devices. The term is also used
    to refer to the packet burst protocol in               In networking, bus refers to a logical and
    NetWare.                                            physical network topology in which mes-
                                                        sages are broadcast along the main cable, so
Burst Speed                                             that all nodes receive each transmission at
                                                        the same time. Standard Ethernet and cer-
    The maximum speed at which a device can             tain ARCnet networks use a bus topology.
    operate without interruption, generally only
    for short periods. This is in contrast to           SEE ALSO
    throughput, which indicates the average               Topology, Bus
    speed at which a device can operate under
    ordinary conditions, such as when transmit-     Bus Interface Board (BIB)
    ting or printing an entire file.
w                                                         BIB (Bus Interface Board)
    In computer hardware, a bus is a path
                                                    Bus Interface Unit (BIU)
    for electrical signals, generally between the
    CPU (central processing unit) and attached          SEE
    hardware. Buses differ in the number of bit           BIU (Bus Interface Unit)
                                                     Byzantine Failure/Byzantine Robustness     117

w                                                    w
Bus Mastering                                        BYTE Information Exchange (BIX)                     a
    In general, bus mastering is a bus-access            SEE                                             b
    method in which a card or device takes con-            BIX (BYTE Information Exchange)
    trol of the bus in order to send data onto the                                                       c
    bus directly, without help from the CPU
    (central processing unit). In a network,
                                                     Byte-Sex                                            d
    the network interface card takes control             For a processor, byte-sex is a feature that     e
    of the bus.                                          describes the order in which bytes are repre-
       Generally, MCA (Microchannel Architec-            sented in a word. Processors may be little-     f
    ture) and EISA (Extended Industry Standard           endian, big-endian, or bytesexual.              g
    Architecture) machines support bus master-              In little-endian representations, the low-
    ing, but ISA (Industry Standard Architec-            order byte in a word is stored at the lower     h
                                                         address. In big-endian processors or con-
    ture) machines do not. VL (VESA local) and
    PCI (Peripheral Component Interconnect)              texts, the high-order byte is stored first.
    buses also support bus mastering.                    Bytesexual is a term used to describe a pro-    j
       Bus mastering can improve throughput              cess that is capable of using either little-
    considerably, but only if the board and the          endian or big-endian representations for        k
    computer support the same bus-mastering              information, depending on the value of a
                                                         flag bit.
    method, and if the bus mastering doesn’t
    conflict with the hard-disk controller.
                                                         SEE ALSO
       Several types of transfer modes are possi-
    ble with bus mastering, including burst
                                                           Big-Endian; Little-Endian; Middle-Endian      n
    mode, streaming data mode, and data              w                                                   o
    duplexing. A particular bus-mastering            Byzantine Failure/Byzantine
    scheme may support some or all of these          Robustness                                          p
    modes.                                               In networking, a situation in which a node      q
                                                         fails by behaving incorrectly or improperly,
Bypass                                                   rather than by breaking down completely         r
    In telephony, a connection with an inter-
                                                         and disappearing from the network. A net-       s
                                                         work that can keep working even if one or
    exchange carrier (IXC) that does not go
    through a local exchange carrier.
                                                         more nodes is experiencing Byzantine failure    t
                                                         has Byzantine robustness.
Byte                                                                                                     v
    A collection of—usually eight—bits (but
    rarely worth a dollar anymore). A byte
    generally represents a character or digit.                                                           x
    120     Cable

                                                        drop cable. The different types of connectors

    Cable                                               are discussed in a separate article.
     It took about 100 years for cable to replace
     the kite string as a medium for electrical       IBM Token Ring
     power, but the change was heartily wel-            IBM Token Ring networks distinguish
     comed, particularly by researchers. Cables         between the main ring path and patch
     are currently the most popular medium for          cables. In this context, patch cables attach
     transmitting information between nodes in          nodes (called lobes in Token Ring networks)
     a network, although wireless transmission          to wiring centers. The wiring centers are
     schemes (radio, infrared, and microwave            called multistation attachment units (MAUs)
     communications) are becoming more widely           in such networks. The patch cables can also
     used.                                              attach to patch panels, which are, in turn,
Network Cabling Schemes                                 connected to MAUs.

     In a network, the cabling scheme connects        Cable Types
     nodes (or stations) and also gives the net-        Four main types of cable are used in
     work its characteristic shape (topology) and       networks:
     features. Network cabling schemes distin-
     guish between main and auxiliary cables.              s   Coaxial cable, also called coax, which
     The main cable provides the path and                      can be thin or thick.
     defines the shape for the network; the auxil-          s   Twisted-pair cable, which can be
     iary cables connect nodes to the main path                shielded (STP) or unshielded (UTP).
     or to wiring centers that are connected to
     the main path. Depending on the architec-             s   IBM cable, which is essentially
     ture, the terminology for such cables differs.            twisted-pair cable, but designed to
                                                               somewhat more stringent specifica-
Ethernet Trunk and Drop Cables                                 tions by IBM. Several types are
                                                               defined, and they are used primarily
     For Ethernet networks, the main cable is                  in IBM Token Ring networks.
     referred to as the trunk cable, and the auxil-
     iary cables are called drop cables. Trunk             s   Fiber-optic cable, which can be single-
     cable forms the backbone, or main cabling                 mode, multimode, or graded-index
     scheme, of an Ethernet network. Because of                multimode.
     its role and location, trunk cable is some-           Coaxial, IBM, and twisted-pair cables
     times called backbone cable. Drop cable            transmit electricity. Fiber-optic cables trans-
     may be used to attach an individual node to        mit light signals. Each of the cable types is
     a network trunk cable. Nodes can also be           subdivided into more specialized categories
     connected to the cable indirectly through a        and has its own design and specifications,
     connector or transceiver rather than with          standards, advantages, and disadvantages.
                                                                                    Cable    121

     Cable types differ in price, transmission      between the insulation and the jacket.
  speed, and recommended transmission dis-          Twisted-pair cable has two conductor wires
  tance. For example, twisted-pair wiring is
  currently the cheapest (and also the most
                                                    twisted around each other. Fiber-optic cable
                                                    may include material to help protect the
  limited in performance). Fiber-optic cable is     fiber from pressure.                               c
  more expensive but much faster and more
  robust. Coaxial cable lies between these two    Conductor                                           d
  types on most performance and price               For electrical cable, the conductor is known      e
  features.                                         as the signal, or carrier, wire, and it may
     This article discusses network cabling in      consist of either solid or stranded wire. Solid
  general. The specific cable types (coaxial,        wire is a single thick strand of conductive       g
  twisted-pair, IBM, and fiber-optic) are cov-       material, usually copper. Stranded wire con-
  ered in more detail in separate articles. In      sists of many thin strands of conductive          h
  addition to this cabling, there is a cable
  infrastructure behind the walls, in shafts,
                                                    material wound tightly together.                  i
                                                       Signal wire is described in the following
  and under the ground. These cables are dis-       terms:                                            j
  cussed under the headings Cable, Horizontal
  and Cable, Backbone.                                s   The wire’s conductive material (for         k
                                                          example, copper)
Cable Components
                                                      s   Whether the wire is stranded or solid
  The different cable types have the following        s   The carrier wire’s diameter, expressed
  components in common:                                   directly (for example, in inches, centi-    n
                                                          meters, or millimeters), or in terms of
    s   A conductor to provide a medium for
                                                          the wire’s gauge, as specified in the
        the signal. The conductor might be a
        copper wire or a glass tube.
                                                          AWG (American Wire Gauge) tables            p
                                                          (see the AWG article for a summary of
    s   Insulation of some sort around the                gauges)                                     q
        conductor to help keep the signal in
        and interference out.
                                                       The total diameter of the strand deter-        r
    s   An outer sheath, or jacket, to
                                                    mines some of the wire’s electrical proper-
                                                    ties, such as resistance and impedance.
        encase the cable elements. The jacket       These properties, in turn, help determine the     t
        keeps the cable components together,        wire’s performance.
        and may also help protect the cable            For fiber-optic cable, the conductor is         u
        components from water, pressure, or         known as the core. The core is a glass or         v
        other types of damage.                      plastic tube that runs through the cable. The
    In addition to these common features,
                                                    diameter of this core is expressed in microns     w
                                                    (millionths of a meter).
  particular types of cable have other compo-                                                         x
  nents. Coaxial cable has one or more shields
 122     Cable

Insulation Layer                                       Code’s communications plenum cable)
                                                       or CL2P (class 2 plenum cable) specifica-
   The insulating layer keeps the transmission
                                                       tions. The cable should also be UL-listed for
   medium’s signal from escaping and also
                                                       UL-910, which subjects plenum cable to a
   helps to protect the signal from outside
                                                       flammability test. The NEC and UL specifi-
   interference. For electrical wires, the insula-
                                                       cations are discussed in the Cable Standards
   tion is usually made of a dielectric (noncon-
   ductor), such as polyethylene. Some types of
   coaxial cable have multiple protective layers
                                                     Nonplenum Cable Jacket
   around the signal wire.
      For fiber-optic cable, the insulation is          Nonplenum cable uses less-expensive mate-
   known as cladding and is made of material           rial for jackets, so it is considerably less
   with a lower refraction index than the core’s       expensive than cable with plenum jackets,
   material. The refraction index is a measure         but it can be used only under restricted con-
   that indicates the manner in which a mate-          ditions. Nonplenum cable jackets are made
   rial will reflect light rays. The lower refrac-      of polyethylene (PE) or polyvinylchloride
   tion index ensures that light bounces back          (PVC), which will burn and give off toxic
   off the cladding and remains in the core.           fumes.
                                                          PVC cable used for networks should meet
Plenum Cable Jacket                                    the NEC’s CMR (communications riser
                                                       cable) or CL2R (class 2 riser cable) specifica-
   The outer casing, or jacket, of the cable pro-
                                                       tions. The cable should also be UL-listed for
   vides a shell that keeps the cable’s elements
                                                       UL-1666, which subjects riser cable to a
   together. Two main classes of jacket are ple-
                                                       flammability test. See the Cable Standards
   num and nonplenum. For certain environ-
                                                       article for a discussion of cable safety stan-
   ments, plenum cable is required by law. It
                                                       dards and performance levels.
   must be used when the cable is being run
   “naked” (without being put in a conduit)          Cable Packaging
   inside walls, and should probably be used
   whenever possible.                                  Cable can be packaged in different ways,
      Plenum jackets are made of nonflam-               depending on what it is being used for and
   mable fluoropolymers (such as Teflon or               where it is located. For example, the IBM
   Kynar). They are fire-resistant and do not           cable topology specifies a flat cable for use
   give off toxic fumes when burning. They are         under carpets. Some fiber-optic trunks con-
   also considerably more expensive (by a fac-         tain thousands of fibers, each of which can
   tor of 1.5 to 3) than cables with nonplenum         carry multiple messages.
   jackets. Studies have shown that cables with           The following types of cable packaging
   plenum jackets have less signal loss than           are available:
   nonplenum cables.                                      Simplex cable: One cable within one
      Plenum cable used for networks should                  jacket, which is the default configu-
   meet the NEC’s CMP (National Electric                     ration. The term is used mainly for
                                                                                               Cable      123

       fiber-optic cable to indicate that the
       jacket contains only a single fiber.
                                                      Cable Properties
                                                          Cable is described in terms of the size and
  Duplex cable: Two cables, or fibers,                     makeup of its components, as well as in
    within a single jacket. In fiber-optic
    cable, this is a common arrangement.
                                                          terms of its performance. For example, elec-             c
                                                          trical cable specifications include the gauge,
    One fiber is used to transmit in each                  or diameter, of the signal wire.                         d
    direction.                                               The cable’s electrical and physical proper-           e
  Multifiber cable: Multiple cables, or                    ties determine the performance you can
   fibers, within a single jacket. For fiber-               expect and the range of conditions under                 f
                                                          which you can use the cable. Cables differ in
   optic cable, a single jacket may contain
                                                          the electrical properties (signal loss, imped-
   thousands of fibers; for electrical cable,
   the jacket will contain at most a few                  ance, and so on) they offer. The table “Cable            h
   dozen cables.                                          Properties” lists some of the features that
                                                          distinguish cables.                                      i
                        MEASUREMENT                                                                                k
P R O P E RT Y          OR DESCRIPTION             COMMENT

Conductor               Millimeters (mm), inches   For stranded wire, this represents the total diameter of the
wire diameter           (in), or gauge (AWG)       entire cluster of strands.                                      n
Core fiber diameter      Microns                    Some core diameters have desirable properties in terms of
                                                   the paths certain wavelengths of light take in the core. For
                                                   example, diameters of 62.5 and 100 microns for multimode        p
                                                   fiber and of under 10 microns for single-mode fiber are
                                                   common.                                                         q
Wire insulation         Millimeters or inches      The diameter of the cable’s insulaton layer is needed to        r
diameter                                           calculate certain electrical properties of a cable.

Cladding diameter       Microns                    The cladding diameter varies much less than the core
                                                   diameter, partly because the cladding helps to make the         t
                                                   fiber easier to package if the cladding is of an approximately
                                                   constant size.                                                  u
Wire shield             Millimeters, inches,                                                                       v
diameter                or gauge
Jacket diameter         Millimeters or inches      The diameter of the jacket can be important when installing
                                                   the cable because it may determine space requirements.          x
124     Cable

 P R O P E RT Y       OR DESCRIPTION             COMMENT


 Conductor wire       Materials; solid vs.       Conductor wires may be solid or stranded, or of different
 composition          stranded (# of strands)    types of conductive material (usually copper alone or in
                                                 some variant). If the wire is stranded, the specifications
                                                 should note the number of strands.

 Wire insulation      Materials

 Shield composition   Materials; % area cov-     For coaxial cable only, shield composition refers to the
                      ered by shield mesh        makeup of the protective shield around the conductive

 Jacket composition   Materials; plenum vs.


 DCR (DC              Ohms (Ω) per distance      Refers to the DC resistance for the conductor wire.
 Resistance)          (100 or 1000 feet)

 Shield DCR           Ohms (Ω) per distance      Refers to the DC resistance for the shield.
                      (100 or 1000 feet)

 Impedance            Ohms                       The measure of a wire’s resistance to electrical current,
                                                 which helps determine the wire’s attenuation properties.
                                                 Most networks use cable with a characteristic impedance
                                                 level. There are devices for connecting cable segments that
                                                 have diffferent impedances.

 Capacitance          Picofarads per foot        The measure of the cable’s ability to store up electrical
                      (pF/ft)                    charge or voltage. This charge storage distorts a signal as
                                                 it travels along its course; the lower the capacitance
                                                 the better.

 Attenuation          Maximum decibels           The measure of the signal loss over distance. Data sheets
                      per distance at a given    may include several attenuation values for different frequen-
                      frequency; common dis-     cies. This distinction can be imporant because attenuation
                      tances include 100 feet,   of an electrical signal increases with signal frequency.
                      1000 feet, and 1 kilome-
                      ter, e.g., dB/1000 ft at
                      5 MHz
                                                                                              Cable      125

                      OR DESCRIPTION             COMMENT
 Crosstalk (NEXT)     Minimum decibels per       NEXT (near-end crosstalk) is a common measure of inter-        b
                      distance (1000 or 100
                      feet) (dB/distance)
                                                 ference by a signal from a neighboring cable or circuit. The
                                                 higher the decibel value, the less crosstalk.
 Velocity of          % (values should be        Specifies the maximum signal speed along the wire, as a         d
 Propagation          about 60%; preferably
                      above 80%)
                                                 proportion of the theoretical maximum (the speed of light).    e
 Other Properties                                                                                               f
 Weight               Unit weight per distance                                                                  g
                      (oz/ft; gm/meter)
 Maximum recom-       Distance (feet, meters,
 mended cable         or kilometers)                                                                            i
 segment range
 Bandwidth            Megahertz (MHz) or
                      megabits per second                                                                       k
 Price                Dollars per distance
                      (100 or 1000 feet)
 Performance/Safety   NEC CL2, CMP, and          See the Cable Standards article for information about these    n
 Ratings              CMR; EIA/TIA-568
                      Categories 1-5; UL
                                                 cable safety standards.
                      Levels 1-5; ETL ratings                                                                   p
     You can obtain the specifications for a              clear as at the beginning. Unfortunately, this         r
  specific type of cable from the cable manu-             will not be true.
  facturer or vendor. The table “Cable Com-                 Any transmission consists of signal and             s
  ponent Abbreviations” lists some common                noise components. Even a digital signal
  abbreviations used in cable specifications or           degrades when transmitted over a wire or
  data sheets.                                           through an open medium. This is because                u
                                                         the binary information must be converted to
Factors Affecting Cable Performance
                                                         electrical form for transmission, and because          v
  Cables are good media for signals, but they            the shape of the electrical signal changes             w
  are not perfect. Ideally, the signal at the end        over distance.
  of a stretch of cable should be as loud and                                                                   x
 126      Cable

  A B B R E V I AT I O N   F E AT U R E                                   COMPONENT(S)

  AD                       Air dielectric                                 Insulation

  AL                       Aluminum braid                                 Shield

  ALS                      Aluminum sheath                                Shield

  AWG                      American Wire Gauge (AWG) value for wire       Carrier wire

  BC                       Bare copper braid                              Carrier wire; shield

  CCAL                     Copper-clad aluminum                           Carrier wire

  CCS                      Copper-covered steel                           Carrier wire

  FEP                      Fluorinated ethylene propylene (Teflon)         Insulation; jacket

  FFEP                     Foamed fluorinated ethylene propylene (Teflon)   Insulation

  FP                       Foamed polyethylene                            Insulation

  K                        Kynar/polyvinylidene fluoride (plenum)          Jacket

  PE                       Polyethylene (solid)                           Insulation; jacket

  PVC                      Polyvinylchloride                              Jacket

  PVDF                     Generic polyvinylidene fluoride (plenum)        Jacket

  SC                       Silvered copper braid                          Carrier wire; shield

  TC                       Tinned copper braid                            Carrier wire; shield

  x%                       Percentage of surface area covered by braid    Shield

  #cond                    Number of conductors                           Carrier wire

    Signal quality degrades for several rea-               occurs more quickly at higher frequencies
  sons, including attenuation, crosstalk, and              and when the cable’s resistance is higher.
  impedance.                                                  In networking environments, repeaters
                                                           are responsible for cleaning and boosting a
Attenuation                                                signal before passing it on. Many devices are
                                                           repeaters without explicitly saying so. For
  Attenuation is the decrease in signal
                                                           example, each node in a token ring network
  strength, measured in decibels (dB) per 100
                                                           acts as a repeater. Since attenuation is sensi-
  feet or per kilometer. Such loss happens as
                                                           tive to frequency, some situations require the
  the signal travels over the wire. Attenuation
                                                                                        Cable   127

  use of equalizers to boost different                  made on a cable depend on the location
  frequency signals the appropriate amount.             in which the cable is used and the function
                                                        for which the cable is intended. These
                                                        demands, in turn, determine the features
                                                        a cable should have.                             c
  Crosstalk is interference in the form of a sig-
  nal from a neighboring cable or circuit; for        Function and Location
  example, signals on different pairs of twisted
  wire in a twisted-pair cable may interfere            Here are a few examples of considerations
  with each other. A commonly used measure              involving the cable’s function and location:     f
  of this interference in twisted-pair cable
  is near-end crosstalk (NEXT), which is
                                                           s   Cable designed to run over long           g
                                                               distances, such as between floors or
  represented in decibels. The higher the deci-
  bel value, the less crosstalk and the better
                                                               buildings, should be robust against       h
                                                               environmental factors (moisture, tem-
  the cable.
                                                               perature changes, and so on). This may
     Additional shielding between the carrier
  wire and the outside world is the most com-
                                                               require extra jackets or jackets made     j
                                                               with a special material. Fiber-optic
  mon way to decrease the effects of crosstalk.
                                                               cable performs well, even over dis-       k
                                                               tances much longer than a floor or a
  Impedance, which is a measure of electrical              s   Cable that must run around corners
  resistance, is not directly a factor in a cable’s
  performance. However, impedance can
                                                               should bend easily, and the cable’s       n
                                                               properties and performance should not
  become a factor if it has different levels
                                                               be affected by the bending. For several   o
  at different locations in a network. In order
  to minimize the disruptive effects of differ-
                                                               reasons, twisted-pair cable is probably   p
                                                               the best cable for such a situation
  ent impedances in a network, special de-
                                                               (assuming it makes sense within the       q
  vices, called baluns, are used to equalize
                                                               rest of the wiring scheme). Of course,
  impedance at the connection (at the balun
                                                               another way to get around a corner is
     Impedance does reflect performance indi-
                                                               by using a connector; however, con-       s
                                                               nectors may introduce signal-loss
  rectly, however. In general, the higher the
                                                               problems.                                 t
  impedance, the higher the resistance, and
  the higher the resistance, the greater the               s   Cable that must run through areas in      u
                                                               which powerful engines or motors are
  attenuation at higher frequencies.
                                                               operating (or worse, being turned on
Selecting Cable                                                and off at random intervals) must be      w
                                                               able to withstand magnetic interfer-
  Cables are used to meet all sorts of power
  and signaling requirements. The demands
                                                               ence. Large equipment gives off strong    x
                                                               magnetic fields, which can interfere       y
128   Cable

      with and disrupt nearby signals. In                  the environments and the transition
      commercial and residential settings,                 between them.
      this can be a problem with cable that is
      run, for example, through the elevator      Main Cable Selection Factors
      shaft. Because it is not affected by such
                                                    Along with the function and location con-
      electrical or magnetic fluctuations,
                                                    siderations, cable selections are determined
      fiber-optic cable is the best choice in
                                                    by a combination of factors, including the
      machinery-intensive environments.
  s   If you need to run lots of cables
                                                       s   The type of network you plan to create
      through a limited area, cable weight
                                                           (Ethernet, Token Ring, or another
      can become a factor, particularly if all
                                                           type). While it is possible to use just
      that cable will be running in the ceiling
                                                           about any type of cable in any type of
      above you. In general, fiber-optic and
                                                           network, certain cable types have been
      twisted-pair cable tend to be lightest.
                                                           more closely associated with particular
  s   Cables being installed in barely acces-              network types. For example, Token
      sible locations must be particularly                 Ring networks use twisted-pair cable.
      reliable, and they should probably be
                                                       s   The amount of money you have avail-
      laid with backup cable during the ini-
                                                           able for the network. Keep in mind
      tial installation. Some consultants and
                                                           that cable installation can be an expen-
      mavens advise laying a second cable
                                                           sive part of the network costs.
      whenever you are installing cable, on
      the assumption that the installation is          s   Whatever cabling resources are already
      much more expensive than the cable                   available (and usable). You will almost
      and that installation costs for the sec-             certainly have available wiring that
      ond cable add only marginally to the                 could conceivably be used for a net-
      total cost. Generally, the suggestion is             work. It is almost equally certain,
      to make at least the second cable opti-              however, that at least some of that
      cal fiber.                                            wire is defective or is not up to the
                                                           requirements for your network.
  s   Cables that need to interface with
      other worlds (for example, with a                s   Building or other safety codes and
      mainframe network or a different                     regulations.
      electrical or optical system) may need
      special properties or adapters. For         Connected versus Bulk Cable
      example, UTP cable in a Token Ring
      network needs a media filter between           You can get cable with or without connec-
      the cable and the MAU to which the            tors at either end. Both connected and bulk
      cable is attached. The kinds of cable         cable have advantages and drawbacks.
      required will depend on the details of        Whether connected or bulk cable is better
                                                    depends on how you are going to use it.
                                                                                    Cable    129

     You have much more flexibility to cut or
  reroute with bulk cable, because you are not
                                                   Cabling Tools
                                                     Installation tools for handling cables include
  restricted to a precut cable segment. On the
  other hand, you (or someone you trust) will        wire strippers, dies, and crimping tools for
  need to attach the connectors. This requires       attaching connectors to the end of a stretch     c
  special tools and involves stripping the end       of bulk cable. Such tools are often included
  of the cable and crimping the connector to         in adapter kits, which are configured for         d
  the bare wire.                                     building particular types of cable (for          e
                                                     example, coaxial cable or cable for RS-232
Cable Prices                                         connections). Depending on how compre-           f
                                                     hensive the toolkit is, expect to pay any-
  Cable prices depend on factors such as the
                                                     where from about $30 to $500.
                                                         Testing tools for cables include a whole     h
    s   Type of cable (coaxial, twisted-pair,        range of line scanners and monitors. The
        fiber optic). In general, fiber-optic          simplest of these can tell you whether there     i
        cable is the most expensive but the
        price is dropping rapidly. Fiber-optic
                                                     is any electrical activity between one loca-
                                                     tion in a network (or a cable installation)
        cable is followed closely by thick coax-     and another. The most sophisticated can do       k
        ial cable. STP and thin coaxial follow       just about everything except tell you where
        in roughly that order, but with consid-      you bought the cable.                            l
        erable overlap in prices. UTP is the             The top-of-the-line scanners can test any    m
        least expensive type of cable.               kind of copper-based cable not only for
    s   Whether cable comes in bulk or with
                                                     faults, but also for performance specifica-       n
                                                     tions (NEXT, attenuation, and so on). These
        connectors at either end. While price is
                                                     types of scanners know about the electrical
        an issue, this question will be answered
        mainly by your needs for the cable.
                                                     requirements of the most popular network         p
                                                     architectures (such as Ethernet/802.3 and
    s   Whether the cable is plenum or non-          Token Ring) and are capable of finding            q
        plenum. Plenum versions can cost             faults or deviations from specifications at       r
        from 1.5 to 3 times as much as the           just about any location on the network.
        nonplenum version.                           Of course, you will pay several thousand         s
                                                     dollars for this capability.
     Cable prices change, so do not be sur-                                                           t
  prised to find considerable variation in
  prices when you start getting quotes.
                                                   Cable Vendors and Resources
                                                     Many companies sell both electrical and
     UTP cable is grouped into voice- and
                                                     fiber-optic cable, as well as connectors,
  data-grade. Most telephone wire is just
  voice-grade. Prices for data-grade UTP cable       installation, and testing tools. Some vendors    w
  are a few cents higher per foot.                   specialize in fiber-optic products, others in
                                                     copper-based products, and still others offer    x
                                                     both.                                            y
130      Cable


 Here are some tips on purchasing and installing cabling:

 s   Cables have quite a few properties that should be considered in making decisions.You can find information
     about these cable properties in cable specifications or data sheets, which are available from cable vendors.

 s   In general, cable that meets military specifications (MIL-SPECS) is designed to more stringent require-
     ments, and so is a good choice for networks. This is even more true for connectors, because the military
     specifications insist on durable and reliable connectors. (Connectors are particularly prone to shoddy

 s   Fiber-optic connectors are especially tricky to attach because fiber optics has such exact alignment
     requirements. It’s probably worth your while to let a professional attach these connectors.

 s   When you’re ordering cable, make sure it’s clear whether you want cable with connectors or “raw” (bulk)

 s   Make sure the cable is good quality. Otherwise, you’ll have trouble after a while, as the insulation within
     and outside the cable breaks down.

 s   Test cable both before and after installing it.

 s   While present needs are obviously the major determinant of cabling decisions, future plans should also be
     taken into consideration. In general, at least consider installing cable one level more powerful than you
     think you’ll need.

 s   When adding cable to an existing cabling system, find out exactly what kind of cable is already in place.The
     safest thing is to get the actual part and specification information from the cable jacket, then order exactly
     that from the same distributor (or a certified equivalent from a different manufacturer).

 s   Before adding to existing cable, test it as thoroughly as possible. If the cable seems likely to have a major
     breakdown within a few months, it’s almost certainly better to replace it now.

 s   Protect the cable as much as possible. Such measures should include protecting the cable from tempera-
     ture or moisture changes, which can cause the cable to crack or melt.

 s   Support the cable as much as possible, so that a hanging cable doesn’t stretch because the cable’s own
     weight is pulling it downward.

 s   Velcro cable ties can help make things neater, by enabling you to collect multiple loose wires into a single
     cluster. The Rip-Tie Company in San Francisco is one vendor that offers these neatness aids.
                                                                        Cable, Backbone    131

   When you are ready to start looking for      w
cabling and other components, it will be        Cable, Adapter                                      a
worthwhile getting the cabling guides and
catalogs from several vendors. The guides
                                                    Cable used to connect a Token Ring              b
                                                    network interface card (NIC) to a hub
offer useful general-purpose hints and guide-       or multistation access unit (MAU). IBM          c
lines for selecting and installing cable.
   Here are some cable vendors and their
                                                    Type 1 and Type 6 cable can be used for this
                                                    purpose. The IBM cables have a DB-9 or
telephone numbers:                                  DB-25 connector at the NIC end and an           e
                                                    IBM data connector at the MAU end.
  AMP Incorporated: (800) 522-6752;                                                                 f
   (717) 564-0100                               w
                                                Cable, Backbone                                     g
  Andrew Corporation: (800) 328-2696;
    Fax (708) 349-5673                              Backbone cable refers to the cable that         h
                                                    forms the main trunk, or backbone, of a
  Berk-Tek: (800) 237-5835
                                                    network, particularly an Ethernet network.      i
  Black Box Corporation: (800) 552-6816;            Individual nodes and other devices may be       j
     (412) 746-5500                                 connected to this cable using special adapt-

  Comm/Scope: (800) 982-1708;
                                                    ers (such as transceivers) and a separate       k
                                                    stretch of cable (called the drop cable in an
    (704) 324-2200; Fax (704) 459-5099
                                                    Ethernet network) to the node.
  CSP (Computer System Products):                      More generally, backbone cable is defined     m
    (800) 422-2537; (612) 476-6866;                 by the EIA/TIA-568 committee as any
    Fax (612) 476-6966                              “behind the scenes” cable—cable running         n
  FIS (Fiber Instrument Sales):
                                                    behind walls, in shafts, or under the
                                                    ground—that is not classified as horizontal
     (800) 445-2901; (315) 736-2206;
     Fax (315) 736-2285
                                                    cable. (Horizontal cable is defined by the       p
                                                    EIA/TIA-568 committee as any cable that
  Jensen Tools: (800) 426-1194;                     goes from a wiring closet, or distribution
     (602) 968-6231; Fax (800) 366-9662             frame, to the wall outlet in the work area.)    r
                                                    This includes cable used to connect wiring
  Trompeter Electronics: (800) 982-2639;
                                                    closets and equipment rooms.                    s
    (818) 707-2020; Fax (818) 706-1040
                                                       The EIA/TIA-568 recognizes four
SEE ALSO                                            main types of backbone cable, and several
  Cable, Backbone; Cable, Coaxial; Cable,           optional variants. These types are listed in    u
  Fiber-Optic; Cable, Horizontal; Cable,            the table “EIA/TIA-568 Main and Optional
  IBM; Connector; Connector, Fiber-Optic            Types of Backbone Cable.”                       v
132        Cable, Broadcast-Oriented

    C A B L E TY P E   MAIN                                             OPTIONAL

    UTP                100-ohm, multipair UTP cable, to be used for
                       voice-grade communications only

    STP                150-ohm STP cable, such as that defined in the    100-ohm STP cable
                       IBM Cable System (ICS)

    Coaxial            50-ohm thick coaxial cable, such as the cable    75-ohm (broadband) coaxial cable,
                       used in thick Ethernet networks                  such as CATV cable

    Optical fiber       62.5/125-micron (step- or graded-index)          Single-mode optical fiber
                       multimode optical fiber

    C O M P A RE                                         w
       Cable, Horizontal                                  Cable, CATV (Community Antenna
                                                          Television, or Cable Television)
       Cable                                                 Wiring used for the transmission of cable
                                                             television signals. CATV is broadband
w                                                            coaxial cable and is generally wired for
Cable, Broadcast-Oriented                                    one-directional transmission; that is, from
    Cable that is designed to carry video signals            the cable station, or a head-end, to the
    sent from one location in the network,                   consumer. If the CATV cable is not one-
    known as the head-end. This type of cable is             directional, it may be possible to use it
    generally designed for one-way communica-                for network cabling.
    tion, which makes it of limited value for use
    as network cable.                                     Cable, Coaxial
w                                                            Coaxial cable, often called coax, is used for
Cable, Category x                                            data transmissions. This cable’s remarkably
    A five-level rating system for telecommuni-               stable electrical properties at frequencies
    cations wiring, specified in the EIA/TIA-568              below 1 GHz (gigahertz) makes the cable
    documents. These describe minimum perfor-                popular for cable television (CATV) trans-
    mance capabilities for unshielded twisted-               missions and for creating local-area net-
    pair cable.                                              works (LANs). Telephone company
                                                             switching offices also use coaxial cable to
                                                             route long-distance calls. The figure “Con-
       Cable Standards                                       text and properties of coaxial cable” sum-
                                                             marizes the features of this type of cable.
                                                                                               Cable, Coaxial    133

 Cable                                                                                                                    c
       Twisted-Pair                                                                                                       d
                                                      Coaxial Properties
   Stable and predictable electrical properties
   At least one shield around conductor wire                                                                              i
   Subject to electromagnetic interference
   Variable impedance levels                                                                                              j
   Thin and thick varieties
   Broadband and baseband varieties
   Thin coaxial uses BNC/TNC connectors; thick coaxial uses N-series connectors                                           l
   Twinaxial runs two cables within a single jacket
   Triaxial and quadrax have extra shielding for special uses
                                                        Coaxial Uses
   Ethernet networks                                                                                                      o
   ARCnet networks
   Cable TV lines
   Video cable                                                                                                            q
   IBM mainframe and midrange-based networks (twinaxial)
   Telephone switching offices                                                                                            r
Coaxial Cable Components                                                      There are restrictions regarding the        t
  A coaxial cable consists of the following
                                                                              wire composition for certain network        u
                                                                              configurations. The diameter of the
  layers (moving outward from the center):                                    signal wire is one factor in determining    v
                                                                              the attenuation (loss) of the signal over
     Carrier wire: A conductor wire (the car-
                                                                              distance. The number of strands in a
       rier, or signal, wire) is in the center.
       This wire is made of (or contains)                                     multistrand conductor also affects the      x
       copper and may be solid or stranded.                                                                               y
134    Cable, Coaxial

   Insulation: An insulation layer consists         A COAXIAL CABLE
      of a dielectric (nonconductor) around         H A S F I VE L AYE R S
      the carrier wire. This dielectric is usu-
      ally made of some form of polyethyl-
      ene or Teflon.
   Foil shield: A thin foil shield around the
     dielectric. This shield usually consists
     of aluminum bonded to both sides of a
     tape. Not all coaxial cables have foil
     shielding; some have two foil shield
     layers, interspersed with braid shield
   Braid shield: A braid, or mesh, conductor,
     made of copper or aluminum, that sur-
     rounds the insulation and foil shield.
     This conductor can serve as the ground
     for the carrier wire. Together with the
     insulation and any foil shield, the braid
     shield protects the carrier wire from
     electromagnetic interference (EMI) and
     radio frequency interference (RFI). The
     braid and foil shields provide good
     protection against electrical interfer-
     ence, but only moderate protection
     against magnetic interference.
   Jacket: An outer cover that can be either
      plenum (made of Teflon or Kynar) or
      nonplenum (made of polyethylene             Coaxial Cable Performance
      or polyvinylchloride).                        The main features that affect the perfor-
    The figure “A coaxial cable has five lay-         mance of coaxial cable are its composition,
 ers” shows the makeup of a coaxial cable.          width, and impedance.
 The layers surrounding the carrier wire also          The carrier wire’s composition determines
 help prevent signal loss due to radiation          how good a conductor the cable will be.
 from the carrier wire. The signal and shield       Copper is among the best materials for this
 wires are concentric, or co-axial, hence the       purpose. The IEEE specifies stranded copper
 name.                                              carrier wire with tin coating for thin coaxial,
                                                    and solid copper carrier wire for thick
                                                                            Cable, Coaxial   135

     Cable width helps determine the electrical        For coaxial cable, the following types of
  demands that can be made on the cable. In          connectors are available:
  general, thick coaxial can support a much
  higher level of electrical activity than thin
                                                        s   A BNC (bayonet nut connector) is          b
                                                            used for thin coaxial cable.              c
     Impedance is a measure of opposition                   The N-series connectors are used for
  to the flow of alternating current. The prop-

                                                            thick coaxial cable.
  erties of the dielectric between the carrier
                                                        s   A TNC (threaded nut connector) may
  wire and the braid help determine the cable’s
  impedance. Each type of network archi-
                                                            be used in the same situations as a       f
                                                            BNC, provided that the other connec-
  tecture uses cable with a characteristic
                                                            tor is also using TNC.                    g
     Impedance helps determine the cable’s              Connectors for coaxial cable should be        h
  electrical properties and also limits the con-     plated with silver, not tin. This improves the
  texts in which the cable can be used. For          contact and the durability of the connector.
  example, Ethernet and ARCnet architectures
                                                   Thin versus Thick Coaxial
  can both use thin coaxial cable, but they
  have different impedances; therefore, Ether-       Descriptively, coaxial cable is grouped
  net and ARCnet cables are not compatible.
  In networks, the impedances range from
                                                     mainly into thin and thick varieties. Thin       l
                                                     coaxial cable is 3/16-inch in diameter and
  50 ohms (for an Ethernet architecture) to          is used for various network architectures,       m
  93 ohms (for an ARCnet architecture).              including thin Ethernet (also known as           n
Coaxial Cable Connectors                             10Base2 or CheaperNet) and ARCnet.
                                                     When using this configuration, drop cables        o
  A segment of coaxial cable has an end              are not allowed. Instead, the T-connector
  connector at each end. The cable is                must be connected directly to the network
  attached through these end connectors              interface card (NIC). This means the NIC         q
  to a T-connector, a barrel connector, another      must have an on-board transceiver, known
  end connector, or to a terminator. Through         as a medium attachment unit (MAU) in the         r
  these connectors, another cable or a hard-
  ware device is attached to the coaxial cable.
                                                     IEEE 802.3 standard.
                                                         Thick coaxial cable is 3/8-inch in diame-
     In addition to their function, connectors       ter. It is used for thick Ethernet (also known   t
  differ in their attachment mechanism and           as 10Base5 or ThickNet) networks, cable
  components. For example, BNC connectors            TV (CATV), and other connections. Thick
  join two components by plugging them               coaxial is expensive and is notoriously diffi-    v
  together and then turning the components to        cult to install and work with. It is more
  click the connection into place. Different         likely to be inherited than selected for use     w
  size coaxial cable requires a different-sized
                                                     in a network.
136     Cable, Coaxial

Cable Content Descriptions                               interference can occur; for example, in
                                                         industrial settings.
  Other descriptions of coaxial cable are
  based on the contents of the cable, rather      Baseband versus Broadband Cable
  than its size, as follows:
                                                    Functionally, coaxial cable is grouped into
    Twinaxial: Also known simply as twinax,         baseband and broadband varieties.
      this coaxial cable has two carrier               Baseband coaxial cable has one channel
      wires, each with its own dielectric, or       over which a single digital message can be
      insulation, layer. The wires are gener-       sent, at speeds of up to 80 megabits per
      ally twisted around each other, which         second (Mbps). Thin coaxial is used for
      helps reduce magnetic interference,           baseband cable.
      and are surrounded by a shield and a             Broadband coaxial cable can carry sev-
      jacket whose properties run the same          eral analog signals (at different frequencies)
      gamut as for ordinary coaxial cable.          simultaneously. Each of these signals can be
      This type of cable is used in IBM and         a different message or a different type of
      AppleTalk networks. For example,              information. Thick coaxial cable can be
      twinaxial cable is used to connect IBM        used for broadband transmissions in a
      5250 terminals to System/36 or AS/            network.
      400 computers.                                   Broadband coaxial can use a single
                                                    cable or multiple cables. In single-cable
    Triaxial: Also known simply as triax, this
                                                    broadband coaxial, frequencies are split;
       coaxial cable has extra shielding: an
                                                    for example, into 6 megahertz (MHz) chan-
       inner braid surrounded by an inner
                                                    nels for each station. Some channels are
       (nonplenum) jacket, surrounded by
                                                    allocated for bidirectional communication.
       an outer copper braid. This outer braid
                                                    Dual-cable broadband coaxial uses one
       is, in turn, surrounded by the outer
                                                    cable for sending and one for receiving data;
       jacket. The extra shielding makes a big
                                                    each cable has multiple channels.
       difference because of the grounding
                                                       Note that broadband coaxial requires
       and improved protection.
                                                    much more planning than baseband coaxial.
    Quadrax: This cable is a hybrid of triax-       For example, a broadband setup will prob-
      ial and twinaxial cable. Quadrax has          ably need amplifiers for dealing with the
      the extra carrier wire with dielectric,       different broadband signals.
      and also has the extra shielding of
      triaxial.                                   Coaxial Cable Designations

    Quad shield: This cable has four layers of      The following designations are used for
      shielding: alternating layers of foil and     coaxial cable used in networks. These are
      braid shields. Quad shield cable is used      just a few of the available coaxial cable
      in situations where heavy electrical          types.
                                                                       Cable, Coaxial    137

RG-6: Used as a drop cable for CATV
  transmissions. It has 75 ohms imped-
                                             Advantages of Coaxial Cable
                                               Coaxial cable has the following advantages
  ance, is a broadband cable, and is
  often quad-shielded.                         over other types of cable that might be
RG-8: Used for thick Ethernet. It has 50
                                               used for a network. The advantages are gen-      c
                                               eral and may not apply in a particular situa-
  ohms impedance. The thick Ethernet           tion. Note also that advantages change or        d
  configuration requires other cable and
  a MAU (transceiver). The other cable
                                               disappear over time, as technology advances      e
                                               and products improve.
  required is a twisted-pair drop cable to
                                                 s   Broadband coaxial can be used to
  the NIC. The drop cable off RG-8
  cable uses a 15-pin DIX (or AUI) con-              transmit voice, data, and even video.      g
  nector. RG-8 is also known as N-Series
  Ethernet cable.
                                                 s   The cable is relatively easy to install.   h
RG-11: Used for the main CATV trunk.
                                                 s   Coaxial cable is reasonably priced         i
                                                     compared with other cable types.
  It has 75 ohms impedance and is a                                                             j
  broadband cable. This cable is often       Disadvantages of Coaxial Cable
  quad shielded (with foil/braid/foil/                                                          k
                                               Coaxial cable has the following disadvan-
  braid around the signal wire and
  dielectric) to protect the signal wire
                                               tages when used for a network:                   l
  under even the worst operating                 s   It is easily damaged and sometimes         m
  conditions.                                        difficult to work with, especially in
                                                     the case of thick coaxial.
RG-58: Used for thin Ethernet. It has
  50 ohms impedance and uses a BNC               s   Coaxial is more difficult to work with      o
  connector.                                         than twisted-pair cable.                   p
RG-59: Used for ARCnet. It has 75 ohms
  impedance and uses BNC connectors.
                                                 s   This type of cable cannot be used with
                                                     token ring network architectures.
  This type of cable is used for broad-
                                                 s   Thick coaxial can be expensive to
  band connections and also by cable
  companies to connect the cable net-
                                                     install, especially if it needs to be      s
                                                     pulled through existing cable conduits.
  work to an individual household.
                                                 s   Connectors can be expensive.
RG-62: Used for ARCnet. It has 93 ohms
  impedance and uses BNC connectors.             s   Baseband coaxial cannot carry inte-
  This cable is also used to connect ter-            grated voice, data, and video signals.     v
  minals to terminal controllers in IBM’s
  3270 system configurations.                                                                    w
138      Cable, Data-Grade

                                                             of the tools used for cable testing.) For more
  U S I N G E X I S T I N G C OA X I A L C A B L E
                                                             specialized tasks requiring tools, such as
  It may be tempting to try to use existing coaxial          crimpers and dies for attaching connectors
  cable—which is likely to be CATV cable—for                 to cable, you will need versions specifically
  a network. If you’re considering this, here’s an           designed for coaxial cable.
  important point to keep in mind: Not all CATV                 When in doubt, of course, ask the vendor
  cables are the same.                                       explicitly whether a particular tool will
                                                             work with coaxial cable.
  Broadcast-oriented cables are designed to carry
  video signals sent from one location in the net-           SEE ALSO
  work, known as the head-end. Such cables are                 Cable; Cable, Fiber-Optic; Cable,
  designed for one-way communication, which                    Twisted-Pair; Connector
  makes them useless for data networks. Even if
  a bidirectional CATV cable is available, several
                                                         Cable, Data-Grade
  other considerations must be taken into account
  before you can use this cable for a local-area             Twisted-pair cable of sufficiently high qual-
  network.                                                   ity to use for data transmission. In contrast,
                                                             voice-grade cable is more susceptible to
  If the cable will still be used to transmit TV chan-       interference and signal distortion. In the
  nels, you need to find two frequency bands that             EIA/TIA-568 cable specifications, categories
  won’t be used for TV channels. Each of these               2 through 5 are data-grade cable.
  bands must have at least 18 MHz band width. The
  bands are used by a modem, which modulates                 SEE ALSO
  network data into the appropriate frequency                  Cable, Twisted-Pair
  band at one end. A second modem demodulates
  this signal at the other end. The TV and data net-     Cable, Distribution
  works will be independent of each other.
                                                             In broadband networks, a term for cable
  Because your network may be grafted onto an                used over intermediate distances (up to a
  existing CATV topology, you need to make sure              few hundred yards) and for branches off a
  your system can deal with this. Typically, a CATV          network trunk, or backbone. RG-11 cable
  network uses a tree topology. The head-end is              is commonly used for this purpose.
  the root, and the signal is transmitted along suc-
  cessive branches. For this setup, you need to          w

  make sure that limitations on cable length are not
                                                         Cable, Drop
  exceeded.                                                  Cable used to connect a network interface
                                                             card (NIC) to a transceiver on a thick Ether-
                                                             net network. Drop cable, also known as AUI
Tools for Working with Coaxial Cable
                                                             cable or transceiver cable, has a 15-pin AUI,
  Almost all cable testers can deal with coaxial             or DIX, connector at the NIC end and an
  cable. (See the Cable article for a discussion             N-series connector at the transceiver end.
                                                             This term may also be applied loosely to
                                                                                                    Cable, Fiber-Optic             139

    other cables that connect a network node                              light rather than electricity. Cables of this
    to a wiring center of some sort.                                      type differ in their physical dimensions and
                                                                          composition and in the wavelength(s) of
                                                                          light with which the cable transmits. The fig-
                                                                          ure “Context and properties of fiber-optic                      c
                                                                          cable” summarizes the features of this type
Cable, Feeder                                                             of cable.
    A 25-pair cable that can be used for carrying                            Because fiber-optic communication uses
    both voice and data signals. This cable can
    run from equipment to distribution frame.
                                                                          light signals, transmissions are not subject                   f
                                                                          to electromagnetic interference. This, and
                                                                          the fact that a light signal encounters little                 g
Cable, Fiber-Optic
    C O N T EX T A N D P RO P E R T I E S O F F I B E R - O P T I C C A B L E                                                            h
    Fiber-optic cable, also known as optical
    fiber, provides
    Context        a medium for signals using
      Electrical                                                                                                                         k
          Coaxial                                                                                                                        l
                                                    Fiber-Optic Properties
                                                    Fiber-Optic Properties                                                               o
      Medium for light signals                                                                                                           p
      Light at certain wavelengths is best for signaling purposes
      Comes in single-mode (thin fiber core; single light path) and multi-mode (thick fiber core; multiple light paths) versions         q
      Multimode can be step-index or graded-index
      Cable is very lightweight
      Very high bandwidth                                                                                                                s
      Immune to electromagnetic inteference, eavesdropping
      Very long cable segments possible                                                                                                  t
                                                        Fiber-Optic Uses                                                                 u
      FDDI networks
      Long-haul lines
      To connect network segments or networks                                                                                            w
      To connect mainframes to peripherals
      To connect high-speed, high-performance workstations
140     Cable, Fiber-Optic

  resistance on its path (relative to an electri-      out strong interference when the elevator is
  cal signal traveling along a copper wire),           running.
  means that fiber-optic cable can be used for             One reason fiber-optic networks are slow
  much longer distances before the signal must         to catch on is price. Network interface cards
  be cleaned and boosted.                              (NICs) for fiber-optic nodes can cost several
     Some fiber-optic segments can be several           thousand dollars, compared to street prices
  kilometers long before a repeater is needed.         of about $100 for some Ethernet and ARC-
  In fact, scientists have sent signals over fiber-     net cards. However, when selecting optical
  optic lines for thousands of kilometers with-        fiber, it is not always necessary to use the
  out any signal boosters. In 1990, researchers        most expensive fiber-optic connections. For
  sent a 1 gigabit per second (Gbps) signal            short distances and slower bandwidths,
  almost 8,000 kilometers (about 5,000 miles)          inexpensive cable is just fine. In general, a
  without a boost!                                     fiber-optic cable will always allow a longer
     In principle, data transmission using fiber        transmission than a copper cable segment.
  optics is many times faster than with electri-
  cal methods. Speeds of over 10 Gbps are            Fiber-Optic Cable Components
  possible with fiber-optic cable. In practice,         The major components of a fiber-optic cable
  however, this advantage is still more promise        are the core, cladding, buffer, strength mem-
  than reality, because the cable is waiting for       bers, and jacket. Some types of fiber-optic
  the transmission and reception technology            cable even include a conductive copper wire.
  to catch up.                                         This can be used to provide power; for
     Nevertheless, fiber-optic connections              example, to a repeater. The figure “Compo-
  deliver more reliable transmissions over             nents of a fiber-optic cable” illustrates the
  greater distances, although at a some-               makeup of this type of cable.
  what greater cost. Fiber-optic cables cover
  a considerable price and performance range.        Fiber-Optic Core and Cladding
Uses of Fiber-Optic Cable                              The core of fiber-optic cable consists of one
                                                       or more glass or plastic fibers through which
  Currently, fiber-optic cable is used less often
                                                       the light signal moves. Plastic is easier to
  to create a network than to connect two net-
                                                       manufacture and use but works over shorter
  works or network segments. For example,
                                                       distances than glass. The core can be any-
  cable that must run between floors is often
                                                       where from about 2 to several hundred
  fiber-optic cable, most commonly of the
                                                       microns. (A micron, also known as a
  62.5/125 variety with an LED (light-
                                                       micrometer, is a millionth of a meter, or
  emitting diode) as the light source.
                                                       about 1/25,000 of an inch.)
     Being impervious to electromagnetic
  interference, fiber is ideal for such uses               In networking contexts, the most popular
  because the cable is often run through the           core sizes are 60, 62.5, and 100 microns.
  elevator shaft, and the elevator motor puts          Most of the fiber-optic cable used in net-
                                                       working has two core fibers: one for com-
                                                       municating in each direction.
                                                                   Cable, Fiber-Optic   141

  COMPONENTS OF A FIBER-OPTIC CABLE                                                            a
     The core and cladding are actually manu-    The buffer of a fiber-optic cable is one       s
  factured as a single unit. The cladding is a   or more layers of plastic surrounding the
  protective layer (usually of plastic) with     cladding. The buffer helps strengthen the     t
  a lower index of refraction than the core.     cable, thereby decreasing the likelihood of
  The lower index means that light that hits     micro-cracks, which can eventually grow
  the core walls will be redirected back to      into larger breaks in the cable. The buffer   v
  continue on its path. The cladding will be     also protects the core and cladding from
  anywhere between a hundred microns and a       potential corrosion by water or other mate-   w
  millimeter (1000 microns) or so.               rials in the operating environment. The
                                                 buffer can double the diameter of some
Fiber-Optic Buffer                               cable.                                        y
 142     Cable, Fiber-Optic

     A buffer can be loose or tight. A loose        Single-Mode Cable
  buffer is a rigid tube of plastic with one or
  more fibers (consisting of core and cladding)        In single-mode fiber-optic cable, the core is
  running through it. The tube takes on all the       so narrow (generally less than 10 microns)
  stresses applied to the cable, buffering the        that the light can take only a single path
  fiber from these stresses. A tight buffer fits        through it. Single-mode fiber has the least
  snugly around the fiber(s). A tight buffer           signal attenuation, usually less than 2 deci-
  can protect the fibers from stress due to            bels (dB) per kilometer. This type of cable
  pressure and impact, but not from changes           is the most difficult to install, because it
  in temperature.                                     requires the greatest precision, and it is the
                                                      most expensive of the major fiber-optic
Strength Members                                      types. However, transmission speeds of 50
                                                      Gbps and higher are possible. To get a sense
  Fiber-optic cable also has strength members,        of this magnitude, note that a 10 Gbps line
  which are strands of very tough material            can carry 130,000 voice channels.
  (such as steel, fiberglass, or Kevlar) that pro-         Even though the core of single-mode
  vide extra strength for the cable. Each of the      cable is shrunk to very small sizes, the clad-
  substances has advantages and drawbacks.            ding is not reduced accordingly, nor should
  For example, steel attracts lightning, which        it be. For single-mode fiber, the cladding
  will not disrupt an optical signal but may          diameter should be about ten times the core
  seriously disrupt the people or machines            diameter. This ratio makes it possible to
  sending or receiving such a signal.                 make the cladding the same size as for popu-
                                                      lar multimode fiber-optic cable. This helps
Fiber-Optic Jacket                                    create a de facto size standard. Keeping
  The jacket of a fiber-optic cable is an outer        the cladding large also makes the fiber and
  casing that can be plenum or nonplenum, as          cable easier to handle and more resistant to
  with electrical cable. In cable used for net-       damage.
  working, the jacket usually houses at least
                                                    Multimode Cable
  two fiber/cladding pairs: one for each
  direction.                                          Multimode fiber-optic cable has a wider
                                                      core, so that a beam of light has room to
Single-Mode versus Multimode Cable
                                                      follow multiple paths through the core.
  Fiber-optic cable can be either single-mode         Multiple modes (light paths) in a transmis-
  or multimode. (Modes are the possible paths         sion produce signal distortion at the receiv-
  for the light through a cable.)                     ing end.
                                                         One measure of signal distortion is modal
                                                      dispersion, which is represented in nanosec-
                                                      onds (billionths of a second) of tail per kilo-
                                                      meter (ns/km). This value represents the
                                                      difference in arrival time between the fastest
                                                                         Cable, Fiber-Optic   143

  and slowest of the alternate light paths. The
  value also imposes an upper limit on the
                                                    Graded-Index Cable                                 a
  bandwidth, since the duration of a signal
  must be larger than the nanoseconds of a tail
                                                      Cable with a gradual change in refraction        b
                                                      index is called graded-index cable, or
  value. With step-index fiber, expect between         graded-index multimode. This fiber-optic          c
  15 and 30 ns/km. Note that a modal disper-          cable type has a relatively wide core, like
  sion of 20 ns/km yields a bandwidth of less         single-step multimode cable. The change
  than 50 Mbps.                                       occurs gradually and involves several layers,    e
                                                      each with a slightly lower index of refrac-
Gradation of Refraction: Step-
                                                      tion. A gradation of refraction indexes con-     f
Index Cable versus Graded-Index Cable
                                                      trols the light signal better than the step-     g
  One reason optical fiber makes such a good           index method. As a result, the attenuation is
  transmission medium is because the different        lower, usually less than 15 dB/km. Similarly,    h
                                                      the modal dispersion can be 1 ns/km and
  indexes of refraction for the cladding and
                                                      lower, which allows more than ten times the
  core help to contain the light signal within
  the core. Cable can be constructed by chang-        bandwidth of step-index cable. Graded-           j
  ing abruptly from the core refractive index         index multimode cable is the most com-
  to that of the cladding, or this change can be      monly used type for network wiring.              k
  made gradually. The two major types of
                                                    Fiber Composition                                  l
  multimode fiber differ in this feature.
                                                      Fiber core and cladding may be made of           m
Step-Index Cable                                      plastic or glass. The following list summa-      n
                                                      rizes the composition combinations, going
  Cable with an abrupt change in refraction
  index is called step-index cable. In step-
                                                      from highest quality to lowest:                  o
  index cable, the change is made in a single            Single-mode glass: Has a narrow core,         p
  step. Single-step multimode cable uses this               so only one signal can travel through.
  method, and it is the simplest, least expen-
                                                         Graded-index glass: Not tight enough to
  sive type of fiber-optic cable. It is also the
                                                           be single-mode, but the gradual change      r
  easiest to install. The core is usually between
  50 and 125 microns in diameter; the clad-
                                                           in refractive index helps give more
                                                           control over the light signal.
  ding is at least 140 microns.
     The core width gives light quite a bit of           Step-index glass: The abrupt change from
  room to bounce around in, and the attenua-                the refractive index of the core to that   u
  tion is high (at least for fiber-optic cable):             of the cladding means the signal is less
  between 10 and 50 dB/km. Transmission                     controllable.                              v
  speeds between 200 Mbps and 3 Gbps are
                                                         Plastic-coated silica (PCS): Has a rela-      w
  possible, but actual speeds are much lower.
                                                            tively wide core (200 microns) and a
                                                            relatively low bandwidth (20 MHz).
144       Cable, Fiber-Optic

      Plastic: This should be used only for very                8/125: A single-mode cable with an 8
         short distances.                                         micron core and a 125 micron clad-
                                                                  ding. This type of cable is expensive
     To summarize, fiber-optic cables may
                                                                  and currently used only in contexts
  consist of glass core and glass cladding (the
                                                                  where extremely large bandwidths are
  best available). Glass yields much higher
                                                                  needed (such as in some real-time
  performance, in the form of higher band-
                                                                  applications) or where large distances
  width over greater distances. Single-mode
                                                                  are involved. An 8/125 cable configu-
  glass with a small core is the highest quality.
                                                                  ration is likely to broadcast at a light
  Cables may also consist of glass core and
                                                                  wavelength of 1,300 or 1,550 nm.
  plastic cladding. Finally, the lowest grade
  fiber composition is plastic core and plastic                  62.5/125: The most popular fiber-
  cladding. Step-index plastic is at the bottom                   optic cable configuration, used in most
  of the heap in performance.                                     network applications. Both 850 and
                                                                  1,300 nm wavelengths can be used
  F I B E R - O P T I C C A B L E Q UA L I T Y                    with this type of cable.
  Here are a few points about fiber-optic cable                  100/140: The configuration that IBM
  (other things being equal):                                     first specified for fiber-optic wiring in
                                                                  a Token Ring network. Because of the
  s   The smaller the core, the better the signal.
                                                                  tremendous popularity of the 62.5/125
  s   Fiber made of glass is better than fiber made                configuration, IBM now supports both
      of plastic.                                                 configurations.
  s   The purer and cleaner the light, the better the           Make sure you buy fiber-optic cable
      signal. (Pure, clean light is a single color, with     with the correct core size. If you know what
      minimal spread around the color’s primary              kind of network you plan to build, you may
      wavelength.)                                           be constrained to a particular core size.
                                                             IBM usually specifies a core of 100 microns
  s   Certain wavelengths of light behave better
                                                             for Token Ring networks; other networks
      than others.
                                                             more commonly use cable with a 62.5
                                                             micron core.
Fiber-Optic Cable Designations
                                                           Components of a Fiber-Optic Connection
  Fiber-optic cables are specified in terms
  of their core and cladding diameters. For                  In addition to the cable itself, a fiber-optic
  example, a 62.5/125 cable has a core with a                connection needs a light source to generate
  62.5 micron diameter and cladding with                     the signal, as well as connectors, repeaters,
  twice that diameter.                                       and couplers to route and deliver the signal.
     The following are some commonly used                    The figure “Components of a fiber-optic
  fiber-optic cable configurations:                            connection” illustrates how this works.

                                                                   Cable, Fiber-Optic   145

COMPONENTS OF A FIBER-OPTIC CONNECTION                                                           a
Fiber-optic transmitters convert an elec-       generally have lower output power, but also
tronic signal into light and send this light    less signal attenuation and higher band-         p
signal into the fiber core. The transmitter’s    width. Output power values should not be
light source and output optical power are       too high, since this increases energy require-   q
crucial elements in determining the transmit-
ter’s performance.
                                                ments and also risks frying the components
                                                at the receiving end.
   The transmitter’s output power depends          Transmitters use either digital or            s
on several things, including the fiber and       analog modulation. Analog modulation
cladding sizes and the fiber’s numerical aper-   is used for voice, video, and even radar         t
ture (NA). The NA is a measure of the fiber’s    signals, which require bandwidths ranging        u
ability to gather light and is determined by    from tens of kilohertz to hundreds of mega-
the angle over which light hitting the fiber     hertz, and even as high as a gigahertz.          v
will move through it.
   Output power values range from less than
                                                Digital modulation is used in computer net-
                                                works and in long-haul telephone systems,
50 to over 200 microwatts. Smaller cores                                                         x
 146     Cable, Fiber-Optic

  which require transmission speeds ranging                 lasers (possible at the two highest
  from tens of kilobits to more than a gigabit              wavelengths) make it possible to emit
  per second. Transmitters differ in speed. Not             at a particular wavelength with mini-
  surprisingly, the faster ones are also more               mal spectral width.
                                                        s   A good relationship between the emit-
Light Source                                                ting area and acceptance angle. The
                                                            emitting area is the opening through
  The light source will be a laser or a light-              which the transmitter emits its light.
  emitting diode (LED). A good light source                 This should be small in relation to the
  in a fiber-optic connection should have the                fiber core’s acceptance angle, so that
  following characteristics:                                all the light emitted by the transmitter
     s   Fast rise and fall times. The rise time is         will enter the core. Not surprisingly,
         the time required for a light source to            lasers have a much smaller emitting
         go from 10 to 90 percent of the desired            area than LEDs.
         level. This time limits the maximum            s   Steady, strong output power. The
         transmission rate, so it should be as              higher the output power, the stronger
         short as possible. Lasers have a rise              the signal and the further it can travel
         time of less than a nanosecond; the                without becoming too weak. Laser
         rise time for LEDs ranges from a few               output can be as much as 1000 times
         nanoseconds to a few hundred                       that of LEDs.
                                                        s   A long lifetime. The lifetime of a light
     s   A narrow spectral width. The spectral              source is the amount of time before the
         width refers to the range of wave-                 source’s peak output power is half its
         lengths emitted by the light source, and           original level. This is generally in the
         it should be as narrow as possible.                millions of hours (longer than ours)
         Spectral widths for lasers are 1 to 3              and is typically longer for LEDs than
         nm; for LEDs, they are from 30 to                  for lasers!
         50 nm.
                                                         Although lasers are clearly the light
     s   Light emission at a central wavelength       source of choice, LEDs are generally the
         with minimal spectral width. The cen-        light source of record. The most likely rea-
         tral wavelength is the primary wave-         son for this is price; transmitters that use
         length of the light being emitted. For       LEDs are usually much less expensive. This
         various reasons, wavelengths of 820,         is not a problem for networking purposes,
         850, 1300, and 1550 nm have all been         however, because LEDs operating at 820 or
         used. LEDs are used for the first three       850 nm are fine for the short-distance, fiber-
         of the wavelengths, but rarely for 1550      optic connections currently most popular.
         nm. Lasers can be used at all of these       Despite their performance shortcomings
         wavelengths, and single-frequency
                                                                           Cable, Fiber-Optic    147

  compared with lasers, LEDs are more                   as 10 microwatts, or as a microwatt level
  reliable and less prone to breakdowns.                needed for a given bit error rate (BER).
                                                        Duty Cycle A duty cycle specifies the ratio        b
                                                        of high to low signal values in a digital
                                                        transmission. This is not necessarily equal to
  Fiber-optic receivers undo the work of trans-
  mitters: they accept a light signal and con-          the proportion of 0 and 1 bit values in the       d
  vert this to an electrical signal representing        message, because some signal-encoding
  information in analog or digital form. A              methods will encode a 1 as high at one point
  receiver’s performance depends on how well            in a transmission and as low in another           f
  its three main components work. The fol-              point. (See the Encoding, Signal article for
  lowing are the main components of a fiber-             examples of such methods.) The ideal duty         g
                                                        cycle is 50 percent.
  optic receiver:
                                                            The duty-cycle value is important because
           The photodetector, which “sees” the

           optical signal and converts it into elec-
                                                        receivers use a reference level as the thresh-    i
                                                        old between high and low values. Some
           trical form. This produces a current
                                                        receivers adjust this reference during a trans-   j
           that is proportional to the level of light
                                                        mission. If a duty-cycle value deviates from
                                                        the 50 percent ideal, the altered threshold
     s     The amplifier, which boosts the sig-          level could lead to more erroneous values.        l
           nal and gets it into a form ready for        For example, if a threshold is adjusted
           processing.                                  downward because of a 20 percent duty
     s     The processor, which tries to repro-
                                                        cycle, low signals that are marginally but not    n
                                                        significantly higher than normal may be mis-
           duce the original signal.
                                                        interpreted as high values. There are two         o
     The receiver also includes interfaces for          strategies for getting around the potential
                                                        error problem: signal encoding and reference
  the cable carrying the light signal and the
  device to which the electrical signal is being        levels.                                           q
  passed.                                                   Certain signal-encoding methods, such as
     The photodetector and amplifier pro-                the Manchester and differential Manchester
  cesses are essentially identical for analog and       methods used in Ethernet and Token Ring           s
  digital signals. The main differences are in          networks, always have a 50 percent duty
  the processor.                                        cycle. The tradeoff for this nice behavior        t
                                                        is that these encoding methods require a
     There are several classes of photodetec-
  tors, each suitable for different speed and           clock that runs at twice the data rate (since
  distance configurations. The receiver sensi-           every interval is associated with two electri-    v
  tivity specifies the weakest signal that the           cal levels).
  photodetector can detect. This information                It is possible to build a receiver that has   w
  may be expressed as an absolute value, such           an absolute reference level; that is, one that    x
                                                        will always correspond to the level of a
 148       Cable, Fiber-Optic

  50 percent duty cycle. This is accomplished         s   Low power loss. There should be mini-
  by coupling the receiver to a DC power sup-             mal loss of signal power going across
  ply. The tradeoff for this is that the receiver         the connection or splice. For networks
  has higher power requirements; it requires a            and short-distance connections, the
  signal that is 6 to 8 dB (roughly, four to              loss should be less than 1 dB; for long-
  eight times) stronger than for an ordinary              haul connections, there should be less
  receiver.                                               than 0.2 dB loss.
                                                      s   Durability. The connector should be
                                                          capable of multiple matings (connec-
  A fiber-optic transceiver includes both a                tions) without loosening or becoming
  transmitter and a receiver in the same com-             unreliable. Durability values typically
  ponent. These are arranged in parallel so               range between about 250 and 1000
  that they can operate independently of each             matings.
  other. Both the receiver and the transmitter        s   Ease of use. The connector or splice
  have their own circuitry, so that the com-
                                                          should be easy to install.
  ponent can handle transmissions in both
  directions.                                         s   Low price. The less expensive, the bet-
                                                          ter, provided all the preceding features
Repeater                                                  are satisfactory.

  Like a transceiver, a fiber-optic repeater            There are many types of connector
  includes both a transmitter and a receiver        designs used for fiber-optic cable. Some of
  in the same component. However, in the            the most commonly used ones in networking
  repeater, these components are arranged in        are SC, ST, SMA, and the MIC connector
  series, separated by circuitry for cleaning       specified for the FDDI (Fiber Distributed
  and boosting the signal. The receiver gets the    Data Interface) network architecture. See the
  signal and passes it through the booster to       Connectors, Fiber-Optic article for more
  the transmitter.                                  information about fiber-optic connectors.
                                                       If a fiber-optic connection is more or less
Connectors and Splices                              permanent, it may make more sense to splice
                                                    the cable segments together. Splicing tech-
  Connectors serve to link two segments of          niques are more reliable and precise than
  cable or a cable and a device. A connector is     connectors. Because of this, signal loss at
  used for temporary links. To link two sec-        splices is much lower (almost always less
  tions of cable permanently, use a splice; to      than 1 dB, and often less than 0.25 dB) than
  link more than two sections of cable, use a       at connectors. Splicing is almost always used
  coupler. In general, use a splice when possi-     for long-haul, fiber-optic cable.
  ble; use a connector when necessary.                 Fusion and mechanical splices are the two
     A good connector or splice should have         most common splicing methods. Of the two,
  the following properties:                         fusion gives the better splices.
                                                                         Cable, Fiber-Optic    149

     A fusion splice welds the two fibers              an optical switch, which allows the light
  together using a high-precision instrument.         to bypass a node and continue around the
  This type of splice produces losses smaller
  than 0.1 dB. The equipment for such splic-
                                                      network.                                          b
  ing is quite expensive, however.                  Fiber-Optic Cable Signal Loss                       c
     A mechanical splice is accomplished by
  fitting a special device over the two fibers to
                                                      As mentioned earlier, light signals can be        d
                                                      diminished by coupling. In addition, factors
  connect them and lock them into place. The          that contribute to signal loss across a stretch   e
  device remains attached to the splice area          of cable include the following:
  to protect the splice from environmental                                                              f
                                                         Pulse dispersion: If the cable’s core width
  effects, such as moisture or pressure.
                                                           is large compared with the light’s
  Mechanical splices have higher signal losses
  than fusion splices, but these losses may still          wavelength, light enters the core at dif-    h
  be less than 0.25 dB.                                    ferent angles and will travel different
                                                           distances to the destination. As             i
Couplers                                                   explained earlier, the difference in
                                                           arrival times between the fastest and
  Fiber-optic couplers route an incoming sig-              slowest signals in a group is measured       k
  nal to two or more outgoing paths. Coup-                 in nanoseconds of tail over the dis-
  lers are needed in fiber-optic networks.                  tance the light must travel. This value      l
  When an electrical signal is split and sent
  along parallel paths, each derived signal is
                                                           limits the maximum transmission rate,        m
                                                           because signal pulses must be sepa-
  the same strength. This is not the case with             rated by at least the nanoseconds of         n
  light signals.                                           tail time. For example, if a signal
     After the signal is split, the derived opti-          acquires 10 nanoseconds of tail over
  cal signals are each weaker than the original            the required distance, the maximum           p
  signal. For example, if a fiber-optic coupler             transmission rate is 100 Mbps.
  splits a signal into two equal signals, each of                                                       q
  those derived signals loses 3 dB relative to           Attenuation: Loss of signal strength that
  the original signal, just from the signal halv-          occurs because some of the light is          r
  ing. Couplers can be designed to split a sig-            absorbed by the cladding, and some
                                                           light is scattered as a result of imper-
  nal equally or unequally. See the Coupler,
  Fiber-Optic article for more information.                fections in the fiber.                        t
Optical Switches
                                                         Fiber bending: Signal loss can occur           u
                                                            because the fiber is bent in particular
  Couplers used in networks need some type                  ways. Multiple bands of light (known        v
  of bypass mechanism, so that the coupler                  as modes) enter a core, each at slightly
                                                            different angles. Bending the fiber can
  can be disconnected if the coupler’s target
  nodes are not on the network. This discon-                enable certain modes to escape from         x
                                                            the core. Since the modes that escape
  nection capability is accomplished with                                                               y
150     Cable, Fiber-Optic

        will not be random, fiber bending can            s   Fiber-optic lines are much harder to
        introduce systematic loss of certain sig-           tap, so they are more secure for private
        nal components. Simply rolling fiber                 lines.
        cable onto a spool for distribution can
                                                        s   Light has a much higher bandwidth, or
        introduce fiber bending. Cable manu-
                                                            maximum data-transfer rate, than elec-
        facturers design their cable spools
                                                            trical connections. (This speed advan-
        carefully, and some even publish speci-
                                                            tage has yet to be realized in practice,
        fications for the spool.
    Microbending: Microbends are tiny kinks
                                                        s   The signal has a much lower loss rate,
      that can arise in the cable as a result of
                                                            so it can be transmitted much further
      various stresses (for example, attach-
                                                            than it could be with coaxial or
      ing a connector at the end of a cable).
                                                            twisted-pair cable before boosting is
      Microbends in the fiber can cumulate,
      and the presence of many kinks can
      significantly increase the signal loss             s   Optical fiber is much safer, because
      from bending.                                         there is no electricity and so no danger
                                                            of electrical shock or other electrical
    Fiber ovality: If the fiber’s core and clad-
       ding are not round, the nonuniform
       shape will distort the signal. This can          s   Fiber-optic cable is generally much
       happen, for example, if the cable was                thinner and lighter than electrical
       squashed with a heavy weight, so that                cable, and so it can be installed more
       the core and cladding are partially                  unobtrusively. (Fiber-optic cable
       flattened.                                            weighs about an ounce per meter;
                                                            coaxial cable weighs nearly ten times
Advantages of Fiber-Optic Cable                             that much.)
  Fiber-optic connections offer the following           s   Cable making and installation are
  advantages over other types of cabling                    much easier than they were in the early
  systems:                                                  days.
    s   Light signals are impervious to             Disadvantages of Fiber-Optic Cable
        interference from EMI or electrical
        crosstalk. Light signals do not interfere     The disadvantages of fiber-optic connections
        with other signals. As a result, fiber-        include the following:
        optic connections can be used in
                                                        s   Fiber-optic cable is currently more
        extremely adverse environments, such
                                                            expensive than other types of cable.
        as in elevator shafts or assembly
        plants, where powerful motors and               s   Other components, particularly NICs,
        engines produce lots of electrical noise.           are very expensive.
                                                                     Cable, Fiber-Optic    151

     s   Certain components, particularly cou-          Such a toolkit will include cable strip-
         plers, are subject to optical crosstalk.       pers, scissors, crimping tools, epoxy,
     s   Fiber connectors are not designed to
                                                        pliers, canned air (for cleaning fibers
                                                        after polishing), inspection micro-
         be used as often as you would like.
         Generally, they are designed for fewer
                                                        scope, polishing materials, and so on.      c
         than a thousand matings. After that,       s   Optical power meter, which is a device      d
         the connection may become loose,               that can read levels of optical signals
         unstable, or misaligned. The resulting         on a fiber-optic line. Using sensors
         signal loss may be unacceptably high.          attached to the cable, this device can      f
                                                        report absolute or relative signal levels
     s   Many more parts can break in a
                                                        over a range of 110 dB (which means         g
         fiber-optic connection than in an
                                                        that the weakest and strongest detect-
         electrical one.
                                                        able signals differ by a factor of over
Fiber-Optic Cable Tools                                 10 billion). An optical power meter         i
                                                        can also be used to measure light at
  It is only fitting that the most complex               specific wavelengths.                        j
  wiring technology should also have the
  most sophisticated tools. Optical fiber
                                                    s   An OTDR (optical time domain reflec-         k
                                                        tometer), which is a device that can
  undergoes an extensive set of tests and
                                                        measure the behavior of the light sig-
  quality-control inspections before it
  even leaves the manufacturer.                         nals over time and create graphical         m
                                                        representations of these measurements.
      The manufacturers’ tests are designed to
  get complete details about the cable’s physi-         An OTDR can be used to measure              n
  cal and optical properties. Optical properties        signal loss along a stretch of cable and    o
  include attenuation, dispersion, and refrac-          to help locate a fault in a fiber-optic
  tive indexes of the core and cladding layers.         connection.                                 p
  Physical properties include core and clad-        s   Splicer, which is used to create splices,   q
  ding dimensions, numerical aperture and               or permanent connections in an optical
  emitting areas, tensile strength, and changes         fiber. Fusion splicers are the most
  in performance under extreme temperature
  and/or humidity conditions (or as a result of
                                                        expensive devices of this sort.             s
  repeated changes in temperature). The val-
                                                    s   Polishers, which are used to prepare
                                                        fiber ends for splicing or connection.
  ues for these properties are used to evaluate
  cable performance.                                s   A microscope, so you can inspect the
      The equipment you might need to test              results of a splicing or polishing opera-   v
  fiber-optic cables in a network setting                tion. A microscope may be included in
  includes the following:                               an installation toolkit.                    w
     s   An installation kit—a general-purpose                                                      x
         tool set for dealing with optical fiber.
152     Cable, Horizontal

Fiber-Optic Cable Vendors                                w
                                                          Cable, Horizontal
  Many vendors sell both electrical and                      Horizontal cable is defined by the EIA/TIA-
  fiber-optic cable, as well as connectors,                   568 committee as any cable that goes from a
  installation, and testing tools. The                       wiring closet, or distribution frame, to the
  following vendors offer an extensive                       wall outlet in the work area. Distribution
  selection of fiber-optics products. (See                    frames from a floor or building are con-
  the Cable article for other cable vendors.)                nected to other frames using backbone
    AMP Incorporated: (800) 522-6752;                        cable.
     (717) 564-0100                                             In a sense, horizontal cable is the most
                                                             crucial in the entire network cabling struc-
    CSP (Computer System Products): (800)                    ture. Since it is installed in the walls, floors,
      422-2537; (612) 476-6866; Fax (612)                    ceiling, or ground, the installation process
      476-6966                                               can be difficult and expensive. Moreover, the
    FIS (Fiber Instrument Sales): (800) 445-                 cable should be able to handle future stan-
       2901; (315) 736-2206; Fax (315)                       dards and technology.
       736-2285                                                 The EIA/TIA-568 recognizes four main
                                                             types of horizontal cable, and several
                                                             optional variants. These types are listed in
    Cable; Cable, Coaxial; Cable, Twisted-                   the table “EIA/TIA-568 Main and Optional
    Pair; Connector, Fiber-Optic; Coupler,                   Types of Horizontal Cable.” The EIA/TIA
    Fiber-Optic; FDDI (Fiber Distributed                     specifications call for at least two cables
    Data Interface)

  TY P E S O F H O R I Z O N T A L C A B L E
  CABLE TYPE          MAIN                                            OPTIONAL

  UTP                 100-ohm, four-pair UTP cable                    100-ohm, 25-wire-pair UTP cable

  STP                 150-ohm STP cable, such as that defined in       100-ohm STP cable
                      the IBM Cable System (ICS)

  Coaxial             50-ohm, thin coaxial cable, such as the cable   75-ohm (broadband) coaxial cable, such as
                      used in thin Ethernet networks                  CATV cable

  Optical fiber        62.5/125-micron (step- or graded-index)         Multimode fiber with other core/cladding
                      multimode optical fiber                          ratios of 50/125-micron, 100/140-micron,

  Undercarpet                                                         Flat cable (such as Type 8 in the ICS) that
                                                                      can be run under carpet without posing a
                                                                                           Cable, IBM   153

from this list to be run to every wall outlet.                 w
At least one of these should be unshielded                      Cable, IBM                                      a
twisted-pair (UTP).                                                The IBM Cable System (ICS) was designed      b
                                                                   by IBM for use in its Token Ring networks
                                                                   and also for general-purpose premises wir-   c
    Cable, Backbone
                                                                   ing. The figure “Context and properties of
                                                                   the IBM Cable System” summarizes the fea-
    Cable                                                          tures of this type of cable.                 e
      Twisted-Pair                                                                                              j
  Optical                                                                                                       k
                                            IBM Cable System Properties
  Comprises Types 1 through 9 (of which all types but 4 and 7 are defined)                                      n
  Type 5 is fiber-optic
  Type 3 is unshielded twisted-pair (UTP)
  Remaining types are shielded twisted-pair (STP)                                                               p
  Type 1 is most common in Token Ring Networks
  Type 3 is not recommended for 16 Mbps networks
  Type 3 cable generally requires a media filter                                                                r
  Type 6 is used mainly as short-distance patch cable
  Type 8 is flat cable for use under a carpet                                                                   s
                                                IBM Cable System Uses                                           t
  IBM Token Ring networks                                                                                       u
  10BaseT Ethernet networks
  ARCnet networks                                                                                               v
  ISDN lines
  Some IBM 3270 networks
154     Cable, IBM

     IBM has specified nine types of cable,           transmission) and two pairs of shielded 22-
  mainly twisted-pair, but with more stringent       gauge solid wire (for data). Although not
  specifications than for the generic twisted-        required by the specifications, a plenum
  pair cabling. The type taxonomy also               version is also available, at about twice
  includes fiber-optic cable, but excludes co-        the cost.
  axial cable. The twisted-pair versions differ
  in the following ways:                           Type 3 Cable

    s   Whether the type is shielded or              Type 3 is unshielded twisted-pair (UTP),
        unshielded                                   with two, three, or four pairs of 22- or 24-
                                                     gauge solid wire. The pairs have at least two
    s   Whether the carrier wire is solid            twists per foot. This category requires only
        or stranded                                  voice-grade capabilities, and so may be used
    s   The gauge (diameter) of the carrier          as telephone wire for voice transmissions.
        wire                                         Type 3 is not recommended for 16 Mbps
                                                     Token Ring networks.
    s   The number of twisted pairs                      Although not required by the specifica-
     Specifications have been created for seven       tions, a plenum version is also available, at
  of the nine types. Types 4 and 7 are unde-         about twice the cost.
  fined; presumably, they are reserved for                Type 3 cable is becoming more popular as
  future use.                                        adapter cable, which is used to connect a
                                                     node to a MAU. You must use a media filter
Type 1 Cable                                         if you are using Type 3 cable to connect a
                                                     node to a MAU or if you need to switch
  Type 1 cable is shielded twisted-pair (STP),
                                                     between UTP and STP in a Token Ring net-
  with two pairs of 22-gauge solid wire. It is
                                                     work. However, you should not mix Type 1
  used for data-quality transmission in IBM’s
                                                     and 3 cable in the same ring. Mixing cable
  Token Ring network. It can be used for the
                                                     types makes trouble-shooting difficult.
  main ring or to connect lobes (nodes) to
                                                         Some manufacturers offer higher-quality
  multistation attachment units (MAUs),
                                                     Type 3 cable for greater reliability. Such
  which are wiring centers.
                                                     cable has more twists per foot, for greater
     Although not required by the specifica-
                                                     protection against interference. Many ven-
  tions, a plenum version is also available, at
                                                     dors recommend that you use Category 4
  about twice the cost of the nonplenum cable.
                                                     cable (with 12 twists per foot). This category
  Compare Type 1 with Type 6.
                                                     of cable costs about 20 percent more than
Type 2 Cable                                         ordinary Type 3 cable, but is rated for higher
                                                     speeds. The category value represents a clas-
  Type 2 is a hybrid consisting of four pairs of     sification system for the performance of UTP
  unshielded 22-gauge solid wire (for voice          cable. See the Cable Standards article for
                                                     more information.
                                                                             Cable, Quadrax      155

                                                       distances usually involved in under-the-
Type 5 Cable
                                                       carpet cabling.
  Type 5 is fiber-optic cable, with two glass
  fiber cores, each with a 100-micron diameter      Type 9 Cable                                           b
  and a 140-micron cladding diameter. (IBM
                                                       Type 9 is STP cable, with two pairs of
  also allows the more widely used 62.5/125-
  micron fiber.)
                                                       26-gauge solid or stranded wire. This type         d
                                                       is covered with a plenum jacket and is
     This type is used for the main ring
                                                       designed to be run between floors.                  e
  path (the main network cabling) in a Token
  Ring network to connect MAUs over greater            SEE ALSO                                           f
  distances or to connect network segments
  between buildings. Plenum versions of
                                                         Cable, Twisted-Pair                              g
  Type 5 cable are available at only a slightly
                                                   Cable, Patch                                           h
  higher cost.
                                                       Cable used to connect two hubs or multi-           i
Type 6 Cable                                           station attachment units (MAUs). IBM Type
                                                       1 or Type 6 patch cables can be used for
  Type 6 is STP cable, with two pairs of 26-
  gauge stranded wire. This type is commonly
                                                       Token Ring networks.                               k
  used as an adapter cable to connect a node
  to a MAU. In that type of connection, the
                                                       SEE ALSO                                           l
                                                         Cable, IBM
  PC end of the cable has a male DB-9 or                                                                  m
  DB-25 connector, and the MAU end has
  a specially designed IBM data connector.
                                                   Cable, Plenum                                          n
     Type 6 cable is also used as a patch cable;       Cable that has a fire-resistant jacket, which       o
  for example, to connect MAUs. For this               will not burn, smoke, or give off toxic fumes
  use, the cable has IBM data connectors at            when exposed to heat. The cable goes               p
                                                       through a plenum, a conduit, or shaft, run-
  each end.
     Because Type 6 is used mostly for shorter         ning inside a wall, floor, or ceiling. Fire regu-
  distances, the price per foot tends to be            lations generally stipulate that cable running     r
  higher than for other cable types.                   through such conduits must be fireproof.
Type 8 Cable                                           SEE ALSO
                                                         Cable                                            t
  Type 8 is STP cable, with two pairs of flat,
  26-gauge solid wire. This type is specially      w                                                      u
                                                   Cable, Quadrax
  designed to be run under a carpet, so the                                                               v
  wires are flattened. This makes the cable             A type of coaxial cable. Quadrax cable,
  much more prone to signal loss than Type 1           sometimes known simply as quadrax, is              w
                                                       a hybrid of triaxial and twinaxial cable. Like
  or Type 2 cable; however, the performance
  of Type 8 cable is adequate for the short            twinaxial cable, quadrax has the extra
156       Cable, Quad Shield

    carrier wire with dielectric; like triaxial         often distinguished in terms of the demands
    cable, quadrax has extra shielding.                 made on the cable. The standards also spec-
                                                        ify the minimum behavior required under
                                                        extreme conditions, such as fire.
       Cable, Coaxial
                                                           The most commonly used safety stan-
w                                                       dards in the United States are those specified
Cable, Quad Shield                                      in the National Electric Code and in docu-
    A type of coaxial cable with four layers of         ments from Underwriters Laboratories.
    shielding: alternating layers of foil and braid     Other standards are specified by the Elec-
    shields. Quad shield cable, sometimes               tronic Industries Association/Telecommuni-
    known simply as quad shield, is used in             cations Industries Association, Electrical
    situations where heavy electrical interference      Testing Laboratory, and Manufacturing
    can occur, such as in industrial settings.          Automation Protocol.

    SEE ALSO                                          The National Electric Code (NEC)
       Cable, Coaxial
                                                        The NEC is published by the National Fire
w                                                       Protection Agency (NFPA, 617-770-3000),
Cable, Riser                                            and specifies safety standards for general-
    Cable that runs vertically; for example,            purpose cables in commercial and residential
    between floors in a building. Riser cable            environments, and also specifically for
    often runs through available shafts (such as        cables used for communications. The
    for the elevator). In some cases, such areas        Class 2 (CL2x) standards apply to general-
    can be a source of electrical interference.         purpose cables, and the Communications
    Consequently, optical fiber (which is imper-         (CMx) standards apply to special-purpose
    vious to electromagnetic interference) is gen-      cables capable of carrying data.
    erally used as rise cable.                             Of the CL2 standards, the most strin-
                                                        gent ones apply to Class 2 plenum cable
w                                                       (CL2P). Cable that meets or exceeds these
Cable Standards                                         standards is said to be CL2P compliant.
    Several cable standards are concerned with          CMP-compliant cable meets the correspond-
    the performance and reliability of cables           ing standard for plenum communications
    under actual working conditions. In particu-        cable.
    lar, these standards specify the cable’s mini-         The less stringent CL2R standards apply
    mal acceptable behavior under adverse               to riser cable (cable that can be used, for
    working conditions; for example, in manu-           example, in a vertical utility shaft between
    facturing or industrial environments, where         floors in a building). The corresponding
    heavy machinery is turned on and off during         standard for communications riser cable
    the course of operations. Such actions can          is CMR.
    generate strong interference and power-                Be wary if you intend to use cable that is
    supply variations. Cable environments are           neither CMx- nor CL2x-compliant. Older
                                                                                Cable Standards      157

  cable that is already in the walls may be           cable is UL-listed but is not suitable for envi-
  noncompliant.                                       ronments that demand fire protection.
Underwriters Laboratories (UL)
                                                         For most networking applications, cable
                                                      that meets requirements for UL level III or
  UL tests cable and other electrical devices to      above should be adequate.                            c
  determine the conditions under which the
                                                      U N D E RW R I T E R S L A B O R ATO R I E S
  cable or device will function safely and as
  specified. UL-listed products have passed
                                                      (UL) PHONE NUMBERS
  safety tests performed by inspectors at the         East Coast: (516) 271-6200
  Underwriters Laboratories.
                                                      Central: (708) 272-8800
     Two tests are most directly relevant to                                                               g
  network cable:                                      West Coast: (408) 985-2400
    UL-910: Tests smoke emissions and the
      spread of flames for plenum cable.             Electronic Industries                                  i
      This test corresponds to the CL2P level
      of safety standards. A cable that passes      Industries Association (EIA/TIA)
      the UL-910 test is rated as OFNP
                                                      A committee for EIA/TIA has created yet
      (optical fiber, nonconductive plenum)
      by UL.
                                                      another classification system for specifying          l
                                                      the performance of unshielded twisted-pair
    UL-1666: Tests the performance of riser           (UTP) cable. The EIA/TIA taxonomy                    m
      cable in a fire. This test corresponds
      roughly to the CL2R level of safety
                                                      includes the following categories (1 through
                                                      5) whose criteria correspond roughly to
      standards. A cable that passes the UL-          the performance criteria specified for the UL         o
      1666 test is rated as OFNR (optical             levels:
      fiber, nonconductive riser) by UL.
                                                         Category 1: Voice-grade, UTP telephone
      UL also uses a system of markings to cat-            cable. This describes the cable that has        q
  egorize cable as falling into one of five levels          been used for years in telephone com-
                                                           munications. Officially, such cable is
  (I through V). Cables that meet level I and II
  standards meet minimum UL safety require-                not considered suitable for data-grade          s
  ments, but the performance of these cables               transmissions (in which every bit must
  may be inadequate for networking purposes.               get across correctly). In practice,             t
  Cables that meet level III, IV, or V standards           however, it works fine over short dis-           u
  meet both safety and various performance                 tances and under ordinary working
  requirements. Higher levels allow for less               conditions.                                     v
  attenuation and interference due to crosstalk
  than lower levels.
                                                         Category 2: Data-grade UTP, capable of            w
                                                           supporting transmission rates of up to
      Cable should be UL-listed, and just about
                                                           4 megabits per second (Mbps). IBM
  every cable is. However, you need to find out
  which listing applies. For example, OFNR
                                                           Type 3 cable falls into this category.          y
158     Cable Standards

    Category 3: Data-grade UTP, capable of                (StarLAN) Ethernet network devel-
      supporting transmission rates of up to              oped by AT&T.
      10 Mbps. A 10BaseT network requires
                                                        Level 3: Data-grade cable that is capable
      such cable.
                                                          of transmission speeds as high as
    Category 4: Data-grade UTP, capable of                16 Mbps. This level corresponds to
      supporting transmission rates of up to              Category 3 cable in the EIA/TIA-568
      16 Mbps. A 16 Mbps IBM Token Ring                   specifications. Level 3 cable is used in
      network requires such cable.                        4 Mbps or 16 Mbps Token Ring net-
                                                          works, and also in 10BaseT Ethernet/
    Category 5: Data-grade UTP, capable of
                                                          802.3 networks.
      supporting transmission rates of up to
      155 Mbps (but officially only up to                Level 4: Data-grade cable that is capable
      100 Mbps). The proposed CDDI                        of transmission speeds as high as 20
      (Copper Distributed Data Interface)                 Mbps. This level corresponds to Cate-
      networks and 100Base-X network                      gory 4 cable in the EIA/TIA-568 speci-
      architecture require such cable.                    fications. Level 4 cable is used for
                                                          ARCnet Plus, a 20 Mbps version of the
Performance Levels                                        ARCnet network architecture.
  Many cable vendors also use a five-level sys-          Level 5: Data-grade cable that is capable
  tem to categorize their UTP cable. Just as              of transmission speeds as high as 100
  there is overlap in the paths to enlighten-             Mbps. This level corresponds to Cate-
  ment in various religious traditions, there is          gory 5 cable in the EIA/TIA-568 speci-
  some overlap between these levels and the               fications. Level 5 cable is used for
  other systems discussed here. For example,              CDDI (or TPDDI), which are copper-
  the references to Level 4, Category 4 cable             based implementations of the 100
  identify the cable according to the features            Mbps FDDI network architecture.
  described here and also according to the                100Base/X, a proposed 100 Mbps ver-
  features in the EIA/TIA specifications.                  sion of Ethernet, is also intended to
    Level 1: Voice-grade cable, which is suit-            run on this type of cable.
      able for use in the “plain old telephone     Electrical Testing Laboratory (ETL)
      system” (or POTS). Such cable can
      handle data at up to 1 Mbps.                   The ETL is an independent laboratory that
                                                     tests and rates products for manufacturers.
    Level 2: Data-grade cable that is capable
                                                     Vendors specify if their cable has been tested
      of transmission speeds as high as 4
                                                     and verified by ETL.
      Mbps. This level corresponds roughly
      to the Type 3 cable described in IBM’s       Manufacturing Automation Protocol (MAP)
      Cabling System (see the Cable, IBM
      article). Level 2 cable also meets             The most commonly observed performance
      the requirements for the 1Base5                standards, arguably, are those associated
                                                                           Cable, Twisted-Pair   159

    with the MAP. Among other things, this            w
    standard specifies the expected performance        Cable,Triaxial                                      a
    for cables in the highly automated and
    machinery-heavy industrial working envi-
                                                          A type of coaxial cable. Also called triax,     b
                                                          this cable has an inner braid surrounded by
    ronments of the future.                               an inner (nonplenum) jacket, surrounded         c
        Cable that meets MAP standards gener-
    ally has quad shields; that is, four layers of
                                                          by an outer copper braid. The extra shield-
                                                          ing makes a big difference because of the
    shielding around the central core in a coax-          grounding and improved protection.              e
    ial cable. The four layers of shielding help
    protect the cable against signal loss from the        SEE ALSO                                        f
                                                            Cable, Coaxial
    conductor wire and against electromagnetic                                                            g
    interference from the outside world; for          w
    example, from heavy machinery being               Cable,Twinaxial                                     h
    turned on and off. See the MAP article
    for more information.
                                                          A type of coaxial cable. Also called twinax,    i
                                                          this cable has two insulated carrier wires,
w                                                         generally twisted around each other, which      j
Cable Tester                                              helps cut down considerably on magnetic
                                                          interference. Twinaxial cables are used in
    An instrument for testing the integrity and
    performance of a stretch of cable. Cable              IBM and AppleTalk networks.                     l
    testers run various tests to determine the            SEE ALSO                                        m
    cable’s attenuation, resistance, characteristic         Cable, Coaxial
    impedance, and so on. High-end testers can                                                            n
    test cable for conformity to various network      w
    architecture specifications, and can some-         Cable,Twisted-Pair                                  o
    times even identify a particular type of cable.       Twisted-pair cable is very widely used, inex-   p
                                                          pensive, and easy to install. It can transmit
                                                          data at an acceptable rate (up to 100 Mbps
    Cable used to connect a network interface
                                                          in some network architectures). The best-       r
                                                          known example of twisted-pair wiring is
    card to a transceiver, mainly in Ethernet             probably telephone cable, which is              s
    architectures. A transceiver cable usually            unshielded and is usually voice-grade, rather
    has an AUI connector at one end and an                than the higher-quality data-grade cable
    N-series or other type of connector at the
    other end. Coaxial transceiver cable comes
                                                          used for networks. The figure “Context and       u
                                                          properties of twisted-pair cable” summa-
    in thick and thin versions. You can also get          rizes the features of this type of cable.       v
    special cable with a built-in right angle.               In a twisted-pair cable, two conductor
                                                          wires are wrapped around each other. A sig-
                                                          nal is transmitted as a differential between    x
                                                          the two conductor wires. This type of signal
160      Cable, Twisted-Pair




                                    Shielded Twisted-Pair (STP) Properties
  Includes shield around twisted pairs
  150 ohm impedance
  Information in differential signal between wires in a pair
  Subject to near-end crosstalk (NEXT)
  Subject to electromagnetic interference
  Generally uses RJ-xx connectors

                                         Shielded Twisted-Pair (STP) Uses

  IBM Token Ring networks
  ARCnet networks
  Rarely in Ethernet networks

                                   Unshielded Twisted-Pair (UTP) Properties

  No shield around twisted pairs
  100 ohm impedance
  Information in differential signal between wires in a pair
  Subject to near-end crosstalk (NEXT)
  Subject to electromagnetic interference
  Generally uses RJ-xx connectors
  Performance grades specified in EIA/TIA-568 CATEGORIES I-5

                                     Unshielded Twisted-Pair (UTP) Uses

  10BaseT Ethernet networks
  ARCnet networks
  Certain sections of IBM Token Ring networks
  Telephone lines (voice-grade)
                                                                    Cable, Twisted-Pair   161

  is less prone to interference and attenuation,     Jacket: The wire bundles are encased in
  because using a differential essentially gives        a jacket made of polyvinylchloride
  a double signal, but cancels out the random
  interference on each wire.
                                                        (PVC) or, in plenum cables, of a
                                                        fire-resistant material, such as Teflon
      Twisting within a pair minimizes cross-           or Kynar.                                c
  talk between pairs. The twists also help deal
  with electromagnetic interference (EMI) and
                                                      The figure “Components of twisted-pair      d
                                                   cable” shows the makeup of this type of
  radio frequency interference (RFI), as well as
                                                   cable. Note that the shield is not included
  signal loss due to capacitance (the tendency
  of a nonconductor to store up electrical
                                                   for unshielded twisted-pair cable.            f
                                                      Twisted-pair cable comes in two main
  charge). The performance of a twisted-pair
                                                   varieties: shielded (STP) and unshielded      g
  cable can be influenced by changing the
  number of twists per foot in a wire pair.        COMPONENTS OF                                 h
      IBM has developed its own categorization
  system for twisted-pair cable, mainly to
                                                   TWISTED-PAIR CABLE                            i
  describe the cable supported for IBM’s                                                         j
  Token Ring network architecture. The sys-
  tem is discussed in the Cable, IBM article.                                                    k
Twisted-Pair Cable Components                                                                    l
  A twisted-pair cable has the following                                                         m
  components:                                                                                    n
    Conductor wires: The signal wires for this
      cable come in pairs that are wrapped
      around each other. The conductor                                                           p
      wires are usually made of copper. They
      may be solid (consisting of a single                                                       q
      wire) or stranded (consisting of many                                                      r
      thin wires wrapped tightly together).
      A twisted-pair cable usually contains                                                      s
      multiple twisted-pairs; 2, 4, 6, 8, 25,
      50, or 100 twisted-pair bundles are
      common. For network applications,                                                          u
      2- and 4-pair cables are most com-
      monly used.                                                                                v
    Shield: Shielded twisted-pair (STP) cable                                                    w
       includes a foil shield around each pair
       of conductors.
162     Cable, Twisted-Pair

  (UTP). STP contains an extra shield or                Because it lacks shielding, UTP is not as
  protective screen around each of the wire          good at blocking noise and interference as
  pairs to cut down on extraneous signals.           STP or coaxial cable. Consequently, UTP
  This added protection also makes STP more          cable segments must be shorter than when
  expensive than UTP. (The price of coaxial          using other types of cable. For standard
  cable actually lies between UTP and STP            UTP, the length of a segment should never
  prices.)                                           exceed 100 meters (about 330 feet).
                                                        On the other hand, UTP is quite inexpen-
Shielded Twisted-Pair (STP) Cable                    sive, and is very easy to install and work
  STP cable has pairs of conductors twisted          with. The price and ease of installation make
  around each other. Each pair is covered with       UTP tempting, but keep in mind that instal-
  a foil shield to reduce interference and mini-     lation is generally the major part of the
  mize crosstalk between wire pairs.                 cabling expense (so saving on the cable
     STP can handle high-speed transmissions,        won’t necessarily help cut expenses very
  but the cable itself is relatively expensive,      much) and that other types of cable may
  can be quite bulky and heavy, and is rather        be just as easy to install.
  difficult to work with.                                To distinguish varieties of UTP, the
     STP is used in ARCnet and Token Ring            EIA/TIA has formulated five categories.
  networks, although the special cable ver-          These are summarized in the Cable Stan-
  sions developed by IBM are more likely             dards article.
  to be used in the Token Ring networks.           Performance Features
  Several of the types specified in the IBM
  Cable System are STP: Types 1, 2, 6, 8,            Twisted-pair cable is described in terms of
  and 9 (see the Cable, IBM article).                its electrical and performance properties.
                                                     The features that characterize UTP and STP
Unshielded Twisted-Pair (UTP) Cable                  cable include the following:
  UTP cable does not include any extra shield-         Attenuation: This value indicates how
  ing around the wire pairs. This type of cable          much power the signal has lost and is
  is used in some Token Ring networks, usu-              dependent on the frequency of the
  ally those working at slower speeds. UTP               transmission. Attenuation is measured
  can also be used in Ethernet and ARCnet                in relation to a specified distance; for
  architectures.                                         example, 100 meters, 1000 feet, or 1
     UTP is not the primary choice for any               kilometer. Attenuation per 1000 feet
  network architecture, but the IEEE has                 values range from under 10 dB (for
  approved a standard for a 10BaseT Ethernet             Category 4 cable running at 1 MHz) to
  network that uses UTP cabling at 10 Mbps.              more than 60 dB (for Category 5 cable
  Networking mavens are divided as to                    running at 100 MHz). With attenua-
  whether 10BaseT and the use of UTP                     tion, a lower value is better.
  cable in general are welcome additions
  or dead-ends.
                                                                           Cable, Twisted-Pair    163

                                                               better performance. Typical values are
CABLE WIRES                                                    between 15 and 25 pF/ft.
Most telephone cable is UTP, and many tele-                Impedance: All UTP cable should have an         b
phone cables have extra wires because the cable              impedance of 100 +/− 15 ohms.                 c
comes with four pairs and the telephone com-               NEXT: The near-end crosstalk (NEXT)
pany needs only two of the pairs for your tele-              indicates the degree of interference
phone connection. (Any additional lines or
intercoms require their own wire pairs.)
                                                             from neighboring wire pairs. This is          e
                                                             also measured in decibels per unit dis-
If there are unused wire pairs, you may be able to           tance, but because of notation and            f
use these for your network cabling. While this is            expression conventions, a high value is       g
a tempting possibility, consider the following               better for this feature. NEXT depends
points carefully:                                            on the signal frequency and cable cate-       h
                                                             gory. Performance is better at lower
s   The cable might not run conveniently for                 frequencies and for cables in the higher
    your needs, so you may need to add cable
                                                             categories.                                   j
s   Make sure you test all the cable you’ll be
                                                       Twisted-Pair Cable Advantages                       k
    using, and don’t be surprised if some of it is       Twisted-pair cable has the following              l
    defective.                                           advantages over other types of cables
s   The telephone cable may be the lower-quality,
                                                         for networks:                                     m
    voice-grade type, and you really should be             s   It is easy to connect devices to twisted-   n
    using data-grade cable, unless you’re transmit-            pair cable.
    ting over very short distances.
                                                               If an already installed cable system,
If you’re going to use already installed cable for

                                                               such as telephone cable, has extra,
your network, make sure all of it works properly.              unused wires, you may be able to use        q
Use a cable tester, which can provide detailed                 a pair of wires from that system. For
information about the cable’s physical and electri-            example, in order to use the telephone      r
cal properties. When you’re dealing with a long
cable system, the chances are good that at least
                                                               cable system, you need telephone cable
                                                               that has four pairs of wires, and there
parts of it will be faulty. Find and replace the bad           can be no intercoms or second lines to      t
cable before you set everything up.                            use the two pairs not needed for the
                                                               telephone connection.                       u
    Capacitance: This value indicates the                  s   STP does a good job of blocking             v
      extent to which the cable stores up
      charge (which can distort the signal).
                                                               interference.                               w
                                                               UTP is quite inexpensive.
      Capacitance is measured in picofarads
      (pF) per foot; lower values indicate
                                                               UTP is very easy to install.
 164     Cable, Voice-Grade

     s   UTP may already be installed (but                Check the wiring sequence before you
         make sure it all works properly and           purchase cable. Different wiring sequences
         that it meets the performance specifi-         can lurk behind the same modular plug in a
         cations your network requires).               twisted-pair cable. (A wiring sequence, or
                                                       wiring scheme, describes how wires are
Twisted-Pair Cable Disadvantages                       paired up and which locations each wire
  Twisted-pair cable has the following dis-            occupies in the plug.) If you connect a plug
  advantages compared with other types of              that terminates one wiring scheme into a
  cable:                                               jack that continues with a different
                                                       sequence, the connection may not provide
     s   STP is bulky and difficult to work             reliable transmission. See the Wiring
         with.                                         Sequence article for more information.
     s   UTP is more susceptible to noise and             You should find out which wiring scheme
         interference than coaxial or fiber-optic       is used before buying cable, and buy only
         cable.                                        cable that uses the same wiring scheme. If
                                                       you are stuck with existing cable that uses
     s   UTP signals cannot go as far as they          an incompatible wiring scheme, you can use
         can with other cable types before they        a cross wye as an adapter between the two
         need cleaning and boosting.                   schemes.
     s   A skin effect can increase attenuation.          If any of your cable purchases include
         This occurs when transmitting data at         patch cables (for example, to connect a com-
         a fast rate over twisted-pair wire.           puter to a wallplate), be aware that these
         Under these conditions, the current           cables come in two versions: straight
         tends to flow mostly on the outside            through or reversed. For networking appli-
         surface of the wire. This greatly             cations, use straight-through cable, which
         decreases the cross-section of the wire       means that wire 1 coming in connects to
         being used for moving electrons, and          wire 1 going out (rather than to wire 8 as in
         thereby increases resistance. This, in        a reversed cable), wire 2 connects to wire 2
         turn, increases signal attenuation,           (rather than to wire 7), and so on. The tools
         or loss.                                      for installing and testing twisted-pair cable
                                                       are the same as those used generally for net-
Selecting and Installing Twisted-Pair Cable            work cables. (See the Cable article for a dis-
                                                       cussion of cable tools.)
  When you are deciding on a category of
  cable for your needs, take future develop-           SEE ALSO
  ments—in your network and also in technol-             Cable; Cable, Coaxial; Cable, Fiber-Optic
  ogy—into account. It is a good idea to buy
  the cable at least one category above the one    w

  you have selected. (If you selected Category
  5 cable to begin with, you should seriously          Old-time, unshielded twisted-pair, tele-
  consider fiber-optic cable.)                          phone cable; category 1 in the EIA/TIA-568
                                                                        Cache Buffer Pool    165

    specifications. This cable is suited to the        on a hard or floppy disk with much slower
    transmission of voice signals. Officially,         access, a system’s performance can be
    such cable is not considered suitable for
    data-grade transmissions. In practice, it
                                                      improved greatly.
                                                         As a verb, cache refers to the process of
    generally works fine at low speeds, over           putting information into a cache for faster    c
    short distances, and under ordinary working       retrieval. Directory information and hard
    conditions.                                       disk contents are examples of data likely to
                                                      be cached. The figure “Disk cache” shows        e
                                                      an example of this process.
      Cable, Twisted-Pair                                                                            f
w                                                 Cache Buffer Pool                                  g
    As a noun, a cache, also known as a disk
                                                      In Novell’s NetWare, the cache buffer pool     h
                                                      is the amount of memory available for the
    cache, is an area of RAM (random-access
                                                      network operating system (NOS) after the       i
    memory) set aside for holding data that is
                                                      server module has been loaded into memory.
    likely to be used again. By keeping fre-                                                         j
    quently used data in fast RAM, instead of
166       Call

    The memory in this pool can allocated for             by a considerable distance. Unlike a wide-
    various purposes:                                     area network, however, a campus network
                                                          does not require remote communications
      s   To cache the file allocation tables
                                                          facilities, such as modems and telephones.
          (FATs) for each NetWare volume
      s   To create a hash table containing
                                                      Campus-Wide Information System
          directory information
      s   To provide memory for NetWare                   SEE
          Loadable Modules (NLMs) that are
                                                            CWIS (Campus-Wide Information
    A request from one program or node to
                                                          Capacitance is the ability of a dielectric
    begin a communication with another node.
                                                          (nonconductive) material to store electricity
    The term is also used to refer to the resulting
                                                          and to resist changes in voltage. In the pres-
    communications session.
                                                          ence of a signal (a voltage change), the
w                                                         dielectric will store some of the charge.
Caller ID                                                 Capacitance is usually measured in micro-
    In ISDN and some other telecommunica-                 farads or picofarads (millionths or trillionths
    tions environments, a feature that includes           of a farad, respectively).
    the sender’s identification number (such as               Other things being equal, the lower the
    telephone number) in the transmission so              capacitance, the better the cable. A higher
    that the receiver knows who is calling.               capacitance means that more of the charge
    Caller ID is also known as ANI (automatic             can be stored in the dielectric between two
    number identification) and CLID (calling               conductors, which means greater resistance.
    line identification).                                  At higher frequencies, high capacitance
                                                          results in greater signal attenuation.
Call Setup Time                                           SEE ALSO
    The amount of time needed to establish a
    connection between two nodes so they can          w
    communicate with each other.                      Capacitor
w                                                         An electrical component in line conditioners,
Campus Area Network (CAN)                                 surge protectors, and other equipment.
                                                          Capacitors help clean incoming power by
    A network that connects nodes (or possibly
                                                          absorbing surges and noise from electromag-
    departmental local-area networks) from
                                                          netic and radio frequency interference. Com-
    multiple locations, which may be separated
                                                          pare it with inductor and MOV (metal oxide
                                                    CAT (Common Authentication Technology)         167

w                                                         well-defined properties, but conveys no
Carrier Band                                              information (content). Information is sent
    A communications system in which the                  by modifying (modulating) some feature of
                                                          the carrier signal, such as the amplitude, fre-
    entire bandwidth is used for a single trans-
    mission and in which a signal is modulated            quency, or timing, to represent the values        c
                                                          being transmitted.
    before being transmitted. This is in contrast
    to baseband systems, which do not modu-
    late the signal, and to broadband systems,
                                                       Carrier Wire                                         e
    which divide the total bandwidth into multi-
    ple channels.
                                                          A conductive wire (capable of carrying an         f
                                                          electrical signal); for example, the central
w                                                         wire in a coaxial cable, which serves as the
Carrier Frequency                                         medium for the electrical signal.                 h
    The rate at which the carrier signal repeats,
    measured in cycles per second, or hertz. In
                                                          SEE ALSO                                          i
    communications, the carrier signal is modu-                                                             j
    lated, or altered, by superimposing a second      w
    signal, which represents the information           CAS (Communicating Application                       k
    being transmitted. In an acoustic signal, the      Specification)
    frequency represents the signal’s pitch.              An interface standard for fax modems devel-
                                                          oped by Intel and DCA. This proposed stan-        m
                                                          dard competes with the Class x hierarchy
Carrier On
                                                          developed by EIA.
    In carrier sense, multiple access (CSMA)
    media-access methods, a signal that indi-         w
    cates the network is being used for a trans-       CAT (Common Authentication
    mission. When a node detects this signal, the
    node waits a random amount of time before             In the Internet community, CAT is a specifi-
    trying again to access the network.                   cation for distributed authentication under       r
                                                          development. CAT supports authentication
Carrier Pulse                                             measures based on either public- or private-      s
                                                          key encryption strategies.
    A signal, consisting of a series of rapid, con-
                                                             With CAT, both client and server pro-
    stant pulses, used as the basis for pulse mod-
    ulation; for example, when converting an
                                                          grams must use the services of a common           u
                                                          interface, which will provide the authentica-
    analog signal into digital form.
                                                          tion services. This interface will connect to     v
Carrier Signal
                                                          either DASS (Distributed Authentication
                                                          Security Service), which uses public-key
    An electrical signal that is used as the              encryption, or Kerberos, which uses private-      x
                                                          key encryption.
    basis for a transmission. This signal has                                                               y
168       CAU (Controlled Access Unit)

    B RO A D E R C A T E G O R Y                   w
       Authentication                              CBMS (Computer-Based Messaging
       DASS (Distributed Authentication                An older term for a Message Handling
       Security Service); Kerberos                     System (MHS), or for electronic mail.
CAU (Controlled Access Unit)                             E-Mail

    In IBM Token Ring networks, the term for       w
    an intelligent hub. CAUs can determine         CBR (Constant Bit Rate)
    whether nodes are operating, connect and           An ATM connection that uses Class A ser-
    disconnect nodes, monitor node activity, and       vice, which is designed for voice or other
    pass data to the LAN Network Manager               data that are transmitted at a constant rate.
    program.                                           Compare it with VBR (variable bit rate).
w                                                  w
CAU/LAM (Controlled Access Unit/                   CC (Clearing Center)
Lobe Attachment Module)
                                                       In EDI, a message-switching element
    In IBM Token Ring networks, a hub (the             through which documents are passed on
    CAU) containing one or more boxes                  the way to their destinations.
    (the LAM) with multiple ports to which
    new nodes can be attached.                         SEE ALSO
                                                         EDI (Electronic Document Interchange)
CBC (Cipher Block Chaining)                        w
                                                   CCIR (International Consultative
    An operating mode for the DES.
                                                   Committee for Radiocommunication)
    SEE                                                An ITU (International Telecommunication
       DES (Data Encryption Standard)                  Union) agency that is responsible for defin-
                                                       ing standards for radio communications.
CBEMA (Computer Business                               In 1993, the CCIR—together with the IFRB
Manufacturers Association)                             (International Frequency Registration
                                                       Board)—was replaced by the ITU-R
    An organization that provides technical
                                                       (International Telecommunication Union—
    committees for work being done by other
                                                       Radiocommunication Standardization
    organizations; for example, the committee
    for the FDDI standard published by ANSI.
                                                       SEE ALSO
                                                     CCS (Common Communications Support)         169

w                                                            In March 1993, the CCITT was officially
CCIS (Common Channel                                      renamed the International Telecommunica-
Interoffice Signaling)                                     tion Union-Telecommunication Standardi-
                                                          zation Sector (ITU-T, sometimes written as
    In telephone communications, a transmis-
    sion method that uses different channels for          ITU-TS or ITU-TSS). However, since the          c
                                                          CCITT name is so familiar and is likely to
    voice and control signals. The control sig-
    nals are sent by a fast, packet-switched              remain in widespread use for some time, the
    method, which makes it possible to include            older name is used throughout this book.        e
    extra information (such as caller ID and bill-
    ing information) in the control channel.
                                                       CCRSE (Commitment, Concurrency,
    SEE ALSO                                           and Recovery Service Element)                      g
      CCS 7                                               In the OSI Reference Model, an application-     h
                                                          layer service that is used to implement
CCITT (Consultative Committee                             distributed transactions among multiple         i
for International Telegraphy and                                                                          j
    The CCITT is a permanent subcommittee of
                                                            ASE (Application Service Element)
    the ITU (International Telecommunications
    Union), which operates under the auspices
    of the United Nations. The committee con-
                                                       CCS (Common Channel Signaling)                     m
                                                          A signaling method in which control signals
    sists of representatives from 160 member
                                                          are sent across different channels than voice
    nations, mostly from national PTT (Postal,
    Telephone, and Telegraph) services.
                                                          and data signals. This makes it possible to     o
                                                          include various types of extra information in
       The CCITT is responsible for dozens of
    standards used in communications, telecom-
                                                          the control signal.                             p
    munications, and networking, including the            SEE ALSO                                        q
    X.25 and X.400 standards, the V.42 and                  CCS 7
    V.42bis standards for modems, and the
    I.xxx series of documents on ISDN (Inte-          w
                                                       CCS (Common Communications                         s
    grated Services Digital Network).
       The CCITT works closely with the
                                                       Support)                                           t
    ISO (International Standardization                    One of the pillars of IBM’s SAA specifica-
    Organization), so that many standards and             tions. CCS includes support for data links,
    recommendations will appear in documents              application services, session services, and     v
    from both groups. CCITT recommendations               data streams.
    appear every four years, with 1992 (the                                                               w
                                                          SEE ALSO
    white books) being the most recent.
                                                            SAA (Systems Application Architecture)        x
170       CCS (Continuous Composite Servo)

w                                                      magnetic technology, compact discs are
CCS (Continuous Composite Servo)                       recorded using optical methods.
    A compact disc recording technique in                 To produce a master disc for commer-
    which the contents are stored on separate          cially produced CDs, a laser literally burns
    tracks laid out in concentric circles.             the information into the disc by creating tiny
                                                       pits in the surface. This changes the reflec-
    C O M P A RE
                                                       tive properties of the disc at these locations
      SS (Sampled Servo)                               relative to the surrounding surface. The
                                                       information is read by using a laser so that
CCS (Hundreds of Call Seconds)                         there is never any physical contact during
                                                       the reading process. The information on a
    In telephone communications, a measure of
                                                       CD is actually contained in the transitions
    line activity. One CCS is equivalent to 100
                                                       between the pits and the non-pit areas
    seconds of conversation on a line, so that an
                                                       (known as the lands).
    hour of line usage is 36 CCS; 36 CCS is
                                                          CD technology has undergone several
    equal to one Erlang, and indicates continu-
                                                       revisions and advancements since the first
    ous use of the line.
                                                       digital audio (DA) discs were developed over
w                                                      10 years ago.
CCS 7 (Common Channel Signaling 7)
                                                     CD Variants
    A version of the CCITT’s Signaling System 7
    (SS7); a transmission method in ISDN that          The following standards and variants have
    makes special services (such as call forward-      been created and used over the years. Most
    ing or call waiting) available anywhere in         of these standards are still in use, and many
    a network. CCS 7 is an extension of the            current CD drives can read several of the
    CCIS method for transmitting control               standards. In addition, newer standards
    information.                                       (such as CD-XA) are often back-compatible
                                                       with earlier standards (such as CD-ROM).
CD (Carrier Detect)                                       CD standards are distinguished by the
                                                       color of the laser used in that particular
    A signal sent from a modem to a PC, to indi-       technology—for example, red, yellow, and
    cate that the modem is on line and ready for       green. Collectively, these standards docu-
    work.                                              ments are known as the Rainbow Books.
w                                                      The following standards are among the most
CD (Compact Disc)                                      popular:
    Compact discs are the product of a record-           CD-DA (Digital Audio) (Red Book) This
    ing and storage technology that makes it               was the first compact disc standard,
    possible to fit over half a gigabyte of digital         and was developed for recording musi-
    data on a disc about the size of a floppy               cal discs. CD-DA discs can hold about
    disk. Unlike floppy or hard disks, which use            74 minutes of music recorded at
                                                          CD (Compact Disc)     171

  44,100 samples per second (known as          storage method, but one that could be
  the scanning frequency), using PCM           made back-compatible with earlier
  (pulse code modulation) as the digiti-
  zation method, and allocating 16 bits
                                               standards. In addition to providing a
                                               new, more flexible sector format, CD-
  for each sample. (With 16 bits, each         ROM/XA uses a different digitization      c
  sample can take on any of 65, 536 (or        method and compresses the audio
  216 values). These bits can be allocated     data—decompressing the audio on the
  in whatever manner one chooses, pro-         fly if the audio should ever be needed.    e
  vided the resulting split is meaningful.     At its lowest scanning frequency and
  For example, by allocating 8 bits            highest compression, a CD-ROM/XA          f
  to each channel, you can get stereo.         disc can hold over nine hours of stereo   g
  CD-DA was not developed for record-          music—compared to just under 1.25
  ing data. CD-DA is what everyone cor-        hours for CD-DA. In addition, CD-         h
  rectly thinks of as audio CD.                ROM/XA uses a new sector format,
                                               which allows a file to be nested inside
CD-ROM (Read Only Memory) (Yellow
  Book) This standard was designed
                                               another. Even though it uses special      j
                                               hardware, CD-ROM/XA technology
  to enable CD technology to be used
  with computers—and for storing huge
                                               is back compatible with CD-DA and         k
                                               ordinary CD-ROM. (Fortunately, most
  amounts of data. Because error rate
                                               CD drives available today include this
  requirements for data are much more
  stringent than for music, the bits in a
                                               extra hardware, so that these drives      m
                                               can read most kinds of CDs.) CD-
  CD-ROM sector are allocated differ-
                                               ROM discs can hold up to 660 MBytes       n
  ently than for a musical performance.
  Whereas a CD-DA sector has 2352
                                               of data.                                  o
  bytes available for storing music in       Photo-CD This disc format was created
                                               by Kodak to provide a way for cus-
  each sector, CD-ROM has only 2048,
  because 280 extra bits had to be allo-       tomers to digitize their photos and to    q
  cated for error-detection and correc-        use them at work or home. The Photo-
  tion. CD-ROM actually does have a            CD technology combines the XA stan-
  less stringent mode, known as mode           dards with multisession technology. A     s
  2 (in contrast to the mode 1 used            session is a recording period. Origi-
  for data). This makes 2336 bytes             nally, CD’s could record only once,       t
                                               which meant that all data or pictures
  per sector available for use (at the
  cost of a considerable amount of             had to be recorded in a single session.
  error correcting).                           With a multisession disc, on the other    v
                                               hand, a customer can have pictures
CD-ROM/XA (Extended Architecture)
                                               recorded several times up to the disc’s   w
  (Yellow Book and some of the Green
  Book) This standard was designed to
                                               capacity.                                 x
  provide a more efficient and flexible                                                    y
172   CD (Compact Disc)

  CD-WO (Write Once) and CD-MO                  surface. This makes it possible to work
    (Magneto-optical) (Orange Book)             with a much weaker laser. CD-R discs
    These are specifications for recordable      can hold up to 660 MBytes of infor-
    CDs. CD-WO—also known as CD-                mation. These discs are, in essence, just
    WORM (Write once, read many)—               ordinary CD-ROM discs produced by
    is the older standard. It can create        special means.
    discs with capacities of 128 Mbytes,
                                              CD-I (Interactive) (Green Book) This
    650 Mbytes, or 6.5 GBytes, depending
                                                standard allows branching based on
    on the disc’s size. CD-WO discs
                                                interaction between the user and the
    require a magneto-optical drive and
                                                material. CD-I drives connect to a tele-
    are not compatible with CD-ROM
                                                vision set. Any computing capabilities
    technology. CD-MO discs can hold
                                                required to run the software are built
    128-, 230-, 600-, 650-, or 1300
                                                into the drive. You cannot use or even
    MBytes, and they must also be read by
                                                read CD-I discs in ordinary CD-ROM
    a special magneto-optical drive. Unlike
                                                drives. 3DO is a proprietary variant of
    CD-WO, however, CD-MO discs can
                                                the CD-I standard.
    be recorded multiple times. Because of
    this, MO discs are also known as EO       High density CD (Blue Book) This tech-
    (erasable optical) discs.                   nology is still being developed. When
                                                perfected, this standard is expected to
  CD-R (Recordable) (Orange Book) This
                                                increase the capacity of a disc ten-
    is a variant of the WO standard.
                                                fold—to about 6.5 GBytes. Look for
    Unlike CD-WO, however, discs
                                                this technology in the next year or so.
    recorded using CD-R technology can
    be read on ordinary CD-ROM drives.        Hybrid standards Several variants have
    Until recently, CD-R machines were          been developed for special purposes or
    much too expensive for personal use;        to make use of particular technology.
    this has begun to change, and such          In general, such discs require special
    devices are becoming very popular for       hardware. Hybrids include CD+G,
    business use. Discs for use in a CD-R       CD-MIDI, CD-EB, and CD-V. CD+G
    drive are distinguished by their gold       (for graphics) is basically an audio CD
    surface, as opposed to the silvery sur-     with additional information such as
    face of a commercially produced disc.       text or graphics. CD-MIDI (for Musi-
    One reason for this is that CD-R discs      cal Instrument Digital Interface) is an
    are created using a somewhat different      audio disc with MIDI information.
    process than commercial CDs. Instead        CD-EB (for Electronic Book) is special
    of burning pits into the surface, the       size and format that is used mainly to
    recording laser in a CD-R drive simply      store reference materials. CD-V (for
    changes the optical properties of an        video) is an audio disc with video
    organic paint in the disc’s recording       information recorded in analog form.
                                                            CDPD (Cellular Digital Packet Data)    173

          The laserdisc is actually a CD-V                 appropriate spread will be meaningful; the
          variant.                                         other signals will be received as noise.
       The logical structure of the material on a
                                                              CDMA uses a soft-handoff when switch-
                                                           ing a transmission from one cell to another
    CD is defined in the ISO 9660 documents.
    These, in turn, are based on the earlier High
                                                           to ensure that no bits are lost in the trans-    c
                                                           mission. In this type of handoff, both cells
    Sierra specifications.
                                                           transmit the transitional bits at the same
w                                                          time and on the same frequency. This way,        e
CDDI (Copper Distributed Data                              one of the transmissions will be within range
Interface)                                                 of the receiver.                                 f
    A networking configuration that implements                 This method is not compatible with the        g
    the FDDI architecture and protocols on                 TDMA (time division multiple access)
    unshielded twisted-pair (UTP) cable—that               method that was adopted as a standard in         h
    is, on electrical (rather than optical) cable. A
    related implementation is SDDI (shielded
                                                           B RO A D E R C A T E G O R Y
    distributed data interface), which uses                   Cellular Communications                       j
    shielded twisted-pair (STP) cable. Also
    known as copper-stranded distributed data              C O M P A RE                                     k
    interface and as TPDDI (twisted-pair DDI).                TDMA (Time Division Multiple Access)          l
    SEE ALSO                                           w
                                                       CDPD (Cellular Digital Packet Data)
       FDDI (Fiber Distributed Data Interface)
                                                           A cellular communications technology that        n
CDFS (CD-ROM File System)                                  sends digital data over unused cellular
                                                           (voice) channels. CDPD data can be trans-
    A file structure used for storing information           mitted at 19.2 kbps, but only in service areas   p
    on a compact disc. The file allocation table            that support CDPD. Currently only a few
    (FAT) system may not be efficient or even               dozen of the major service areas around the
    feasible for such a disc because of the large
    number of files the disc may contain.
                                                           country provide direct CDPD support.             r
                                                              CDPD can be used as a mobile computing
w                                                          strategy to stay connected with the company      s
                                                           network back at the office. Essentially, a
CDMA (Code Division Multiple Access)
                                                           mobile user needs a special CDPD modem
    In cellular communications, CDMA is a pro-
    posed transmission method that uses special
                                                           and the appropriate software. The user gets      u
                                                           an IP (Internet protocol) address, which
    codes to fit up to ten times as much informa-
                                                           makes it possible to communicate as well as      v
    tion into a channel. Each signal that comes
    in on a given frequency is “spread” using a
                                                           to make use of Internet services.                w
                                                              Mobile users can remain connected even
    different code. When the receiver decodes
    the received signals, only the signal with the
                                                           when they are not using their computers and      x
                                                           even when they are outside the range of a
174       CD-ROM Drive

    cell that supports CDPD. The CDPD specifi-      which have a huge capacity (660
    cations support a “sleep” mode for the com-    megabytes).
    puter. The network signals periodically to        Several features distinguish CD-ROM
    sleeping devices, and a device will “wake”     drives from each other:
    if the signal includes the device’s name or
                                                     s   Transfer rate, which represents the
    address. The monitoring for each device is
                                                         amount of data that the drive can read
    done by the MDIS (mobile data intermediate
                                                         from the disc in a second. Speeds are
                                                         based on a base rate of 150 kbytes
        Similarly, the MDIS allows a user to
                                                         per second, which is known as a
    remain connected even beyond areas that
                                                         single-speed drive. Double speed and
    support CDPD through a technology known
                                                         quad-speed drives can transfer 300
    as switched CDPD. If the user is outside a
                                                         and 600 kbytes per second, respec-
    service area with CDPD capabilities when
                                                         tively. Quad speed drives are the
    called, the MDIS opens a circuit-switched
                                                         current norm, but 6x drives (not
    connection over the channel. The connection
                                                         yet known as “hex speed”) are also
    is circuit-switched as far as the cellular
    network is concerned, but is essentially
    packet-switched as far as the device is con-     s   Access time, which represents the aver-
    cerned. This is because the MDIS closes the          age time it takes to find a specified item
    connection whenever there is silence, and            of information on the disc. Currently,
    reopens it whenever there is activity.               access times of less than 200 msec are
        CDPD supports data compression and               considered standard.
    encryption. This cuts down on transmission
                                                     s   Compatibility with various CD
    times (and costs) and also helps keep snoop-
                                                         standards, which indicates the types
    ers from getting access to the data. In the
                                                         of CDs the drive can read. The CD
    CDPD specification, the data are first
                                                         (compact disc) article summarizes
    compressed and then encrypted.
                                                         these. Briefly, drives should be able to
        The CDPD specification is being
                                                         read CD-XA (extended architecture)
    formulated under the auspices of the
                                                         discs and should support multisession
    CDPD Forum, which you can contact at
    info@forum.cdpd.net or at 800-335-CDPD
    (2373).                                          s   Number of discs the drive can handle.
                                                         Multidisc systems can hold 3, 6, or
                                                         even 18 discs, and can switch between
      Cellular Communications
                                                         them within a few seconds. The drive
w                                                        can only read one disc at a time,
CD-ROM Drive                                             however.
    CD-ROM stands for compact-disc,                   A CD-ROM drive may be connected
    read-only memory. A CD-ROM drive               to a network, making any available CDs
    is a peripheral device for reading CDs,
                                                                            Cell Loss Priority   175

shareable resources. With the appropriate            w
server and drivers, users can share access to        CD-ROM File System (CDFS)                          a
the disc currently loaded in the drive. A CD-
ROM drive can be accessed just like any
                                                         SEE                                            b
                                                           CDFS (CD-ROM File System)
other volume, except that you can only read                                                             c
from it. If there are licensing restrictions on      w
the use of a disc, it is essential that the server   Cell                                               d
software be able to restrict simultaneous                In communications or networking, a packet,     e
access to the licensed number of users.                  or frame, of fixed size. In general, fast
   Like any other type of hardware device,               packet-switching technologies—such as          f
CD-ROM drives require hardware drivers to                ATM (Asynchronous Transfer Mode) and           g
communicate. In addition, a special driver               SDMS (Switched Multimegabit Digital Ser-
containing extensions is required. These                 vice)—use cells. Slower packet-switching       h
extensions are specific to the operating                  technologies—such as X.25—are more
system, such as DOS, OS/2, or NT, with                   likely to use variable-sized packets.
which the CD-ROM is working. Microsoft                      In cellular communications, a cell refers   j
has provided such a driver for MS DOS,                   to a geographic area. Each cell has its own
called MSCDEX, which can be used with                    transmitter and receiver, through which sig-   k
most CD-ROM drives. Some hardware
manufacturers have also created their own
                                                         nals can be distributed throughout the cell.
                                                         Transmissions must be “handed off” from
proprietary drivers. If you are connecting a             one cell to another when a mobile telephone    m
CD-ROM drive to a workstation, you will                  or networking caller actually moves from
need to load both the driver’s regular hard-             one cell to another.
ware driver and either MSCDEX or the
manufacturer’s own extensions driver.
   If you want to make a CD-ROM drive
                                                     Cell, ATM                                          p
                                                         In the broadband ATM (Asynchronous
available as a shared volume on a NetWare
3.12 or NetWare 4.x network, you do not                  Transfer Mode) network architecture, cell
load the MSCDEX driver. Instead, load the                refers to a packet. ATM cells are each 53      r
CD-ROM driver’s regular hardware drivers                 octets, of which five octets are header and
and Novell’s CDROM.NLM. This NLM                         48 are data.                                   s
manages the interface between the drive and
NetWare and enables the CD-ROM device
                                                         SEE ALSO                                       t
                                                           ATM (Asynchronous Transfer Mode)
to be viewed and accessed by multiple users,                                                            u
just like any other NetWare volume.                  w
   Note that the drivers available for a given       Cell Loss Priority                                 v
CD-ROM drive may or may not work with                    In an ATM network, a bit value that            w
your system. Verify that the drive is compat-            specifies whether a cell can be discarded
ible before you install it.                              if advisable; for example, if the network      x
176       Cellular Communications

    gets too busy. A value of 1 indicates an          w
    expendable cell.                                  Central Office
                                                          The telephone switching station nearest to a
                                                          customer (residential or business). Custom-
      ATM (Asynchronous Transfer Mode)
                                                          ers are connected directly to a CO, which
w                                                         connects them to other points in the tele-
Cellular Communications                                   communications hierarchy. The CO pro-
    Cellular communications is a wireless com-            vides services such as switching, dial tone,
    munications technology. The communica-                private lines, and centrex.
    tions area is divided into smaller areas,
    called cells, and transmissions are passed        Central Processing
    from cell to cell until they reach their desti-
                                                          Central processing, also known as central-
    nations. Each cell contains an antenna and
                                                          ized processing, is a network configuration
    transmission facilities to pick up signals
                                                          in which a single server processes tasks for
    from another cell or from a caller and to
                                                          multiple stations, all of which can commu-
    pass them on to an adjacent cell or to a
                                                          nicate with the server. In such a setup, the
    callee within the cell. Cells can be anywhere
                                                          nodes must share the computing power of
    from a few kilometers to 32 kilometers
                                                          the central processor. One consequence is
    (20 miles) in diameter.
                                                          that the more tasks, the slower things get
       One cellular communications method,
    called CDPD (Cellular Digital Packet Data)
                                                             Central processing can be compared with
    transmits data over any cellular channels
                                                          distributed processing, in which tasks are
    that are not being used. CDPD uses tele-
                                                          performed by specialized nodes somewhere
    phone (voice) channels, but can switch to a
                                                          on a network. A station that needs some-
    new frequency, if necessary, when a voice
                                                          thing done sends a request onto the net-
    transmission begins in the cell being used.
                                                          work. The server responsible for the service
    CDPD was developed to provide data com-
                                                          takes on the task, does it, and returns the
    munications in the cellular frequency range
                                                          results to the station. The client station need
    without interfering with voice calls.
                                                          never know who actually did the work.
CELP (Code Excited Linear Predictive                  w

Coding)                                               CERT (Computer Emergency
                                                      Response Team)
    A variant of the LPC voice encoding algo-
                                                          In the Internet community, CERT is a
    rithm. CELP can produce digitized voice
                                                          group formed in 1988 (by DARPA) to help
    output at 4,800 bits per second.
                                                          respond to, and deal with security problems
    SEE ALSO                                              that may arise on the Internet. The group
      LPC (Linear Predictive Coding)                      also provides Internet administrators with
                                                                                      Character    177

    information and assistance to help avoid             P R I M A R Y S O U RC E S
    security problems.                                      You can find the current form of the CGI          a
       Tools and documents related to network
    security are available through Anonymous
                                                            specifications at
    FTP from CERT’s database in cert.org.                      overview.html
    See the Protocol, FTP article for more
    information.                                     w
    B RO A D E R C A T E G O R Y
                                                     Channel                                                 e
                                                         A channel is a physical or logical path for a
       Network Security
                                                         signal transmission. Two particularly impor-        f
w                                                        tant channels in networking are the commu-          g
CFB (Cipher Feedback)                                    nications channel and the disk channel.
    An operating mode for the DES.                          A communications channel is a path               h
                                                         through which data or voice can be trans-
    SEE                                                  mitted; for example, in a network or a
       DES (Data Encryption Standard)                    telephone call. In telecommunications, a            j
                                                         single cable may be able to provide multiple
CGI (Common Gateway Interface)                           channels.                                           k
    An interface specification that defines the
                                                            A disk channel, in a hard-disk configura-
                                                         tion, consists of the components that con-
    rules of communication between informa-
    tion servers, such as HTTP (Hypertext
                                                         nect a hard disk drive to an operating              m
                                                         environment, such as DOS, OS/2, NetWare,
    Transport Protocol) servers on the World
                                                         or VINES. These components include cables
    Wide Web and gateway programs. More
    specifically, the CGI is used when such a
                                                         and a hard disk adapter or controller. A sin-       o
                                                         gle channel can accommodate multiple hard
    server needs to pass a user request to a gate-
                                                         disks. A computer may have multiple disk            p
    way program. Being able to pass work off to
    the gateway program helps take some of the
                                                         channels.                                           q
    workload off the server.
       The gateway program is generally
                                                         C O M P A RE
    designed to provide a mechanism for getting
    input from a user—for example, so an
    authorized user can complete an authentica-
                                                     Channel Bank                                            t
                                                         A device that multiplexes low-speed signals
    tion form in order to get access to restricted
                                                         into a single high-speed signal.
    areas. Among other things, the CGI specifi-
    cations define the mechanisms by which            w
    information can pass from the server to the
    gateway program and back.
                                                     Character                                               w
                                                         A byte with an identity. A group of bits, usu-
       The CGI specifications, along with many
                                                         ally, seven or eight bits, that represents a sin-   x
    of the other specifications related to HTTP
    environments, are still undergoing revisions.
                                                         gle letter, digit, special symbol, or control       y
178       Checksum

    code in an encoding scheme, such as ASCII        w
    or EBCDIC.                                       CHRP (Common Hardware Reference
Checksum                                                 A set of specifications for PowerPC systems.
                                                         CHRP is being developed by Apple, IBM,
    Checksum is a simple error-detection strat-
                                                         and Motorola, and is designed to enable
    egy that computes a running total based on
                                                         such a machine to run multiple operating
    the byte values transmitted in a packet, and
                                                         systems and cross-platform applications.
    then applies a simple operation to compute
                                                            While specifications have not been final-
    the checksum value.
                                                         ized, a minimum machine will have at least
        Checksums are very fast and easy to
                                                         8 MB of RAM and a 1 MB cache; CHRP
    implement, and they can detect about
                                                         machines will use the PowerPC 604 or later
    99.6 percent of errors in a packet. This
                                                         chip, and will support the PCI (Peripheral
    reliability level is acceptable for most sim-
                                                         Component Interconnect) bus standard.
    ple communications situations, but is less
                                                         CHRP machines will support at least the
    reliable than the more sophisticated CRC
                                                         following environments:
    (Cyclical Redundancy Check) calculations,
    which have an accuracy of more than 99.9               s   AIX (IBM’s UNIX port)
    percent.                                               s   IBM OS/2 for PowerPC
        The receiver compares the checksums
    computed by the sender and by the receiver.            s   Mac OS (Apple’s new Macintosh
    If they match, the receiver assumes the trans-             operating system)
    mission was error-free. If they do not match,          s   Novell NetWare
    there was an error.
                                                           s   Solaris (from SunSoft)
    B RO A D E R C A T E G O R Y
       Error Detection and Correction                      s   Microsoft Windows NT

    C O M P A RE                                     w
       CRC (Cyclical Redundancy Check);              CICS (Customer Information Control
       Parity                                        System)
                                                         A terminal that provides transaction pro-
                                                         cessing capabilities for IBM mainframes.
Chromatic Dispersion
                                                         CICS supports the SNA (Systems Network
    In a fiber-optic transmission, the dispersion         Architecture).
    of a light signal because of the different
    propagation speeds of the light at different     w
    wavelengths; also known as material disper-      CIDR (Classless Interdomain Routing)
    sion. The wavelengths around which disper-           CIDR is a routing strategy that was devel-
    sion is minimal, such as those around 1300           oped as a partial solution to two difficulties
    or 830 nanometers, are commonly used for             that have developed as the number of
                                                  CIDR (Classless Interdomain Routing)    179

networks connected to the Internet has            of Class C networks as belonging to the
grown very large. One problem was that            same “supernetwork”—as indicated by their
routers had to deal with too many network
addresses and were choking on their routing
                                                  common value in the higher order address
                                                  bits (known as the prefix bits in this con-
tables. The second problem was that the           text). By routing just on the (fewer) higher-    c
supply of Class B network addresses was           order bits, routers can fulfill their functions
being used up too quickly. Class B networks       without having to store all the networks to
can have up to 65,536 hosts, but there can        which they are routing.                          e
be only 16,384 Class B network addresses.            For CIDR to be successful, several things
This address class (see IP Address for a more     are required:                                    f
detailed discussion) is useful for companies
                                                    s   The internal and external gateway pro-     g
or organizations that have large networks
                                                        tocols need to be able to represent the
with thousands of hosts for each network.
                                                        “supernetwork cluster” groupings.
While there are many companies with a few
thousand hosts on their networks, there are
                                                        Earlier gateway protocols (such as         i
                                                        BGP-3, IGRP, and RIP-1) cannot do
few that have anywhere near 65,000.
                                                        this; newer versions (such as BGP-4,       j
   Because of the way address classes are
defined, this situation leads to a lot of poten-
                                                        EIGRP, IS-IS, OSPF, and RIP-2) can.
                                                        The protocol situation is in transition
tial addresses being wasted. The next
address class—C—supports networks with
                                                        because newer protocols are, in some       l
                                                        cases, just becoming available.
256 or fewer hosts. There can be more than                                                         m
2 million Class C addresses. So, whereas            s   Class C addresses must be assigned
Class B address spaces are too big, those               consecutively, as assumed in the CIDR      n
for Class C are somewhat small for many                 strategy. While this can be done easily    o
businesses and organizations. When a mid-               in some areas, it’s much more difficult
size company asks for an Internet address, it           in others. One important and sticky        p
must be given either a Class B address from
a dwindling supply or several (perhaps sev-
                                                        issue is how to deal with address
                                                        owners who move, as such a move
eral dozen) Class C addresses. For example,             could entail a switch in providers,        r
a company with just over 8,000 hosts would              which would undoubtedly lead to
need 32 Class C addresses. In contrast, by              routing changes. If the address that’s     s
taking a Class B address, it would waste
more than 55,000 potential addresses.
                                                        moving happens to be in the middle of
                                                        a “supernetwork,” the abbreviated
   CIDR is designed to make a happy                     addressing scheme falls apart.             u
medium possible by assigning consecutive
Class C addresses to organizations or corpo-
                                                    s   An effective strategy must be worked       v
                                                        out for assigning addresses. Two possi-
rations that have more than 256 machines,
                                                        ble basic approaches are provider-         w
but that may not be large enough to merit a
Class B address. CIDR takes advantage of
                                                        based and geographically based. In the     x
                                                        former, networks that share a provider
the assignment scheme and treats the cluster                                                       y
180       Cipher Block Chaining (CBC)

          get addresses close to each other,               will mark the user’s extra packets to indicate
          regardless of whether these networks             that they can be discarded if necessary.
          are physically near each other. The
          geographically based approach would
          assign addresses within a block to net-
          works in the same geographical area.             A closed path through which electricity
          The current Internet is closer to the            can flow. The term is also used to refer to
          provider-based variant.                          components (such as chips) capable of creat-
                                                           ing such a path.
Cipher Block Chaining (CBC)                            w
                                                       CIS (CompuServe Information
    An operating mode for the DES.
    SEE                                                    CIS, better known simply as CompuServe, is
       DES (Data Encryption Standard)                      the oldest of the major online services, and
                                                           is still the largest, although America Online,
Cipher Feedback (CFB)                                      or AOL, is gaining rapidly. CompuServe
                                                           supports DOS, Windows, and Macintosh
    An operating mode for the DES.
                                                           users. It offers the usual forums, electronic
    SEE                                                    mail, financial and news services, and soft-
       DES (Data Encryption Standard)                      ware to download or use online. For a flat
                                                           monthly fee, users have unlimited access to
w                                                          basic services; special services incur addi-
                                                           tional fees. For a fee, users can also get
    Text that has been encrypted to make it                access to the Internet.
    unintelligible to anyone who lacks essential
                                                           SEE ALSO
    information about the encryption scheme.
    The required information is generally a spe-              AOL (America Online); Prodigy
    cific value, known as the encryption (or                FO R I N F O R M A T I O N
    decryption) key. Conventional-, public-,                  Call (800) 848-8199
    or private-key encryption strategies may
    be used to create ciphertext.                      w
                                                       CISC (Complex Instruction Set
    SEE ALSO                                           Computing)
                                                           CISC is a processor design strategy that pro-
w                                                          vides the processor with a relatively large
CIR (Committed Information Rate)                           number of basic instructions, many of which
    In frame-relay networks, a bandwidth, or               are complex but very powerful. These com-
    information rate, that represents the average          plex instructions may require several clock
    level for a user. If the user’s network activity       cycles to complete, which can slow down
    exceeds this rate, the frame-relay controller          overall processing.
                                                                                        Client   181

       CISC is in contrast to the RISC (reduced       w
    instruction set computing) design strategy. A     Class A Certification                                a
    RISC chip uses a small number of sim-
    ple operations to do its work. These simple
                                                          An FCC certification for computer or other       b
                                                          equipment intended for industrial, commer-
    operations are optimized for speed, and               cial, or office use, rather than for personal    c
    most require only a single clock cycle for
                                                          use at home. The Class A commercial certifi-
                                                          cation is less restrictive than the Class B
                                                          certification.                                   e
CIX (Commercial Internet Exchange)                    w                                                   f
    CIX is an association of domestic Internet        Class B Certification
    access providers that provides connection             An FCC certification for computer equip-
    points between commercial traffic and the              ment, including PCs, laptops, and portables     h
    Internet. The CIX was formed to route                 intended for use in the home rather than in a
    commercial traffic back when such traffic               commercial setting. Class B certification is     i
    was not allowed according to the AUP
    (acceptable use policy) for the Internet. CIX
                                                          more restrictive than the commercial Class      j
                                                          A certification.
    members agree to carry each others’ traffic
    when requested. Contact Gopher or Web
    servers at cix.org for more information
                                                      Clearing Center (CC)                                l
    about CIX.                                            SEE                                             m
                                                            CC (Clearing Center)
    In fiber-optic cable, the material (usually        CLID (Calling Line Identification)                   o
    plastic or glass) surrounding the fiber core.
    The cladding has a lower index of refraction
                                                          In ISDN and some other telecommunica-           p
                                                          tions environments, a feature that includes
    than the core, which means that light hitting         the sender’s identification number (such as      q
    the cladding will be reflected back into the           telephone number) in the transmission so
    core to continue its path along the cable.            that the receiver knows who is calling. It
                                                          is also known as ANI (automatic number          s
                                                          identification) and caller ID.
      Cable, Fiber-Optic                                                                                  t
Clamping Time
                                                      Client                                              u
    In power protection, the amount of time
                                                          A client is a machine that makes requests of    v
                                                          other machines (servers) in a network or
    needed for a surge protector to deal with
    a voltage spike or surge; that is, to bring the
                                                          that uses resources available through the       w
    voltage within acceptable levels.
                                                             For example, workstations are network
                                                          clients because they use services from the      y
182       Client-Based Application

    server. As another example, a client applica-   provides the user with an interface for giving
    tion is an application that makes requests      commands and making requests. The appli-
    of other applications, on the same or on        cation’s real work is done by the back end,
    different machines, for services, informa-      which processes and carries out the user’s
    tion, or access to resources.                   commands.
                                                       Client/server computing allows for sev-
    C O M P A RE
                                                    eral types of relationships between the server
                                                    and client, including the following:
                                                      s   Stand-alone (non-networked) client
      Workstation                                         applications which do not request
w                                                         access to server resources. For
Client-Based Application                                  example, a local word processor might
                                                          be a stand-alone client application.
    An application that executes on the client
    machine (the workstation) in a network.           s   Applications that run on the client but
                                                          request data from the server. For
Client/Server Computing                                   example, a spreadsheet program might
                                                          run on a workstation and use files
    Client/server computing is a network-                 stored on the server.
    ing arrangement with the following
    characteristics:                                  s   Programs where the physical search of
                                                          records takes place on the server, while
      s   Intelligence, defined either as process-         a much smaller program running on
          ing capabilities or available informa-          the client handles all user-interface
          tion, is distributed across multiple            functions. For example, a database
          machines.                                       application might run this way on the
      s   Certain machines—the clients—can                server and client.
          request services and information from       s   Programs that use server capabilities to
          other machines—the servers. For                 share information between network
          example, a server may have quick                users. For example, an electronic-mail
          access to huge databases that can be            system may use the server this way.
          searched on behalf of the client.
                                                       The figure “Client/server computing
      s   The server does at least some of the      arrangements” illustrates these different
          processing for the client.                arrangements.
       Applications capable of running in a         SEE ALSO
    client/server environment can be split into a
                                                      Back End; Front End
    front end that runs on the client and a back
    end that runs on the server. The front end
                                                                                          Clock Speed   183

    C L I E N T / S E R VE R C O M P U T I N G A R RA N G E M E N T S                                          a
w                                                              contrast, for wide-area networks (WANs),
CLNS (Connectionless Mode                                      CONS (connection-oriented network               p
Network Service)                                               service) is more popular.
    In the OSI Reference Model, CLNS is a
                                                               P R I M A R Y S O U RC E
    network-layer service in which data                           ISO document 8348                            r
    transmission can take place without a fixed
    connection between source and destination.                 B RO A D E R C A T E G O R Y                    s
    Individual packets are independent, and they
    may reach the destination through different
                                                                  Connectionless Service
    paths and in a mixed order. In this type of
                                                               C O M P A RE
                                                                  Connection-Oriented Service
    transmission service, each packet must carry
    its own destination address and information            w
    about the packet’s relative position in the             Clock Speed                                        w
    message.                                                   Activities carried out by and for the proces-
        CLNS is the most common operating                      sor must all be carefully timed and coordi-
    mode for local-area networks (LANs). In                    nated. To make this possible, each processor    y
184       CLP (Cell Loss Priority)

    has a clock associated with it. This clock     w
    serves as a timing reference by slicing time   Cluster
    into very short intervals. The clock speed         In a network, particularly in a mainframe-
    is defined as the number of such slices in          based network, a group of I/O (input/ouput)
    a second.                                          devices, such as terminals, computers, or
       Clock speed is expressed in millions of         printers, that share a common communica-
    cycles per second (megahertz, or MHz). For         tion path to a host machine. Communica-
    example, the CPU in the original IBM had a         tions between the devices in a cluster and the
    clock speed of 4.77 MHz. This is painfully         host are generally managed by a cluster con-
    slow when compared to today’s processors,          troller, such as IBM’s 3274 controller.
    with clock speeds that can be in the 100
    MHz range.                                     w
                                                   Cluster Controller
                                                       A device that serves as an intermediary
CLP (Cell Loss Priority)
                                                       between a host machine, such as a
    In an ATM network, a bit value that                mainframe, and a group (cluster) of I/O
    specifies whether a cell can be discarded           (input/ouput) devices, such as terminals,
    if advisable; for example, if the network          computers, or printers. The IBM 3274 is an
    gets too busy. A value of 1 indicates an           example of such a device. This controller has
    expendable cell.                                   been superseded by the 3174 establishment
      ATM (Asynchronous Transfer Mode)             w
                                                   CMC (Common Mail Calls)
CLTS (Connectionless Transport                         An API (Application Program Interface)
Service)                                               developed by the X.400 API Association
                                                       (XAPIA) to enable message-handling
    In the OSI Reference Model, a transport-
                                                       agents—for example, in an email system—to
    layer service that does not guarantee deliv-
                                                       communicate with message stores, or post
    ery, but makes a best effort, does error
                                                       offices. The calls in the API are designed
    checking, and uses end-to-end addressing.
                                                       to be independent of hardware platforms,
w                                                      operating systems, email systems, and mes-
CLU (Command Line Utility)                             saging protocols. The API is also referred
    In Novell’s NetWare and in other operating         to as common messaging calls.
    and networking environments, a program
    that can be executed at the appropriate        CMIP (Common Management
    command-line prompt. Examples of com-          Information Protocol)
    mand line utilities in NetWare include
                                                       A network management protocol for the
    NCOPY and FLAG for manipulating
                                                       OSI Reference Model. CMIP, pronounced
    files and file attributes, respectively.
                                                       “see-mip,” defines how management
                                 CMISE (Common Management Information Service Element)        185

    information can be communicated between              control services. CMIS, pronounced “see-
    stations. CMIP is functionally comparable            miss,” is documented in CCITT recommen-
    to the older, and arguably more widely used,
    SNMP (Simple Network Management
                                                         dation X.710 and ISO document 9595.          b
                                                         SEE ALSO
                                                           CMISE; Network Management
      Network Management
                                                     CMISE (Common Management                         e
w                                                    Information Service Element)
CMIPDU (Common Management                                In the OSI network management model, a
Information Protocol Data Unit)                          CMISE is an entity that provides network     g
    In the OSI network management model,                 management and control services. Seven
    a packet that conforms to the CMIP. The              types of CMISEs, pronounced “see-mize,”      h
    packet’s contents depend on the requests             are specified:                                i
    from a CMISE, which relies on the CMIP to
    deliver the user’s requests and to return with
                                                           s   Event report
    answers from the appropriate application or                Get
                                                           s   Cancel get
                                                           s   Set
      CMISE (Common Management                                                                        m
      Information Service Element);                        s   Action
      Network Management                                   s   Delete
w                                                          s   Create
CMIPM (Common Management
Information Protocol Machine)                               The services provided by CMISEs are
    In the OSI network management model,                 used by the system management functions      q
    software that accepts operations from a              (SMFs). The SMFs are in turn used to carry
    CMISE user and initiates the actions needed          out the tasks specified for the five system    r
    to respond and sends valid CMIPDUs                   management functional areas (SMFAs)
                                                         defined in the OSI network management
    (CMIP packets) to a CMISE user.
                                                         model. The figure “Major components in        t
    SEE ALSO                                             the ISO-OSI network management model”
      CMISE; Network Management                          shows this relationship.                     u
w                                                        SEE ALSO                                     v
CMIS (Common Management
Information Service)
                                                           Network Management                         w
    In the OSI network management model,                                                              x
    a standard for network monitoring and
186       CMOS (Complementary Metal-Oxide Semiconductor)

    M A J O R C O M P O N E N T S I N T H E I S O - O S I N E T WO R K M A N A G E M E N T M O D E L

w                                                           maintained by battery power, and they
CMOS (Complementary Metal-Oxide                             are generally not accessible to the operating
Semiconductor)                                              system.
    CMOS, pronounced “see-moss,” is a logic                 C O M P A RE
    family for digital circuits. CMOS logic is not             TTL (Transistor-Transistor Logic)
    exceptionally fast, but it has relatively low
    power consumption, which makes it ideal             w
    for such items as battery-powered PCs.               CMOT (Common Management
       CMOS is used for RAM chips that                   Information Services and Protocol
    need to retain information, such as configu-          Over TCP/IP)
    ration data or date and time information.               An effort to implement the OSI framework’s
    The values stored in these RAM chips are                CMIS and CMIP services on the Internet
                                                         CNI (Certified NetWare Instructor)     187

    community’s TCP/IP protocol suite, rather           SEE ALSO
    than on OSI layer protocols. For various              CNE; CNI; ECNE                                 a
    reasons, including the popularity of SNMP
    and the difficulty of porting the OSI model
                                                    w                                                    b
                                                    CNE (Certified NetWare Engineer)
    to a TCP/IP environment, CMOT was never                                                              c
    completed.                                          A title given to people who successfully com-
                                                        plete a whole series of Novell-authorized        d
CMS (Conversational Monitor System)
                                                        courses on becoming technicians or consul-
                                                        tants for NetWare networks and/or pass a
    A subsystem in IBM’s SNA.                           comprehensive exam about this topic. The         f
                                                        CNE program is designed for people who
                                                        are responsible for designing and installing
      SNA (Systems Network Architecture)
                                                        NetWare networks, and also for the low-          h
w                                                       level maintenance tasks such as diagnostics,
CN (Common Name)                                        troubleshooting hardware or networking           i
                                                        software, and so forth. Separate tracks are
    In the NetWare Directory Services (NDS) for
                                                        available for NetWare 2.2, 3.11, and 4.x. In
    Novell’s NetWare 4.x, a name associated
    with a leaf object in the NDS Directory tree.       addition to demonstrating mastery of basic       k
                                                        and advanced topics related to NetWare,
    For a user object, this would be the user’s
    login name.                                         successful CNE candidates must demon-            l
                                                        strate mastery of networking technology
                                                        and operating system concepts.
CNA (Certified NetWare
                                                        SEE ALSO
    A title given to people who successfully com-
                                                          CNA; CNI; ECNE                                 o
    plete Novell-authorized courses on adminis-     w                                                    p
    tering a NetWare network and/or pass a          CNI (Certified NetWare Instructor)
    comprehensive exam about this topic. The            A title given to people who successfully com-
    CNA program is designed for people who              plete a comprehensive and rigorous training      r
    are responsible for the day-to-day opera-           program in order to teach Novell courses.
    tions and high-level maintenance of their           Candidates who are accepted for the CNI          s
    networks. CNAs must know how to add
    and remove users, grant user rights, load
                                                        program must demonstrate a proficiency
                                                        in their area of specialization by attending
    applications, do backups and other mainte-          each course they want to teach and passing       u
    nance tasks, and maintain network security.         the course test at a more stringent level than
    Separate tests are required and degrees are         is required of ordinary (CNA or CNE)
    offered for NetWare 2.2, 3.11, and 4.x
                                                        students.                                        w
                                                            As a final requirement, candidates must
       CNAs are discussed in The CNA Study              pass an IPE (instructor performance evalua-      x
    Guide (James Chellis, et al. Network Press,
                                                        tion). Among other things, candidates must       y
188       CO (Central Office)

    set up a classroom or lab, and then teach a     are used in digital telephone systems, such as
    45- to 60-minute section of the course for      ISDN (Integrated Services Digital Network),
    which the candidate wants to become an          so that voice signals can be transmitted over
    instructor. Candidates do not know which        digital lines. The name is a contraction of
    section they will be asked to teach until the   coder/decoder.
    day before their evaluation.                        To make the conversion, a codec must
                                                    use some type of signal-sampling technique.
                                                    These samples are converted into discrete
      CNA; CNE; ECNE
                                                    signals for transmission across the digital
w                                                   lines.
CO (Central Office)                                      The most common conversion method
    The telephone switching station nearest to a    is PAM (pulse amplitude modulation), in
    customer (residential or business). Custom-     which samples of the analog signal’s ampli-
    ers are connected directly to a CO, which       tude are converted into discrete signals
    connects them to other points in the tele-      whose amplitude corresponds to the analog
    communications hierarchy. The CO pro-           signal’s amplitude at sampling time. To
    vides services such as switching, dial tone,    reproduce the original signal accurately,
    private lines, and centrex.                     PAM devices must sample the analog signal
                                                    at a rate at least twice the frequency’s signal.
w                                                   For example, for voice signals, which have a
Coax Booster                                        4 kilohertz bandwidth, the PAM device must
    A device that strengthens the signal in a       sample at least 8,000 times.
    coaxial cable, thereby making it possible           The discrete amplitude value is modu-
    to run a cable over greater distances.          lated one more time to make it compatible
                                                    with the digital circuits. PCM (pulse code
                                                    modulation) converts the PAM signals into a
COCF (Connection-Oriented
                                                    stream of binary values. To make this con-
Convergence Function)
                                                    version, the range of amplitudes in a PAM
    In the DQDB (Distributed Queue Dual-            signal is divided into 128 discrete quantizing
    Bus) network architecture, a function that      levels.
    prepares data coming from or going to a             To represent 128 possible amplitude val-
    connection-oriented service. The service first   ues, seven bits are needed for each PAM sig-
    establishes a fixed, but temporary, connec-      nal. This means that PCM must work at 56
    tion, then transmits the data, and finally       kilobits per second (kbps) or faster. Digital
    breaks the connection.                          channels in North America provide a 64
                                                    kbps capacity, which means 8 kbps can be
Codec                                               used for administrative and system control
    A codec is a device for converting analog
    signals to digital form. For example, codecs
                                                           Collision Detection and Avoidance    189

    B RO A D E R C A T E G O R Y                          Transfer: A coding used by the applica-
       Digital Communication                                tions at both ends of the connection or
                                                            by the translation program. Transfer
                                                            coding may be needed if the source
                                                            and target codings are different.            c
Code Excited Linear Predictive Coding
                                                        SEE ALSO                                         d
                                                          ASCII; ASN.1; EBCDIC
(CELP)                                                                                                   e
       CELP (Code Excited Linear Predictive
                                                    Cold Boot Loader                                     f
       Coding)                                          In Novell’s NetWare, a program on the file        g
                                                        server’s hard disk that will automatically
w                                                       load NetWare after a cold boot.                  h
    Coding is a general term for a representa-      w
                                                    Collision Detection and Avoidance
    tion, usually by means of a predefined syn-                                                           j
    tax or language. For example, in the OSI            In an Ethernet network, a collision is the
    Reference Model, an application layer               simultaneous presence of signals from two        k
    packet, or protocol data unit (APDU), will          nodes on the network. A collision can occur
                                                        when two nodes each think the network
    have a coding that depends on the applica-
    tion involved.                                      is idle and both start transmitting at the       m
                                                        same time. Both packets involved in a colli-
       ASCII and EBCDIC are two widely used
    codings. Abstract Syntax Notation One               sion are broken into fragments and must be
    (ASN.1) coding is used in many contexts             retransmitted.                                   o
    that adhere to the OSI Reference Model,
    such as in network management tasks.
                                                    Collision Detection                                  p
       In a communications setting, several             To detect a collision, nodes check the DC        q
    types of coding are distinguished, and each         voltage level on the line. A voltage level two
    type may occur dozens of times:                     or more times as high as the expected level      r
       Source: The coding used by the applica-          indicates a collision, since this means there
                                                        are multiple signals traveling along the wires
         tion that initiates a transmission. That
         application must be running on an end          at the same time. Collision detection in         t
                                                        broadband networks involves a separate
         system—that is, on a network node
                                                        bandwidth for collision detection and is
         capable of using all seven layers in the
         OSI Reference Model.                           somewhat more complex, since there may           v
                                                        not be any DC voltage to test.
       Target: The coding used by the applica-             In the CSMA/CD (carrier sense multiple        w
         tion that receives a transmission. The
         receiving application must be running
                                                        access/collision detection) media-access
                                                        method, for example, collision detection
         on an end system.                              involves monitoring the transmission line for    y
190        COM (Common Object Model)

    special signals that indicate that two packets            To provide the cross-platform capabili-
    were sent onto the network at the same time            ties, COM uses OLE COM, a protocol
    and have collided. When this happens, spe-             based on the DCE/RPC (Distributed
    cial actions are taken (as described in the            Computing Environment/Remote Procedure
    CSMA/CD article).