1

Two key network-layer functions

• forwarding: move packets from router’s input to appropriate router output

• routing: determine route taken by packets from source to dest.

– routing algorithms

analogy:

routing: process of planning trip from source to dest

forwarding: process of getting through single interchange 1

23

0111

value in arriving

packet’s header

routing algorithm

local forwarding table

header value output link

0100

0101

0111

1001

3

2

2

1

Interplay between routing and forwarding

routing algorithm determines

end-end-path through network

forwarding table determines

local forwarding at this router

2

Network vs Transport Layer

• network vs transport layer connection service:– network: between two hosts

– transport: between two processes

Network service model

Q: What service model for “channel” transporting

datagrams from sender to receiver?

example services for individual datagrams:

guaranteed delivery

guaranteed delivery with less than 40 msec delay

example services for a flow of datagrams:

• in-order datagram delivery

• guaranteed minimum bandwidth to flow

• restrictions on changes in inter-packet spacing

3

Network layer service models:

Network

Architecture

Internet

ATM

ATM

ATM

ATM

Service

Model

best effort

CBR

VBR

ABR

UBR

Bandwidth

none

constant

rate

guaranteed

rate

guaranteed

minimum

none

Loss

no

yes

yes

no

no

Order

no

yes

yes

yes

yes

Timing

no

yes

yes

no

no

Congestion

feedback

no (inferred

via loss)

no

congestion

no

congestion

yes

no

Guarantees ? Destination Address Range

11001000 00010111 00010000 00000000

through11001000 00010111 00010111 11111111

11001000 00010111 00011000 00000000

through11001000 00010111 00011000 11111111

11001000 00010111 00011001 00000000

through11001000 00010111 00011111 11111111

otherwise

Link Interface

0

1

2

3

Q: but what happens if ranges don’t divide up so nicely?

Datagram forwarding table

4

Longest prefix matching

Destination Address Range

11001000 00010111 00010*** *********

11001000 00010111 00011000 *********

11001000 00010111 00011*** *********

otherwise

DA: 11001000 00010111 00011000 10101010

examples:

DA: 11001000 00010111 00010110 10100001 which interface?

which interface?

when looking for forwarding table entry for given destination address, use longest address prefix that matches destination address.

longest prefix matching

Link interface

0

1

2

3

Router architecture overviewtwo key router functions: run routing algorithms/protocol (RIP, OSPF, BGP)

forwarding datagrams from incoming to outgoing link

high-speed switching

fabric

routing processor

router input ports router output ports

forwarding data

plane (hardware)

routing, management

control plane (software)

forwarding tables computed,

pushed to input ports

5

line

termination

link layer

protocol(receive)

lookup,

forwarding

queueing

Input port functions

decentralized switching:• given datagram dest., lookup output

port using forwarding table in input port memory (“match plus action”)

• goal: complete input port processing at ‘line speed’

• queuing: if datagrams arrive faster than forwarding rate into switch fabric

physical layer:

bit-level reception

data link layer:

e.g., Ethernet

see chapter 5

switchfabric

Switching fabrics

transfer packet from input buffer to appropriate output buffer

switching rate: rate at which packets can be transfer from inputs to outputs often measured as multiple of input/output line rate

N inputs: switching rate N times line rate desirable

three types of switching fabrics

memory

memory

bus crossbar

6

Switching via memory

first generation routers:• traditional computers with switching under direct

control of CPU

• packet copied to system’s memory

• speed limited by memory bandwidth (2 bus crossings per datagram)

inputport(e.g.,

Ethernet)

memory

outputport(e.g.,

Ethernet)

system bus

Switching via a bus

datagram from input port memory

to output port memory via a shared bus

bus contention: switching speed limited by bus bandwidth

32 Gbps bus, Cisco 5600: sufficient speed for access and enterprise routers

bus

7

Switching via interconnection network

overcome bus bandwidth limitations

banyan networks, crossbar, other interconnection nets initially developed to connect processors in multiprocessor

advanced design: fragmenting datagram into fixed length cells, switch cells through the fabric.

Cisco 12000: switches 60 Gbps through the interconnection network

crossbar

Output ports

buffering required when datagrams arrive from fabric faster than the transmission rate

scheduling discipline chooses among queued datagrams for transmission

linetermination

link layer

protocol(send)

switchfabric

datagram

buffer

queueing

This slide in HUGELY important!

Datagram (packets) can be lost due to congestion, lack of buffers

Priority scheduling – who gets best performance, network neutrality

8

Output port queueing

• buffering when arrival rate via switch exceeds output line speed

• queueing (delay) and loss due to output port buffer overflow!

at t, packets more

from input to output

one packet time later

switch

fabric

switch

fabric

How much buffering?• RFC 3439 rule of thumb: average buffering

equal to “typical” RTT (say 250 msec) times link capacity C– e.g., C = 10 Gpbs link: 2.5 Gbit buffer

• recent recommendation: with N flows, buffering equal to

RTT C.

N

9

17

Routing & IP Routing Protocols

18

Routing: Problem Definition• You are a router in a packet switched network and you receive a

packet destined to some remote node– E.g., router A below receives a packet destined to node F

• Question: How does A know where to send this packet?– Does A send it to B? C? or D?

• Answer: Recall from our earlier discussion in IP that router A consults a forwarding table to make this decision

• Routing Problem: How does router A build this forwarding table?– Built by a routing algorithm (protocol): The job of the routing algorithm

is to determine the next hop router for ALL destinations in the network

A

ED

CB

F

2

2

13

1

1

2

53

5

Dest. Next Hop CostB B 2D D 1C D 3E D 2F D 4

Forwarding Table in A

10

19

End-to-End Path Determination: Routing principles

Graph abstraction for routing algorithms:

• graph nodes are routers

• graph edges are physical links– link cost: delay, $ cost, or

congestion level (amount of traffic carried on the link)

Goal: determine “good” path(sequence of routers) thru

network from source to dest.

Routing protocol

A

ED

CB

F

2

2

13

1

1

2

53

5

• “good” path:– typically means minimum

cost path

20

Forwarding vs Routing

• Distinction between “Forwarding” and “Routing”– Forwarding consists of taking a packet, looking at its destination

address, consulting the forwarding table, and sending the packet in a direction determined by the table

• Very easy once the forwarding table has been built

– Routing is the process by which the forwarding table is built• Need a routing protocol to dynamically build and maintain the table

A

ED

CB

F

2

2

13

1

1

2

53

5

Dest. Next Hop Cost

B B 2D D 1C D 3E D 2F D 4

Forwarding table in A

11

21

Routing Algorithm classification

Global or decentralized information?Global:

• all routers have complete topology, link cost info

• “link state” algorithms

Decentralized:

• router knows physically-connected neighbors, link costs to neighbors

• iterative process of computation, exchange of info with neighbors

• “distance vector” algorithms

22

Link-State Algorithms: General Idea

• Have each router build the complete topology of the network

• Once the complete topology is built, have each router run an algorithm to compute the shortest path from itself to all other routers (nodes) in the network

A

ED

CB

F

2

2

13

1

1

2

53

5

2 Issues

• How does a router build the complete topology of the network?

• How does a router compute the shortest path to all other nodes in the network using this topology information?– Dijkstra’s Shortest Path Algorithm

12

23

Building the Network Topology• At the heart of a link state algorithm is the discovery

of a node’s links’ states– Each node is assumed to capable of finding out the state of

the link to its neighbors (up or down) and the cost of each link

– Each node creates an update packet, also called a link-state packet (LSP) and periodically sends this information to all of its neighbors

– Node’s neighbors send the packet to their neighbors and so on until the LSP is received by all nodes in the network

A

ED

CB

F

2

2

13

1

1

2

53

5LSP Contents• ID of the node -- A

• List of neighbors and the cost to each neighbor – (B, 2), (C, 5), (D, 1)

• A sequence number

• A time-to-live (TTL) 24

Reliable Flooding• Reliable flooding is the process of making sure that all nodes get a

copy of LSP from all other nodes in the network

• When a node X receives a copy of an LSP that originated at some other node Y, it checks to see if it has already stored a copy of an LSP from Y.

• If not, it stores the LSP.

• If it already has a copy, it compares the SeqNos; if the new LSP has a larger SeqNo, it is assumed to be more recent, and the last LSP is stored replacing the old one. Otherwise the new LSP is discarded

A

ED

CB

F

2

2

13

1

1

2

53

5• The new LSP is then forwarded on all neighbors of X except the neighbor from which the LSP was just received

13

25

Computing the Shortest Path: Dijkstra’s Algorithm

Notation:• Cost(i,j): link cost from node i to j.

– Cost(A, B) = 2– Cost(A, C) = 5– Cost(A, F) = infinity

• Distance(v): current value of cost of path from source to destination V– Distance(D) = 1

• Pred(v): predecessor node of v along path from source to v– Pred(D) = A

• N: set of nodes whose least cost path definitively known

A

ED

CB

F

2

2

13

1

1

2

53

5

26

Dijsktra’s Algorithm – where A is the source node

1 Initialization:

2 N = {A}

3 for all nodes v

4 if v adjacent to A

5 then Distance (v) = cost(A,v)

6 else Distance (v) = infinity

7

8 Loop

9 find w not in N such that Distance(w) is a minimum

10 add w to N

11 update Distance(v) for all v adjacent to w and not in N:

12 Distance(v) = min( Distance(v), Distance(w) + cost(w,v) )

13 /* new distance to v is either old distance to v or known

14 shortest path distance to w plus cost from w to v */

15 until all nodes in N

14

27

Dijkstra’s algorithm: example

Step

0

1

2

3

4

5

N

A

AD

ADE

ADEB

ADEBC

ADEBCF

D(B),p(B)

2,A

2,A

2,A

D(C),p(C)

5,A

4,D

3,E

3,E

D(D),p(D)

1,A

D(E),p(E)

infinity

2,D

D(F),p(F)

infinity

infinity

4,E

4,E

4,E

A

ED

CB

F

2

2

13

1

1

2

53

5

Algorithm complexity: n nodes• each iteration: need to check all nodes, w, not in N• n*(n+1)/2 comparisons: O(n2) – O(nlogn) algorithm possible 28

Distance Vector Algos: General Idea

• We do NOT need to know the complete topology of the network to build the forwarding table!

• If a node X simply tells its neighbor Y the cost of reaching another node Z via itself (X), then– neighbor Y can compute the cost of reaching Z via X by

simply adding the cost of reaching X from Y and the cost of reaching Z from X, which was advertised by X

A

ED

CB

F

2

2

13

1

1

2

53

5Example

• If D tells A that it can reach E with a cost of 1, then A knows it can reach E via D with a cost of 1+1 = 2

• If D tells A that it can reach F with a cost of 3, then A knows it can reach F via D with a cost of 1+3 = 4

15

29

Distance Vector Routing Algorithm• Each node tells its

neighbors the cost of reaching every other node via itself

• A node computes its cost of reaching a destination via each of its neighbors and picks the best one

distributed:• each node communicates

only with directly-attached neighbors

iterative:• continues until no nodes

exchange new information• self-terminating: no

“signal” to stop

D (Y,Z)X

distance from X to

Y via Z as next hop

c(X,Z) + D (Z)Y

=

=

D (Y)X

min {D (Y,w)}X

w=

X Y ZC(X,Y)

D (Z)Y

T

D (Z)T

30

Distance Table StructureDistance Table Data Structure• each node has its own distance table• row for each possible destination• column for each directly-attached

neighbor to node• example: in node X, for dest. Y via

neighbor Z:

A

E D

CB7

8

1

2

1

2

D ()

A

B

C

D

A

1

7

6

4

B

14

8

9

11

D

5

5

4

2

Ecost to destination via

16

31

Distance Table gives Forwarding Table

D ()

A

B

C

D

A

1

7

6

4

B

14

8

9

11

D

5

5

4

2

Ecost to destination via

A

B

C

D

A,1

D,5

D,4

D,4

Outgoing link

to use, cost

Distance Table Forwarding (Routing) Table

32

Distance Vector Routing: overviewEach node:

wait for (msg from neighbor)

recompute distance table

if least cost path to any dest

has changed, notifyneighbors

Initialize the Distance Table Structure

17

33

Initialization

1 Initialization:

2 for all adjacent nodes v:

3 D (*,v) = infinity /* the * operator means "for all rows" */

4 D (v,v) = c(X,v)

5 for all destinations, y

6 send D (y) to each neighbor /* Cost of reaching Y via X */

X

X

X

At all nodes, X:

34

Distance Vector Algorithm: Example

X Z

12

7

Y

This columnshows theInit. results

InitInfo sent

We need to adjustvalues

if necessary

18

35

Distance Vector Algorithm (cont.):8 loop

9 wait (until I receive update from neighbor V)

10 if (update received from neighbor V wrt destination Y)

11 /* shortest path from V to some Y has changed */

12 /* V has sent a new value for its DV(Y) */

13 /* call this received new value is "newval" */

14 for the single destination Y: D (Y,V) = c(X,V) + newval

15

16 if we have a new D (Y) for any destination Y

17 send new value of D (Y) to all neighbors

18

19 forever

X

X

X

36

X Z

12

7

Y

D (Y,Z)X

c(X,Z) + D (Y)=

= 7+1 = 8

Z

D (Z,Y)X

c(X,Y) + D (Z)=

= 2+1 = 3

Y

Distance Vector Algorithm: Exampleinfo sent

Initial values Adjusted new value

19

37

Distance Vector Algorithm: Example

X Z

12

7

Y

This columnShows theInit. results

InitInfo sent

Adjusted new

values

NewInfo sent

Adjustvalues

if necessary

38

Hierarchical RoutingOur routing study thus far assumed

• all routers to be “identical” and the network to be “flat”

• 2 Problems exist with such a model:

1. scale: with 200 million destinations:– can’t store all destinations in routing tables!

– routing table exchange would swamp links!

2. administrative autonomy– Can’t assume that a network with the scale of Internet will be

administered by a single authority

• Solution?– Internet is NOT flat, but is a network of networks

– each network admin controls routing in its own network – next

– http://www.ilkertemir.com/backbone-map/• Turkey’s past and current Internet backbone map

20

39

Hierarchical Routing

a

b

b

aaC

A

Bd

A.a

A.c

C.bB.a

c

b

c

• aggregate routers into regions, “autonomous systems” (AS)

• routers in same AS run same routing protocol– “intra-AS” routing protocol

– routers in different AS can run different intra-AS routing protocol

gateway routers• special routers in AS

• run intra-AS routing protocol with all other routers in AS

• also responsible for routing to destinations outside AS

– run inter-AS routingprotocol with other gateway routers 40

Intra-AS Routing

• Also known as Interior Gateway Protocols (IGP)

• Most common Intra-AS routing protocols:

– OSPF: Open Shortest Path First

– RIP: Routing Information Protocol

– IGRP: Interior Gateway Routing Protocol (Cisco proprietary)

21

41

OSPF (Open Shortest Path First)

• “open”: publicly available

• Uses Link State algorithm – LS advertisement dissemination to entire AS via flooding

– Topology map at each node

– Route computation using Dijkstra’s algorithm

– Carried in OSPF messages directly over IP • OSPF has its own network layer protocol number like IP!

42

RIP (Routing Information Protocol)

• Distance vector algorithm

• Included in BSD-UNIX Distribution in 1982

• Distance metric: # of hops (max = 15 hops)– 16 is the infinity to eliminate routing loops

• Distance vectors: exchanged among neighbors every 30 sec via Response Message (also called advertisement)

• Each advertisement: list of up to 25 destination nets within AS– If more destinations, send multiple advertisements

22

43

RIP: Example

Destination Network Next Router Num. of hops to dest.

w A 2y B 2z B 7x -- 1…. …. ....

w x y

z

A

C

D B

Routing table in D

44

RIP: Example

Destination Network Next Router Num. of hops to dest.

w A 2y B 2z B A 7 5x -- 1…. …. ....

Routing table in D

w x y

z

A

C

D B

Dest Next hopsw - -x - -z C 4…. … ...

Advertisementfrom A to D

23

45

RIP Table processing

• RIP routing tables managed by application-levelprocess called route-d (daemon)

• advertisements sent in UDP packets, periodically repeated

physical

link

network forwarding(IP) table

Transprt(UDP)

routed

physical

link

network(IP)

Transprt(UDP)

routed

forwardingtable

46

Inter-AS routing in the Internet: BGP

Figure 4.5.2-new2: BGP use for inter-domain routing

AS2 (OSPF

intra-AS

routing)

AS1 (RIP intra-AS

routing) BGP

AS3 (OSPF intra-AS

routing)

BGP

R1 R2

R3

R4

R5

• Each BGP router communicates only with its neighbors– R1 with R2, R3 with R4

– Global info about routes to destination networkspropagates in an AS-by-AS manner via the exchange of BGP messages

24

47

Why different Intra- and Inter-AS routing ?

Policy:

• Intra-AS: single admin, so no policy decisions needed

• Inter-AS: admin wants control over how its traffic routed, who routes through its net.

Performance:

• Intra-AS: can focus on performance

• Inter-AS: policy may dominate over performance