Docstoc

CE 4226 Network Systems Analysis and Design

Document Sample
CE 4226 Network Systems Analysis and Design Powered By Docstoc
					CE4226
Network Systems
Analysis and Design

 Switching and Routing Protocols
Virtual LAN

 An emulation of a standard LAN that
  allows data transfer to take place without
  the traditional physical restraints placed
  on a network
 A VLAN is a set of LAN devices that
  belong to an administrative group
 Group membership is based on
  configuration parameters and
  administrative policies rather than
  physical location
                                           2
Virtual LAN (Contd.)

 Members of a VLAN communicate with
  each other as if they were on the same
  wire or hub, when in fact they may be
  located on different physical LAN
  segments
 Members of a VLAN communicate with
  members in a different VLAN as if they
  were on different LAN segments, even
  when they are located in the same
  switch
                                           3
             Virtual LAN (Contd.)

                Switch A                               Switch B




Station A1    Station A2   Station A3   Station B1   Station B2   Station B3


               Network A                              Network B



                                                                       4
Virtual LAN (Contd.)
                    VLAN A

     Station A1     Station A2       Station A3




    Station B1    Station B2     Station B3

                                                  5
                    VLAN B
Virtual LAN (Contd.)

 VLANs can span multiple switches
 Trunk links VLANs that span multiple
  switches together
     IEEE 802.1Q standard
     Cisco Inter-Switch Link (ISL) protocol




                                               6
              Virtual LAN (Contd.)
                   VLAN A                                        VLAN A


   Station A1     Station A2     Station A3      Station A4     Station A5     Station A6




   Switch A                                                                    Switch B




Station B1      Station B2     Station B3     Station B4      Station B5     Station B6


                  VLAN B                                        VLAN B
                                                                                            7
Virtual LAN (Contd.)

 Each router interface defines a
  broadcast domain boundary
 Routers prevent broadcast traffic from
  propagating across interfaces
 VLANs define broadcast domain
  boundaries in a layer 2 network
 If the VLAN switch receives a broadcast
  on one port, it determines what other
  ports are allowed to receive

                                            8
Virtual LAN (Contd.)
             Router




                       Router




                                9
Virtual LAN (Contd.)




                       10
Virtual LAN (Contd.)




                       11
Virtual LAN (Contd.)

 Broadcast domain in switch network is difficult
  to see without access to the configuration files
 Each virtual bridge created within a switch
  defines a broadcast domain (VLAN)
 VLAN is a group of devices participating in the
  same broadcast domain
 They can communicate with each other without
  needing to communicate through a router



                                                    12
Virtual LAN (Contd.)

 Traffic from one VLAN cannot pass
  directly to another VLAN (between
  broadcast domains) within the same
  switch
 Use layer 3 devices to interconnect the
  VLANs




                                            13
VLAN Memberships

 Port-based VLANs
     A switch port is manually configured to be a
      member of a VLAN
     All machines on the port belong to the same VLAN
     Cisco’s Catalyst
 MAC-based VLANs
     VLAN membership is based on the MAC address
      of the workstation
     The switch has a table listing of the MAC address
      of each machine, along with the VLAN to which it
      belongs
                                                          14
VLAN Memberships

 Protocol-based VLANs
   Layer 3 data within the frame is used to
    determine VLAN membership
   For example, IP machines can be
       classified as the first VLAN, and AppleTalk
       machines as the second
      The major disadvantage of this method is
       that it violates the independence of the
       layers, so an upgrade from IPv4 to IPv6,
       for example, will cause the switch to fail

                                                 15
    Spanning Tree Protocol (STP)
                            Host A


LAN X




        Switch 1                     Switch 2




LAN Y




                   Host B

                                                16
    Spanning Tree Protocol (STP)
                            Host A


LAN X




        Switch 1                     Switch 2




LAN Y




                   Host B

                                                17
Spanning Tree Protocol (STP)
(Contd.)
 Spanning Tree Protocol (STP)
      IEEE802.1d
 Rapid Spanning Tree Protocol (RSTP)
      IEEE802.1w
      Provide rapid convergence of the spanning tree
       after a topology change
      Supersede the STP in IEEE802.1d 2004 edition
 Common Spanning Tree Protocol (CSTP)
      IEEE802.1q trunking protocol
      One instance of STP for all VLANs

                                                        18
Spanning Tree Protocol (STP)
(Contd.)
 Per VLAN Spanning Tree (PVST)
     Cisco’s proprietary protocol
     Build a separate logical tree topology for each
      VLAN
     Allow load sharing by having different forwarding
      paths per VLAN
     Use ISL trunking
 Per VLAN Spanning Tree+ (PVST+)
     provide the same functionality as PVST
     Use 802.1Q trunking rather than ISL

                                                          19
Spanning Tree Protocol (STP)
(Contd.)
 Multiple Spanning Tree Protocol (MSTP)
      IEEE 802.1s
      Use RSTP for rapid convergence
      Allow several VLANs to be mapped to a reduced
       number of spanning tree instances
      Reduce the number of spanning trees required to
       support a large number of VLANs
      Provide multiple forwarding paths for load sharing




                                                            20
Spanning Tree Protocol (STP)
(Contd.)
 Multi-Instance Spanning Tree Protocol
  (MISTP)
     Cisco’s proprietary protocol
     Allow a set of VLANs to be grouped into a
      single spanning tree




                                                  21
Redundancy and Load
Sharing in VLANs
 It is common practice to design
  redundant links between LAN switches
 STP avoids loops but allows only one
  active path
 STP provides redundancy, but not load
  sharing
 MSTP, PVST+, and MISTP offer both
  load sharing and redundancy

                                          22
Virtual LAN (Contd.)

 Advantages:
   Group users together, even though they
    weren't physically located together
   Simplify moves, adds, and changes
      Divide physical LANs into many logical
       LANs or broadcast domains
      a VLAN usually has its own IP subnet
      A router (or a routing module within a
       switch) provides inter-VLAN
       communication

                                                23
Routing Choices

 Routing
     Static or dynamic
     Distance-vector and link-state protocols
     Interior and exterior
     Etc.




                                                 24
Selecting a Routing Protocol

 Convergence time
 Resource consumption
 Bandwidth consumption
 Support VLSM
 Multipath load sharing
 Scalability
 Open standard
 Authentication
                               25
Next Hop Routing

 Each router makes routing decisions
  about how to reach a destination based
  on its routing table
 Router just selects the next hop leading
  to the destination
 This concept relies on the next hop
  router to select a further hop closer to a
  destination

                                               26
Next Hop Routing

 This independent hop-by-hop routing
  requires all routers have a consistent
  view of network
 To achieve consistency:
     Manually configure
     Use routing protocol




                                           27
Static Routing

 Network administrator compute the
  routing table for all routers in advance
 Easy to configure because we just
  simply tell each router how to reach
  every indirect network segment
 It is predictable and controllable
 No overhead on routers and links



                                             28
    Static Routing Example
                         172.16.5.0/24                       172.16.3.0/24


172.16.2.0/24                       Router 2
                                        3 Com         172.16.3.1

            172.16.2.1     172.16.5.1                          172.16.3.2
    3 Com
                                                172.16.1.2
                                                                        3 Com




                 Router 1                               Router 3
            172.16.1.1                                         172.16.4.1



             172.16.1.0/24                                     172.16.4.0/24
                                                                                29
  Static Routing Example

 Router 1
     ip route 172.16.3.0 255.255.255.0 172.16.1.2
     ip route 172.16.4.0 255.255.255.0 172.16.1.2
     ip route 172.16.5.0 255.255.255.0 172.16.1.2
 Router 2
     ip route 172.16.2.0 255.255.255.0 172.16.1.1
     ip route 172.16.4.0 255.255.255.0 172.16.3.2


                                                 30
  Static Routing Example

 Router 3
     ip route 172.16.1.0 255.255.255.0 172.16.3.1
     ip route 172.16.2.0 255.255.255.0 172.16.3.1
     ip route 172.16.5.0 255.255.255.0 172.16.3.1
 Another way
     ip route 0.0.0.0 0.0.0.0 172.16.3.1




                                                 31
Static Routing

 The price of its simplicity is a lack of
  scalability
 Difficult to update the configuration
 Error prone
 It cannot use redundant links to adapt to
  network failure (unless you configure 2
  or more static routes to the same
  destination)

                                             32
Dynamic Routing

 Main advantages are scalability and
  adaptability
 Not require reconfiguration when
  network changes
 Adapt to partial failure




                                        33
Dynamic Routing

 Routers learn about network topology by
  communicating with other routers by
  using routing protocol
 Each router announces its presence,
  and available routes on the network
 Other routers hear and adjust routing
  table accordingly


                                          34
Dynamic Routing

 Introduce complexity
     Prepare information to send to others
     Select the best route among many
      candidates
     Remove the old or unusable information
 Bandwidth overhead to exchange
  information


                                               35
Hybrid Routing

 Some parts of the network use static
  routing, and some other parts use
  dynamic routing
     Static routing in the access network
     Dynamic routing in the core and
      distribution networks
 Use static routing for stub networks
 Use dynamic routing when there are
  many router connections

                                             36
Dynamic Routing Protocols

 Exterior, interior protocols
      Exterior Gateway Protocol (EGP)
      Interior Gateway Protocol (IGP)
 Distance vector, link state protocols




                                          37
Exterior VS Interior Protocols

 Exterior gateway protocol
     Carry routing information between two
      independent administrative entities, such
      as two corporations, two universities
     Border Gateway Protocol (BGP) is the
      most widely used




                                                  38
Exterior VS Interior Protocols

 Interior gateway protocol
     Used within a single administrative domain
      or among closely cooperative groups
     Routing Information Protocol (RIP), Open
      Shortest Path First (OSPF), Interior
      Gateway Routing Protocol (IGRP),
      Enhanced Interior Gateway Routing
      Protocol (EIGRP) are common


                                               39
Exterior VS Interior Protocols

 It is possible to use IGP as EGP and
  vice versa, but not a good idea
 EGPs are designed to scale to the
  largest of networks
 Therefore, EGPs has complexity and
  overhead
 IGPs are fairly simple and have little
  overhead, but they don’t scale well

                                           40
Distance Vector VS Link State
Protocols
 Another way to classify is by what information
  the routers tell each other, and how they use
  the information to form their routing tables
 Distance vector protocol
      A router periodically sends all of its neighbors two
       pieces of information about the destinations it
       knows how to reach
         Distance to the destination

         Vector (direction) to the destination indicating the
          next hop


                                                                 41
Distance Vector VS Link State
Protocols
   Simple to configure and understand
   Updates are periodically broadcasted
   Slow convergence
   Less processor intensive
   No complete network topology
       Loop
       Count to infinity
       Split horizon
       Poisonous reverse
       Holddown timer
       Triggered update

                                           42
Distance Vector VS Link State
Protocols
 Link state protocol
     A router provides information about the
      topology of the network in its vicinity to all
      routers in the network
        Links it is attached to
        State of those links




                                                       43
Distance Vector VS Link State
Protocols
 Complex
 Update only when state changes
 Fast convergence
 High processing power
 Have complete network topology
     Loop-free
 Scale better than DV


                                   44
RIP

 Routing Information Protocol
 RIP was developed originally for the
  Xerox Network System (XNS) protocols
  and was adopted by the IP community in
  the early 1980s
 RIP version 1 (RIPv1): RFC 1058
 RIP version 2 (RIPv2): RFC 2453



                                         45
RIP
 RIP broadcasts its routing table every 30
    seconds
   RIP sends triggered updates when the metric of
    a route changes (only the changes)
   RIP allows 25 routes per update packet
   Easy to configure and troubleshoot for flat or
    edge network
   Uses a single routing metric (hop count) to
    measure the distance to a destination network
   The hop count can not go above 15
   A hop count of 16 means the distance to the
    destination is infinity (unreachable)
                                                 46
RIP
 RIP uses split horizon with poison reverse
    route updates are sent out an interface with an
     infinite metric for routes learned (received) from the
     same interface
 RIP supports equal-cost load balancing to a
    destination
   RIP has an administrative distance of 120
   RIP summarizes IP addresses at network
    boundaries
   RIP uses UDP port 520
   RIPv1 sends update to 255.255.255.255
   RIPv2 sends update to 224.0.0.9
                                                         47
RIP Timer

 Update Timer
    Specifies the frequency of the periodic broadcasts
    By default, the update timer is set to 30 seconds

 Invalid Timer
    When the invalid timer expires, the route is marked
     as invalid (possibly down)
    The router marks the route invalid by setting the
     metric to 16
    The route is retained in the routing table and
     packets can be forwarded as normal
    By default, the invalid timer is 180 seconds, or six
     updates periods (30 x 6 = 180)
                                                          48
RIP Timer

 Holddown Timer
    Cisco implements an additional timer for RIP
    The holddown timer stabilizes routes
    The holddown timer starts when route is
     unreachable
         After invalid timer expires
         After receiving update that route is
          unreachable
      Router accepts no updates for the route until the
       holddown timer expires
      By default, the holddown timer is 180 seconds

                                                           49
RIP Timer

 Flush Timer
     A route entry marked as invalid is retained
      in the routing table until the flush timer
      expires
     By default, the flush timer is 240 seconds,
      which is 60 seconds longer than the invalid
      timer
     Start automatically after last update is
      received (along with invalid timer)

                                                50
RIP

 RIPv1 is a classful routing protocol
 Discontiguous subnets are not visible to
  each other, and VLSM is not supported
 RIPv2 is classless
 RIPv2 supports authentication
     simple plain-text password
     MD5 authentication
 It still sends updates every 30 seconds
  and retains the 15-hop limit
                                             51
RIP

 Downed link
      This failure is detected by the interface hardware
       and indicated to the router
      This change is sent out as a triggered update
 Downed router
    Downed router couldn’t inform neighboring routers

      It takes up to 3 minutes for the neighboring routers
       to mark the routes learned from the downed router
       as unreachable
      Then use triggered updates

                                                            52
IGRP
 Interior Gateway Routing Protocol (IGRP)
  developed by Cisco to overcome RIP
 IGRP update timer is 90 second
 IGRP has a maximum hop limit of 100, by
  default, and can be configured to support a
  network diameter of 255
 IGRP uses a vector metric and a composite
  metric based on the following factors:
      Bandwidth
      Delay
      Reliability
      Load

                                                53
IGRP Metric

 Bandwidth
     The bandwidth of the lowest-bandwidth link on
      the path
     Equal to 107 / Data rate (kbps)
 Delay
     The sum of all the delays for outgoing interfaces
      in the path
     Delay is not dynamically calculated
     Each router interface has a default delay (can be
      changed by administrator)
     Equal to Delay (microseconds) / 10

                                                      54
IGRP Metric
 Reliability
      Based on the interface reliability reported by
       routers in the path
      255 is 100 percent reliable and 1 is minimally
       reliable
      It is dynamically calculated (5-min interval)
 Load
      Based on the interface load reported by routers
       in the path
      255 is 100 percent loaded and 1 is minimally
       loaded
      Load is dynamically calculated (5-min interval)

                                                         55
        IGRP Metric

                            BW                           k5
metric  k1  BW  k 2              k3  delay 
                         256  load                reliabilit y  k 4

         By default, K1 = K3 = 1 and K2 = K4 = K5 = 0


                    metric  BW  delay

         K1 to k5 can be changed but must be same for
           all routers

                                                                 56
     IGRP Metric
Media        Data rate (kbps) BW     Delay (us) Delay
ATM          155000         65       100       10
155Mb/s
Fast         100000         100      100       10
Ethernet
FDDI         100000         100      100       10
Token Ring   16000          625      630       63

Ethernet     10000          1000     1000      100
T1           1544           6476     20000     2000
E1           2048           5000     20000     2000
DS0          64             156250   20000     2000
56 kbps      56             178571   20000     2000
                                                        57
IGRP Metric

 Link1 is serial 1.544 Mbps
 Link2 is Ethernet 10 Mbps
 IGRPmetric = (10,000,000/1544) +
  (20000 + 1000)/10
 IGRPmetric = 6476 + 2100 = 8576




                                     58
IGRP Metric

 IGRP uses composite metric (CM) to
  make routing decision
 But updates vector metric (VM) to other
  neighbors
     NewVM.BW = min(UpdtVM.BW, Link.BW)
     NewVM.DL = UpdtVM.DL + Link.DL
     NewVM.RL = min(UpdtVM.RL, Link.RL)
     NewVM.LD = max(UpdtVM.LD, Link.LD)
     CM = min(NewCM, OldCM)

                                            59
IGRP Timer

 Update Timer
    IGRP sends its routing table to its neighbors
     every 90 seconds
 Invalid Timer
    IGRP uses an invalid timer to mark a route
     as invalid after 270 seconds (three times the
     update timer)




                                                 60
IGRP Timer

 Holddown Timer
    The router accepts no new changes for the
     route until the holddown timer expires
    This setup prevents routing loops in the
     network
    The default holddown timer is 280 seconds
 Flush Timer
    IGRP uses a flush timer to remove a route
     from the routing table
    The default flush timer is set to 630 seconds


                                                 61
IGRP

 IGRP is classful
 IGRP implements split horizon with poison
  reverse, triggered updates, and holddown
  timers for stability and loop prevention
 By default, IGRP will load-balance traffic if there
  are several paths with equal cost (up to 6)
 IGRP can load-balance over unequal-cost
  paths



                                                   62
IGRP

 104 routes per IGRP message
 Administrative distance is 100
 IGRP uses IP directly, with protocol
  number 9
 IGRP sends update to 255.255.255.255
 No support for authentication
 Largely replaced by EIGRP


                                         63
EIGRP

 Enhanced Interior Gateway Routing
  Protocol (EIGRP)
 EIGRP is compatible with IGRP
 EIGRP can also redistribute routes for
  RIP, IS-IS, BGP, and OSPF
 EIGRP has Protocol-Dependent
  Modules that can deal with AppleTalk
  and Novell’s IPX

                                           64
EIGRP

 EIGRP is a classless protocol
 Hybrid distance-vector protocol
 EIGRP uses hellos and forms neighbor
  relationships (as link-state protocol) to
  detect the new or the loss of a neighbor
 EIGRP uses the composite metric as
  IGRP but multiplied with 256
 EIGRP does not send periodic updates

                                              65
EIGRP

 By default, EIGRP will load-balance
  traffic if there are several paths with
  equal cost
 EIGRP can load-balance over unequal-
  cost paths
 EIGRP supports authentication
     simple plain-text password
     MD5 authentication

                                            66
EIGRP

 EIGRP packet uses both multicast
  (224.0.0.10) and unicast
 Reliable Transport Protocol (RTP)
  ensures delivery in order
 EIGRP uses IP protocol number 88
 Administrative distance
     EIGRP internal routes are 90
     EIGRP external routes are 170
     EIGRP summary routes are 5

                                      67
EIGRP

 EIGRP uses Diffusing-Update Algorithm
  (DUAL) for fast convergence and less
  bandwidth utilization
 Updates are
      Nonperiodic: updates are sent only when a metric
       changes rather than at regular intervals
      Partial: updates include only routes that have
       changed, not every entry in the routing table
      Bounded: updates are sent only to affected routers



                                                        68
OSPF

 Open Shortest Path First (OSPF)
 RFC 2328
 Link state protocol
 Classless protocol
 Hierarchical routing protocol
     Divide into areas
 Support equal cost multipath load
  balancing for up to 6 paths
                                      69
OSPF

 OSPF propagates only changes to minimize
  bandwidth utilization
 OSPF networks converge far more quickly than
  RIP networks
 Within a network multiple areas can be created
  to help ease CPU use in SPF calculations,
  memory use and the number of LSAs being
  transmitted
 60-80 routers are considered to be the
  maximum to have in one area
                                               70
OSPF

 The default area is 0.0.0.0 and should exist
  even if there is only one area in the whole
  network
 All other areas must connect to the backbone
  area (0.0.0.0) via Area Border Router (ABR)
 ABR sends and receives summary route from
  backbone area
 Autonomous System Boundary Router (ASBR)
  connects an OSPF network to another network
  that uses a routing protocol other than OSPF
                                                 71
OSPF




       72
OSPF

 OSPF is sent over IP, protocol type 89
 OSPF uses two multicast addresses for
  broadcast and point to point networks
     224.0.0.5 for all OSPF routers
     224.0.0.6 for all designed routers
 OSPF uses unicast for Non-Broadcast
  Multi-Access (NBMA) networks such as
  ISDN, X.25, frame relay, ATM

                                           73
OSPF

 OSPF calculates cost by
     Cost = 108 / data rate (bps)
     Cost = 1 for data rate > 100 Mbps
 OSPF supports authentication
     simple plain-text password
     MD5 authentication




                                          74
              Summary of Routing Protocol
              Features
Protocol            RIP           OSPF   IGRP   EIGRP
Type                DV            LS     DV     DV
Convergence Time    Slow          Fast   Slow   Fast
VLSM                Yes (v2)      Yes    No     Yes
Bandwidth usage     High          Low    High   Low
Resource usage      Low           High   Low    Low
Multipath support   Yes (Cisco)   Yes    Yes    Yes
Scale well          No            Yes    Yes    Yes
Proprietary         No            No     Yes    Yes
Non-IP protocols    No            No     No     Yes

                                                        75

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:14
posted:8/24/2011
language:English
pages:75