Channel Access Methods by 22zklI

VIEWS: 14 PAGES: 27

									        Channel Access Methods
When several devices are connected to a single channel, there
must be some rules to govern these devices when they access,
transmit, and release the channel .
There are three basic channel access methods to provide for an
orderly and efficient use of that capacity:

   • Contention
   • Polling
   • Token passing

Different access methods have different overhead effects on
network traffic.
                     Contention
• With contention systems, network devices may transmit
  whenever they want.
• With this technique, no control is exercised to determine whose
   turn it is, all stations contend for time. No referee mandates
   when a device may or may not use the channel.
• This scheme is simple to design.
• The scheme provides equal access rights to all stations.
• Stations simply transmit whenever they are ready, without
  considering what other stations are doing.
• Unfortunately, the "transmit whenever ready" strategy has one
  important shortcoming.
• Stations can transmit at the same time.
• When this happens, the resulting co-mingling of signals usually
  damages both to the point that a frame's information is lost.
• This unhappy event is called a "collision."
                          Collision
• Newer contention protocols were developed that called for stations
  to listen to the channel first before transmitting.
• If the listening station detects a signal, it will not start
   transmitting and try again later.
• These protocols are called CSMA (Carrier Sense, Multiple Access
  with collision detection) protocols.
• These protocols will reduce collisions.
• However, collisions may still occur when two stations sense the
  cable, detect nothing, and subsequently transmit.
• In order to reduce collisions, CSMA/CD protocols compute a
  random backoff time before retransmitting the frame (as shown in
  the flow diagram).

 Examples of CSMA/CD protocols :
IEEE 802.3 (Ethernet )
CSMA / CD collision schematic
Frame format of a CSMA/CD bus Networks
                Preamble          7 octets

                   SFD            1 octet
            Destination address   2 or 6 octets
              Source address      2 or 6 octets
             Length indicator     2 octets
                   Data           2- octet field
                                  <= 1500


              Pad (optional)

               Frame Check         4 octets
                 Sequence
Operational parameters of a CSMA/CD
            bus Networks
Bit rate             10 Mbps (Manchester encode)
Slot time            512 bit times
Interframe gap       9.6 ms
Attempt limit        16
Backoff limit        10
Jam size             32 bits
Maximum frame size   1518 octets
Minimum frame size   512 bits
    Frame format & operational parameters

• A medium access control (MAC) unit is responsible for the
  encapsulation and de-encapsulation of frames for transmission
  and reception on the cable, error detection, and
  implementation of MAC algorithm.
• Each frame consists of 8 fields. All the fields are of fixed
  length except the data and padding fields.
• Preamble=> Sent at the head of all frames. ?used to achieve
  bit synchronization before the actual frame contents are
  received. It is a sequence of seven octets. (each equal to the
  binary pattern 10101010).
• SFD (start of frame delimiter)=> Single octet 10101011,
  signals the start of a valid frame t the receiver, immediately
  follows the preamble.
   Frame format & operational parameters
• Destination & Source addresses => Specify the intended
  destination station & originating station. Each address can
  be either 16 or 48 bits. If the first bit in the address field is
  0, it specifies the address is an individual address and the
  transmitted frame is intended for a single destination. If the
  bit is 1, it specifies a group address and the frame is
  intended either for a logically related group or for all others
  stations. In this case, the address field is set to all binary 1s.
• Length indicator => Specifies the number of octets in the
  data field.
• Pad => If the value of length indicator is less than the
  minimum frame size, a sequence of octets is added, known
  as padding.
• FCS => Contains CRC value that is used fir error detection.
CSMA/CD operation: a) transmit
CSMA / CD operation: b) receive
            Advantages

• CSMA/CD control software is relatively
  simple and produces little overhead.
• CSMA/CD network works best on a bus
  topology with bursty transmission. Bursty
  traffic is characterized by short, sporadic
  transmissions. Example: interactive
  terminal-host traffic.
• This technique is efficient for light to
  moderate load.
            Disadvantages

• CSMA/CD protocols are probabilistic and depends on the
   network (cable) loading. Performance tends to collapse
   under heavy load.
• Considered unsuitable for channels controlling automated
  equipment that must have certain control over channel
  access.
• We cannot set priorities to give faster access to some
   devices.
          Polling access method

• Polling is an access method that designates one device
  (called a "controller", "primary", or "master") as a channel
  access administrator.
• This device (Master) queries each of the other devices
  (“secondaries”) in some predetermined order to see
  whether they have information to transmit.
• If so, they transmit (usually through the master).
            Polling access method
• Secondaries may be linked to the master in many different
   configurations.
• One of the most common polling topologies is a star, where the
   points of the star are secondaries and the master is the hub.
• To get data from a secondary, the master addresses a request for
  data to the secondary, and then receives the data from the
  secondary sends (if secondary sends any).
• The primary then polls another secondary and receives the data
  from the secondary, and so forth.
• System limits how long each secondary can transmit on
  each poll.
                   Advantages
• Polling centralizes channel access control.
• Maximum and minimum access times and data rates on the
  channel are predictable and fixed.
• Priorities can be assigned to ensure faster access from some
  secondaries.
• When many stations have data to transmit over an extended
  period of time, round-robin techniques can be very efficient.
• If only a few stations have data to transmit over an extended
  period of time, then there is a considerable overhead in
  passing the turn from station to station, because most of the
  stations will not transmit but simply pass their turns.
• Polling is deterministic and is considered suitable for
  channels controlling some kinds of automated equipment.
                 Disadvantages

• Polling systems often use a lot of bandwidth sending notices
   and acknowledgments or listening for messages.
• Line turnaround time on a half- duplex line further increases
   time overhead.
• This overhead reduces both the channel's data rate under low
  loads and its throughput.
            Token passing System
• In token-passing systems, a small frame (the token) is
   passed in an orderly fashion from one device to another.
• A token is a special authorizing message that temporarily
  gives control of the channel to the device holding the token.
• Passing the token around distributes access control among
  the channel's devices.
• Each device knows from which device it receives the token
  and to which device it passes the token.(see fig.)
• Each device periodically gets control of the token, performs
  its duties, and then retransmits the token for the next device
  to use.
• System rules limit how long each device can control the
  token.
Control token MAC: Token ring
                         Token Ring
• Token ring networks are primarily used in technical and office
  environments.
• Whenever a station wishes to send a frame, it first waits for the
  token. When the station gets the token, it start sending frame.
• The intended recipient retains a copy of the frame and indicates
  by setting the response bits at the tail of the frame.
• A station releases the token in one of the two ways depending
  on the bit rate of the ring.
• In slower ring (4 Mbps), the token is released only after the
  response bits have been received.
• In higher speed rings (16 Mbps), the token is released after
  transmitting the last bit of a frame (early token).
Token Ring
      Token format & Frame format in Token
                      Ring
   Token Format:

          1           1           1     octets

        SD           AC         ED


      Frame Format:
                                                        octets
  1       1      1        2/6   2/6     <5000     4     1        1
SD       AC FC DA SA INFO                        FCS   ED FS

Start of frame                  FCS coverage           End of frame
  Field Descriptions of a Token Ring
JKOJKOOO             Start delimiter (SD)

JK1JK1 I     E       End delimiter (ED)

PPPT M RR R           Access control (AC)

FF    ZZZZZZ         Frame control (FC)

                                   Source and destination
I/G     15/47 bit address
                                   Address (SA/DA)
AC xx ACxx             Frame status (FS)
Control token MAC: Token bus
Slotted ring principles: bit definitions of each slot
Slotted ring principles:Outline
           topology
                       Advantages

• Even though there is more overhead using tokens than using
  CSMA/CD, performance differences are not noticeable with
  light traffic and are considerably better with heavy loads because
  CSMA/CD will spend a lot of time resolving collisions.
• A deterministic access method such as Token Ring guarantees
  that every node will get access to the network within a given
  length of time. In probabilistic access method (such as
  CSMA/CD) nodes have to check for network activity when they
  want to access the network.
              Disadvantages
• Components are more expensive than for
  Ethernet or ARCnet.
• Token Ring architecture is not very easy to
  extend to wide-area networks (WANs).
• Token Ring network is much more expensive
  than Ethernet. This is due to the complex
  token passing protocol.

								
To top