Chapter 7 Packet-Switching Networks - PowerPoint PPT Presentation


PPT – Chapter 7 Packet-Switching Networks PowerPoint presentation | free to download - id: 764402-NzFmO


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation

Chapter 7 Packet-Switching Networks


CHAPTER 7 PACKET-SWITCHING NETWORKS Dept of computer science and engg 6th sem cse SSE Initally all the nodes other than the destination node are at infinite cost to ... – PowerPoint PPT presentation

Number of Views:100
Avg rating:3.0/5.0
Slides: 47
Provided by: MYP101


Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Chapter 7 Packet-Switching Networks

Chapter 7 Packet-Switching Networks
  • Dept of computer science and engg
  • 6th sem cse
  • SSE

Virtual circuit packet switching
  • Virtual circuit packet switching establishes a
    fixed path called

    before any packets flow through the network.
  • Virtual circuit resides in network layer.
  • Figure 2
  • Delay is incurred when a message is broken into 3
    packets and transmitter over the circuit.
  • Same delay as datagram packet switching execpt
    for additional delay used to set up the circuit.

  • Path is determined parameters are set in swiches
    by exchanging Connect request and Connect confirm
  • If the switch does not have enough resources then
    it responds with a connect reject message and the
    setup procedure fails.
  • Figure 3
  • At the input of every switch ,the virtual circuit
    is identified by

  • When packet arrives at the input port the VCI in
    the header is used to access the table.
  • The table lookup provides the output port to
    which the packet is fowarded and the VCI that is
    used at the input port of the next switch.
  • Call setup procedure sets up chain of pointers
    across the network that directs the flow of
  • Figure 4

Structure of packet switch
  • A packet switch will perform two main functions
  • Routing function uses algorithm to find path to
  • Stores result in routing table
  • Forwarding function processes incoming packet
    from input port
  • Forwards packet to appropriate output port based
    on routing table information

Generic packet switch
  • Input port, output port, interconnection fabric,
    switch controller

  • Input ports and output ports are paired
  • Line card
  • Contain several input/output port.
  • Capacity of link connecting the fabric is high
    speed, fully utilized
  • Implements physical, data link and network layer
  • Symbolic timing, line coding, framing, physical
    layer addressing ,error checking

Organization of line card made up of various
Figure 6-gt
  • Line card supports medium access control protocol
    to handle broadcast network.
  • Implemented by special purpose chipset
  • Network layer routing table in line card
  • Need to perform fast table lookup to find output
  • Contain buffers and associated scheduling
  • Network processor performs table lookup and
    packet scheduling

  • Contain general purpose processor
  • This manages control and management functions
  • If controller is connectionless executes routing
  • If connection oriented handles signally messages.
  • Controller communicates with line card and fabric
    to configure internal parameters.

Inter connection fabric
  • Transfers packet between line card
  • Crossbar interconnection fabric transfer packets
    parallel between input and output ports
  • Buffers are added to crossbar to accommodate
    packet contention.
  • Buffer located at input or output port.

Output buffering
  • Crossbar with output buffering should run N times
    faster than port speed as N packets
    simultaneously arrive at a particular output
  • If output is idle, one packet is transmitted rest
    are buffered.
  • Only one packet is allowed to proceed to
    particular output

Input buffering
  • 2 packets at input 2,first packet like to go to
    output 3,second to output 8
  • Packet from input buffer 1 also want to go to
    output 3
  • Frist packet from input buffer should wait ,2nd
    should wait behind 1st even though output 8 is
  • Results in performance degradation .
  • Problem called HEAD of Line Blocking

Banyan switch
  • Solution is building a large switch
  • Composed of 22 switching elements interconnected
    in certain fashion
  • Exactly 1 path will exist from input to each
  • Routing is done in distributed manner
  • Ie appending binary the binary address of the
    output number to each packet
  • Switching element at stage I steers a packet
    based on ith bit of the address.
  • Bit is 0 the element steers packet to upper

  • Input 1 likes to send it to the output destines
    to packet 5
  • The switching element at stage 1 looks at first
  • Of the address and steers the packet to its lower
  • At stage 2 it steers the packet to the upper
    output since the second bit is stage 3
    steers the packet to lower output and sends to
    output 5

Routing in packet networks
  • Routing is concerned with determining feasible
    paths for packet to follow from source to
  • 1)Rapid and accurate delivery of packets
  • Routing algorithm should operate correctly
  • Find path to correct destination
  • Should not take long time to find path
  • 2)Adaptablilty to changes in network topology
  • Network equipment and transmission line will fail
  • Routing algorithm should adapt and reconsider the
  • 3)Adaptability to various source destination
    traffic loads
  • Traffic loads change dynamically, hence should
    adjust path based on the current traffic loads.

  • 4) Ability to route packets away from congested
  • Routing algorithm should avoid congested links
    and balance the load.
  • 5) Ability to determine connectivity of network
  • Routing system should know connectivity to find
    optimal paths
  • 6)Ablility to avoid loops
  • Inconsistent information will lead to routing
    loops. Routing system should avoid the loops in
    presence of distributed routing system.
  • 7)Low Overhead
  • Should obtain connectivity information by
    exchanging control messages with routing system.
    These messages create overhead on bandwidth usage
    and hence should be minimized.

Routing algorithm Classification
  • Routing algorithm is classifies as
  • 1) Static routing
  • 2)Dynamic routing
  • Static routing
  • Paths are precomputed based on the network
    topology, link capacities and other information
  • Computation is completed the paths are loaded in
    the routing table of each node.
  • This remains fixed for a long time.
  • Works when network size is small, and the traffic
    doesnt change in time
  • It becomes cumbersome, as network size increases.
  • Disadvantage is inability react to network

Dynamic routing
  • Each node will continuously learn the state of
    the network by communicating with the neighbors
  • Change in network topology is eventually
    propagated trough nodes
  • Based on the information the node computes the
    best path to the destination
  • Disadvantage is the added complexity with the
  • They can also be distinguished as
  • Centralized routing
  • Distributed routing

Centralized and Distributed routing
  • In Centralized routing the network control
    center computes all the path and then uploads the
    information in the network
  • In Distributed routing nodes cooperate by the
    means of message exchanges and perform their own
    routing computations.
  • Distributed routing is better than centralized
    routing but generate inconsistent results ,loop
    wil develop
  • If A thinks B is the best path to Z and If B
    thinks A is the best path to B
  • Then desired packet to Z have the misfortune of
    arriving at A or B and stuck in a loop between A
    and B

Routing Tables
  • Once the routing algorithm defines the set of
    path the path information is stored in the
    routing table.
  • In virtual circuit packet switching the routing
    table translates each incoming VCI to outgoing
    VCI an identifies the output port to which the
    packet is to be forwarded.
  • Datagram packet switching the routing table
    identifies the next hop to forward the packet
    based on destination address of the packet
  • Virtual circuits are bidirectional.

  • A packet sent by node A with VCI 1 will
    eventually reach node B and from node A with VCI
    5 will reach node D
  • VCI 1 from node A will be translated to 2 and to
    7 and finally to 8 and reaches B
  • When node 1 will receive the header with VCI 1
    that node replaces the incoming VCI with 2 ande
    forwards the packet with node 3

  • If a packet with VCI 5 arrives at node 1 from
    node A,the packets is forwarded to node 3 after
    the VCI is replaced with 3
  • After arriving at node 3 the packet receives the
    outgoing VCI 4 and is forwarded to node 4
  • Node 4 translates the VCI to 5 and forwards the
    packet to node 5.
  • Node 5 translates to VCI 2 and delivers the
    packet to destination node D
  • The routing table uses node numbers to identify
    where the packet comes from and where the packet
    is forwarded.

Routing table for packet switching network
  • Shows routing table for network topology
  • Minimum hoping routing objective
  • Packet destines to node 6 arrives at node 1, the
    packet is fowarded to node 3 based on the
    corresponding routing table entry
  • Node 3 forwards the packet to node 6

Hierarchical routing
  • The size of routing table is reduced
  • Hosts that are near to each other should have
    addresses that have common perfixes
  • Routers will examine the part of address and
    decide to route the packet
  • Example of hierarchal address assignment and flat
    address assignment

  • Forwards an incoming packet to all the ports
    except the one received from
  • Each switch performs the flooding process and
    eventually reach the destination
  • Effective routing approach when the routing
    information is not available
  • Also effective when source need to send packet to
    all the nodes
  • Flooding may easily swamp network

  • To reduce resource consumption in the network
  • Time to live field (TTL) in each packet
  • Source sends the packet TTL is set to some number
  • Each node decrements the TTL by 1 before flooding
    the packet
  • If value reaches 0 the node discards the packet
  • Avoid wastage of bandwidth, minimum hop number
    between two further nodes
  • 2nd ,the node adds identifier to header of packet
  • When node receives the packet that contain the
    identifier it discards it as it knows packet
    already visited the node(prevents looping)
  • Identified with unique sequence number,node will
    record the sequence number and source address
  • If already visited it will discard the packet

Deflection routing
  • Also called hot potato routing. eg Manhatten
    street network
  • They provide multiple path for each source
  • Each path tries to forward the packet to the
    preffered port
  • If preferred port is busy or congested the packet
    is deflected to another port

  • Advantage is node are buffer less, packets do not
    have to wait for a specific port to become
  • If unavailable it will deflect the packet to
    another port, reaches destination
  • Since they take alternate path they cannot
    guarantee in-sequence delivery of packets
  • Used in optical network where optical buffers are
    difficult to find
  • Used to implement high speed switches where
    topology is regular and buffers are expensive

Shortest path routing
  • Routing algorithm is based on
  • Shortest path algorithm
  • Each link represents the cost of using the link
  • Shortest path between node 2 to node 6 is trough
  • 2-4-3-6 and the cost path is 4
  • Metrics used to assign the cost
  • Cost 1/capacity Higher cost to lower capacity
  • Cost packet delay It includes queuing delay in
    switch buffer and propagation delay in link
  • 3 Cost Congestion Congestion measure is
    traffic loading. Avoids Congested link

The Bellman Ford Algorithm
  • If Neighbor of node A knows the shortest path to
    node Z ,then node A can determine the shortest
    path to Z
  • Ie Node A will calculate the cost/Distance to
    node Z trough its neighbors by picking the

  • Dj Current estimate to minimum cost to
    destination node
  • Cij link cost from node I to node j (c13c312)
  • (c110) (c15c23 8)
  • Minimum cost from node 2 to node 6 trough node 1
    ,node 4 and node 5
  • D2 minC21 D1 , C24 D4 ,C23 D5
  • min33,13,42
  • 4
  • 1.Initialization Di 8 for all i ? d
  • Dd0
  • 2. Updating for each I ?d
  • Di Minj Cij Dj for all j ? i

  • Initally all the nodes other than the destination
    node are at infinite cost to node 6.
  • (Iteration 1)Node 3 finds that it is connected to
    node 6 with cost of 1.Node 5 finds it is
    connected to node 6 with cost 2.Node 3 and 5
    update their entries.
  • (Iteration 2)Node 1 finds it can reach node 6 via
    node 3 with cost 3.Node 2 finds it can reach node
    6 via node 5 with cost 6. Node 4 finds it has
    path via 3 and 5 with costs 3 and 7
    respectively.Node 4 selects the path via node
    3.Node 1,2 and 4 update their entries
  • (Iteration 3)Node 2 finds it can reach node 6 via
    node 4 with distance 4.Node 2 changes the entry
    to (4,4) and informs the neighbours
  • Node 1,4,5 processes the new entry from node 2
    but donot find new shortest paths.The algorithm
    is conservged

  • Link connecting node 3 and node 6 breaks. Compute
    minimum cost for each node to destination node.

  • (update 1)As soon as node 3 detects that
    link(3,6) breaks, node 3 recomputed the minimum
    cost to node 6.Node 3 looks for path to 6 trough
    neighbours, Node 1 and Node 4 and its calculation
    indicates that the new shortest path is trough
    node 4 at a cost of 5.Node 3 then sends the new
    routing updates that is node 1 and node 4
  • (update 2)Node 1 and 4 recomputes their minimum
    costs node 1 findsits shortest path is still
    trough node 3 but the cost has increased to
    7node 4 finds the shortest path is trough node 2
    or node 5 with cost 5.node 4 chooses node 2.Node
    1 transmits its routing updates to node 2,node 3
    and node 4
  • Node 4 transmits its routing updates to node

  • (Update 3)Node 1 finds its shortest path is still
    trough node 3.Node 2 finds its shortest path is
    still trough node 4 but the cost has increased to
    6.Node 3 finds its shortest path is still through
    node 4 but the cost has increased to 7.Node 4 and
    Node 5 finds the shortest path is not changed.
    Node 2 transmits its updates to node1,node4 and
    node 5 and transmits its updates to node 1 and 4.
  • (update 4) Node 1 finds its shortest path is
    through either node 2 or node 3 with cost
    9.Suppose that node 1 chooses node 2.Node 4 now
    finds the shortest path through node 5 with cost
    5 since the cost through 2 has incresed to 7.Node
    5 does not change its shortest path.Node 1
    transmits its updates to node 2,node 3,node 4 and
    node 4 transmit its updates to node 1,2,3 and 5
  • (update 5) None of the nodes find short paths

  • Update node1 node2 node3
  • Before break (2,3) (3,2)
  • After break (2,3) (3,2) (2,3)
  • (2,3) (3,4)
  • (2,5) (3,4) (2,5)
  • (2,5) (3,6) (2,5)
  • (2,7) (3,6)
  • (2,7) (3,8) (2,7)

Split Horizon with poison reverse
  • Update Node 1 Node 2 Node3
  • Before break (2,3) (3,2) (4,1)
  • After Break (2,3) (3,2) (-1, )
  • 1 (2,3) (-1, ) (-1, )
  • 2 (-1, ) (-1, ) (-1, )
  • After the link breaks node 3 sets the cost of the
    destination to infinity,since the minimum cost
    node 3 has received from node 2 is also
    infinity.When node 2 received the update message
    it also sets to infinity.Node 1 learns the
    destination is unreachable

Dijkstra s Algortihm
  • Alternate algorithm for finding the shortest
    paths to source to all other nodes.
  • To identify the closest nodes from the source
    node in the order of their increasing path cost.
  • 1 Initialalization
  • Ns
  • DjCsj,for all j ? s
  • 2.Finding the next closest node
  • Dimin Dj
  • Add i to N
  • 3.Update minimum cost from node I to node N
  • DjminDj,DiCij

Source routing versus Hop by Hop Routing
  • Datagram network each node determines the next
    hop along the shotest path ,packet travels from
    source as it follows HOP BY HOP routing to
  • SOURCE ROUTING the path to the destination is
    determined by source.Explicit routing allow a
    perticular packet node to determine the path.
  • Source includes path information in the packet
    header that contain sequence of nodes to traverse
    and sufficient information to nodes till packet
    is fowarded to destination.

  • Each node examines the header,removes address
    identifying the node,forwards packet to next node
  • If we reserve the path information we can
    traverse the node back to the source by reversing
    the path.

Link state routing versus Distance vector routing
  • Distance Vector routing
  • Neighboring routes exchange routing tables that
    sets the vector of known distance to destination.
    Bellman ford algorithm is used to find the best
    path though information from the neighbors. With
    new path the router wil send new vector to
  • This adapts to change in network topology
  • Link State Routing
  • Each router will flood information to the state
    of the links that connects to neighbors
  • Allows routers to construct map of entire network
    and then uses Djikstra s algorithm. If state is
    changed the router detects the change and floods
    the new information to network

ATM Networks
  • Asynchronous transfer mode (ATM) is a method for
    multiplexing and switching that supports broad
    range of services
  • Connection oriented packet switching technique
    that provides quality of service(QoS)
  • Variable bit rate delay bursty
    traffic processing
  • TDM Multirate only low,fixed Inefficient
    Minimal ,high speed
  • Packet Easily handled variable
    efficient header and packet

  • ATM is packet based,can easily handle services
    and generate information in bursty fashion on
    variable bit rates.
  • The abbreviated header of ATM facilitates
    implementations that result in low delay and high

  • Information flow generated from various users is
    converted into cell and sent to ATM multiplexer
  • It arranges the cell into one or more queues and
    implements some scheduling strategy to determine
    which order it is transmitted