Routing4-1 Introduction To Computer Networks מבוא לרשתות תקשורת 89-350 ניתוב (Routing) פרופ ' אמיר הרצברג Chapter 4, Network Layer Computer Networking:

Routing 4-1

Introduction To Computer Networks

-89מבוא לרשתות תקשורת 350

(Routing)ניתוב פרופ' אמיר הרצברג

Chapter 4, Network LayerComputer Networking: A Top Down Approach Featuring the Internet, 3rd edition. Jim Kurose, Keith RossAddison-Wesley, July 2005. ** Group purchase ?? **

Based on foils by Kurose & Ross ©, see: http://www.aw.com/kurose-ross/

My site: http://AmirHerzberg.com

Course site http://www.cs.biu.ac.il/~89-350/

Routing 4-2

Routing: Agenda Network layer services: datagram vs. virtual

circuit Link-state routing: OSPF Distance-vector routing: RIP Hierarchical and Inter-AS routing: BGP

Routing 4-3

Network layer functions

transport packet from sending to receiving hosts

network layer protocols in every host, router

three important functions: path determination: route

taken by packets from source to dest. Routing algorithms

call setup: some network architectures require router call setup along path before data flows

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

networkdata linkphysical

networkdata linkphysical

networkdata linkphysical

networkdata linkphysical

networkdata linkphysical

networkdata linkphysical

networkdata linkphysical

networkdata linkphysical

application

transportnetworkdata linkphysical

application

transportnetworkdata linkphysical

ThisLecture

IP

Routing 4-4

Network service modelQ: What service model for

“channel” transporting packets from sender to receiver?

Two main service abstractions:

Datagram vs. Virtual Circuit

A1: Datagram: best effort only

• IP service

A2: Virtual Circuit… any of: • loss-free delivery• congestion feedback to sender• in-order delivery• guaranteed bandwidth• no jitter• Like phone

Routing 4-5

Virtual Circuits (VC) Service

call setup before data can flow, teardown after Packets carry VC identifier (not destination host ID) every router/switch on source-dest path maintains

“state” for connections transport-layer connection only involved two end systems

link, router resources (bandwidth, buffers) may be allocated to VC for circuit-like performance, guaranteed bandwidth/delay

“like telephone circuit” performance-wise network actions along source-to-dest path

Routing 4-6

Virtual circuits: signaling protocols used to setup, maintain and teardown VC

Teardown may initiate at either end – or any router along VC Exercise: design simple VC protocol

used in ATM, frame-relay, SNA, X.25 Also `premium` Internet services (RSVP, MPLS)

application

transportnetworkdata linkphysical

application

transportnetworkdata linkphysical

1. Initiate call 2. incoming call

3. Accept call4. Call connected5. Data flow begins 6. Receive data

Routing 4-7

Datagram networks: the Internet model no call setup at network layer routers: no state about end-to-end connections

no network-level concept of “connection”

packets forwarded using destination host address packets in same connection may take different paths

application

transportnetworkdata linkphysical

application

transportnetworkdata linkphysical

1. Send data 2. Receive data

Routing 4-8

Datagram or VC network: why?

Datagram (IP) data exchange among

computers “elastic” service, no

strict timing req. “smart” end systems

(computers) can adapt, perform

control, error recovery simple inside network,

complexity at “edge” many link types

different characteristics uniform service difficult

VC (ATM, MPLS) evolved from telephony “dumb” end systems

telephones complexity in network

human conversation: strict timing, reliability

requirements

Guaranteed service

Routing 4-9

Routing: Agenda Network layer services: datagram vs. virtual

circuit Link-state routing: OSPF Distance-vector routing: RIP Hierarchical and Inter-AS routing: BGP

Routing 4-10

Routing

Graph abstraction for routing algorithms:

graph nodes are routers

graph edges are channels In Internet: a network link cost: delay, $ cost,

congestion level, reliability… [measure!]

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 Cost is not always additive Other def’s possible

Routing 4-11

Routing Algorithm classification

Global or decentralized information, decisions?

Global: all routers have complete

topology, link cost info “link state” algorithmsDecentralized: router knows physically-

connected neighbors, link costs to neighbors

iterative process of computation, exchange of info with neighbors

“distance vector” algorithms

Static or dynamic?Static: Does not handle changes

/ failures Implicitly: restart upon

change (failure/recovery)Dynamic: Change routes (due

changes in costs, topology) periodically in response to change

in cost / topology / faults/ recoveries

Routing 4-12

Link-State Routing: Shortest Path AlgorithmsGlobal, Static Link-State Source Routing: Link state is fixed: c)i,j(, cost from i to j (∞ if no link)

Known to each source (A) [E.g. with Flood (PI) or PIF – later] Source routing: find path from each source A to all other

nodes

Single Source Shortest Paths Tree Problem: Given graph G=(V,E), source node AV Find shortest-paths tree TG for A; for each node v:

DA)v(: cost of shortest path from source A to v• We simplify and write D(v) for DA(v)

All paths in T are shortest length: D(v)=∑(u,w)path(A,v)c(u,w) NOT Minimum Spanning Tree (Exercise: show why!) Not same tree for all sources (why?)

Routing 4-13

Dijsktra’s AlgorithmBuilds shortest-paths tree T iterativelyAfter k iterations, T contains k links (and k+1 nodes)

Given graph G=(V,E), source node AV, and link cost function c(x,y)0

let T ={ A }Let D(x)={0 if x=A, ∞ otherwise)while (T does not contain all nodes) { find edge (x,y)E with x in T and y not in T minimizing D(x)+c(x,y) add (x,y) to T D(y)=D(x)+c(x,y) }Output T, D

This usually holds… Why needed?Alg for negative costs?

Routing 4-14

Dijkstra’s algorithm: example

Step012345

NodesA

ADADE

ADEBADEBC

ADEBCF

D)B(222

D)C(5433

D)D(1

D)E(infinity

2

D)F(infinityinfinity

444

A

ED

CB

F

2

2

13

1

1

2

53

5

Routing 4-15

Dijkstra’s algorithm, discussionAlgorithm complexity: n nodes, static link costs each iteration: need to check all nodes, w, not in T n*(n+1)/2 comparisons: O(n2) more efficient implementations (e.g. heap): O(nlogn)

Dynamic costs (e.g. c(i,j)=|traffic(i,j)|) may oscillate: B,C,D send to A (1,e,1 units), each use Dijkstra Initial, sensible state: DA, CBA Oscillate between: BCDA, DCBA !!

A

D

C

B1 1+e

e0

e

1 1

0 0

A

D

C

B2+e 0

001+e1

A

D

C

B0 2+e

1+e10 0

A

D

C

B2+e 0

e01+e1

initially… recompute

routing… recompute … recompute

Routing 4-16

OSPF (Open Shortest Path First)

“Open”: publicly available, standard routing protocol

Link State algorithm LS packet (`advertisement`) flooding Topology map at each node Route computation using Dijkstra’s

algorithm

Routing 4-17

OSPF Flooding Each node x sends Link State Packet (LSP)

periodically / on link state change, with: ID of x, list of direct neighbors and costs Sequence number (64 bit, assume to never wrap around) Time to live

Sequence number used to identify *newer* LSP An older LSP is discarded What if a router crash and sequence number reset to 0?

Receiving node floods LSP to all its neighbors Except neighbor where the LSP came from

Reliability via acknowledgement and retransmission Variant of PI (Propagation of Information) protocol

In recitation: PI, PIF (Prop. of Info. w/ Feedback)

Routing 4-18

OSPF “advanced” features

Multiple cost metrics For different Type Of Service [TOS] E.g., satellite: high bandwidth, but high latency

Collected topology allows… Multiple same-cost pathsMulticast OSPF (MOSPF)

Hierarchical OSPF in large domains. Security: all OSPF messages authenticated

(to prevent malicious intrusion)

Routing 4-19

Routing: Agenda Network layer services: datagram vs. virtual

circuit Link-state routing: OSPF Distance-vector routing: RIP Hierarchical and Inter-AS routing: BGP

Routing 4-20

Distance Vector Routing Algorithm

Iterative, Dynamic: continues until no

nodes exchange info. self-terminating: no

“signal” to stop restart upon changeAsynchronous: nodes need not

exchange info/iterate in lock step!

Distributed: each node

communicates only with directly-attached neighbors

Distance Table data structure

each node has its own table row for each possible

destination column for each neighbor example: in node X, for dest. Y

via neighbor Z:

D )Y,Z(X

distance from X toY, via Z as next hop

c)X,Z( + min {D )Y,w(}Z

w

=

=

Routing 4-21

Distance Table: example

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

dest

inat

ion

D )C,D(E

c)E,D( + min {D )C,w(}D

w== 2+2 = 4

D )A,D(E

c)E,D( + min {D )A,w(}D

w== 2+3 = 5

D )A,B(E

c)E,B( + min {D )A,w(}B

w== 8+6 = 14

loop!

loop!

Routing 4-22

Distance table gives routing 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

dest

inat

ion

A

B

C

D

A,1

D,5

D,4

D,4

Outgoing link to use, cost

dest

inat

ion

Distance table Routing table

Routing 4-23

Distance Vector Routing: overview

Iterative, asynchronous: each local iteration caused by:

local link cost change message from neighbor:

its least cost path change from neighbor

Distributed: each node notifies

neighbors only when its least cost path to any destination changes neighbors then notify

their neighbors if necessary

wait for )change in local link cost or msg from neighbor(

recompute distance table

if least cost path to any dest

has changed, notify neighbors

Each node:

Routing 4-24

Distance Vector Algorithm at node x:Initialization: for all nodes y and all neighbors v: DX)y,v( = {c)x,v( for y=v, else ∞}

On Detect_cost_of_link to neighbor v (i.e. c(x,v)) changes by d /* change cost to all dest's via neighbor v by d */ /* note: d could be positive or negative */ for all destinations y: DX(y,v) = DX(y,v) + d

On Rcv <y, D> from neighbor v: /* shortest path from v to some y has changed */ /* v sent a new value for its minWDV(y,w) */ DX(y,v) = c(x,v) + D

Every T seconds (or always, if T=0): If DX(y)=minWDX(y,w) changed (for any destination y), then send <y, DX(y)> to all neighbors

Routing 4-25

Distance Vector Algorithm: example

X Z12

7

Y

D )Y,Z(X

c)X,Z( + min {D )Y,w(}w=

= 7+1 = 8

Z

D )Z,Y(X

c)X,Y( + min {D )Z,w(}w=

= 2+1 = 3

Y

Routing 4-26

Distance Vector Algorithm: example

X Z12

7

Y

Routing 4-27

Distance Vector: link cost changesLink cost changes: node detects local link cost

change updates distance table if cost change in least cost path,

notify neighbors Example when DX)y,v( reduced:

X Z14

50

Y1

algorithmterminates(converges – no activity)

“goodnews travelsfast”(and noloops)

Routing 4-28

Distance Vector: Convergence If network stabilizes, then DV eventually outputs shortest paths (and halts)

Assuming FIFO, max delay Δ, min weight w [why?] Consider single, fixed target node y; DY(y)=0 Let x(0)=y, x(0), x(1),…x(n) be the nodes by order of min-

distance from y Claim: for all i (0≤i≤n), eventually Dx(i)(y)=D(i), the min-

distance from x(i) to y. Proof: by induction on i. Trivially holds for i=0 (i.e. x(0)=y). Assume from time t, there are no new `detect cost change`

events, and x(0)…x(i) stabilized to D(i). We prove the claim holds for i+1 as well. For every j>i, eventually Dx(j)(y)D(i+1) [why?] Then, eventually x(i+1) receives D(j) from x(j), s.t. j≤i and

D(i)=D(j)+c(x(j),x(i))… Claim follows! But… convergence can be veeery sloooooooow…

Routing 4-29

Slow convergence: “count to infinity”

Link cost changes: bad news travels slow!

Bad news = DX)y,v( increased (e.g. link down) Convergence time depends on cost ratios, not

just on nX Z

14

50

Y60

algorithmcontinues

on!

Packets sent from Z, Y to X will loop btw Z, Y!!

Routing 4-30

Distance Vector: poisoned reverse(fix to most `count to infinity` scenarios)

If Z routes through Y to get to X : Z tells Y its (Z’s) distance to X is infinite

(so Y won’t route to X via Z) Exercises: modify code; re-do convergence analysis Often helps – but complexity still depends on cost

ratios, not just on n; Exercise: demonstrate this!

X Z14

50

Y60

algorithmterminates

Routing 4-31

RIP ( Routing Information Protocol)

Standard distance vector algorithm Distance vectors exchanged among neighbors

every 30 sec via RIP advertisement Distance metric: # of hops

≤15hops; 16=unreachable Each advertisement: list of up to 25

destination nets Sent using UDP (why?)

Routing 4-32

RIP: Example

Destination Network Next Router Num. of hops to dest. w A 2

y B 2 z B 7

x -- 1…. …. ....

w x yA

C

D B

Routing table in D

z

Routing 4-33

RIP: Example

Destination Network Next Router Num. of hops to dest. w A 2

y B 2 z B A 7 5

x -- 1…. …. ....Routing table in D

Dest Next hops w - - x - - z C 4 …. … ...

Advertisementfrom A to D

w x yA

C

D Bz

Routing 4-34

RIP: Link Failure and Recovery If no advertisement heard after 180 sec -->

neighbor/link declared dead routes via neighbor invalidated new advertisements sent to neighbors neighbors in turn send out new advertisements

(if tables changed) link failure info quickly propagates to entire net poison reverse used to prevent ping-pong

loops• Send infinite distance (= 16 hops) to first hop toward

destination

Routing 4-36

Routing: Agenda Network layer services: datagram vs. virtual

circuit Link-state routing: OSPF Distance-vector routing: RIP Hierarchical and Inter-Autonomous Systems

routing: BGP

Routing 4-37

Hierarchical Routing and Autonomous Systems

scale: with 200 million destinations:

can’t store all dest’s in routing tables!

routing table exchange would swamp links!

administrative autonomy internet = network of

networks (Autonomous Systems)

AS= Autonomous System, a network under a single administrative control

Identified by 16-bit AS Number (ASN)

Our routing study thus far - idealization all routers identical network “flat”… not true in practice

Routing 4-38

Hierarchical Routing and AS

Aggregate routers By “autonomous

systems” (AS) Or: within a huge AS

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

protocol routers in different AS

can run different intra-AS routing protocol

special routers in AS run intra-AS routing

protocol with all other routers in AS

Use inter-AS routing protocol to route to Nets in other AS-es

BGP (Border Gateway Protocol): Internet’s inter-AS protocol

gateway routers

Routing 4-39

Intra-AS and Inter-AS routing

Gateways:•perform inter-AS routing amongst themselves•perform intra-AS routing with other routers in their AS

inter-AS, intra-AS routing in

gateway A.c

network layer

link layer

physical layer

a

b

b

aaC

A

Bd

A.a

A.c

C.bB.a

cb

c

Routing 4-40

Intra-AS and Inter-AS routing

Host h2

a

b

b

aaC

A

Bd c

A.a

A.c

C.bB.a

cb

Hosth1

Intra-AS routingwithin AS A

Inter-AS routingbetween A and B

Intra-AS routingwithin AS B

How to publish inter-AS route to hosts? 1st idea: gateway publish it as if it had a link, via Intra-AS protocol (RIP/OSPF/…)

O(E) messages per each entry (network prefix) ?

Routing 4-41

I-BGP: Efficient Distribution of BGP routes

Host h2

a

b

b

aaC

A

Bd c

A.a

A.c

C.bB.a

cb

Hosth1

E-BGP

I-BGP distribution of routes in AS B

How to publish inter-AS route to hosts? Solution: send to each router using I-BGP

No flooding inside AS Identify gateway; use Intra-AS routing to reach it

Routing 4-42

Routing with I-BGP and RIP/OSPF

AS 1 AS 2192.0.2.1

135.207.0.0/16

10.10.10.10

I-BGP: 135.207.0.0/16Next Hop = 192.0.2.1

192.0.2.0/30

Forwarding Table

135.207.0.0/16

destination next hop

10.10.10.10192.0.2.0/30 10.10.10.10

I-BGP

192.0.2.1135.207.0.0/16

destination next hop

+

Intra

10.10.10.10192.0.2.0/30

destination next hop

E-BGP

Routing 4-43

Connecting Autonomous Systems Types of Autonomous Systems

Transit AS (ISP): AS x that transfers traffic AS y to AS z

Non-Transit AS (customer): never transfers…

• Stub AS: small corporation: connected via single ISP • Multihomed AS: large corporation; connected via

multiple connections to other AS’s (at least one ISP – Transit AS)

Use Border Gateway Protocol (BGP) Focus on policies, not optimization E.g., allow multihomed AS to refuse to transit

Routing 4-44

Inter-AS routing with BGP(Border Gateway Protocol)

Path Vector protocol similar to Distance Vector protocol Broadcast to neighbors (peers) Broadcast announcements: list of

<Path,Attributes> Path is sequence of ASes to destination

(network prefix, e.g. xx.yy/16) Attributes provide information to select best

route

Routing 4-45

Internet inter-AS routing: BGP

Suppose: gateway X send its path to peer gateway W

W may or may not select path offered by Xcost, policy (don’t route via competitors

AS), loop prevention reasons.

If W selects path advertised by X, then:Path (W, xx.yy/16) = w, Path (X, xx.yy/16)

Control incoming traffic by controlling announcements to peers…

Routing 4-46

Customers and Providers

Customer pays provider for access to the Internet

provider

customerIP traffic

provider customer

Routing 4-47

Customers Don’t Always Need BGP

provider

customer

Default route 0.0.0.0/0pointing to provider.

Configured route 192.0.2.0/24pointing to customer

192.0.2.0/24

Static routing is the most common way of connecting anautonomous routing domain to the Internet. This helps explain why BGP is a mystery to many …

Routing 4-48

Customer-Provider Hierarchy

IP trafficprovider customer

Routing 4-49

The Peering Relationship

peer peer

customerprovider

Peers provide transit between their respective customers

Peers do not provide transit between peers

Peers (often) do not exchange $$$trafficallowed

traffic NOTallowed

Routing 4-50

Peering Provides Shortcuts

Peering also allows connectivity betweenthe customers of “Tier 1” providers.

peer peer

customerprovider

4-51

BGP Operations (Simplified)

Establish session on TCP port 179

Exchange all active routes

Exchange incremental updates

AS1

AS2

While connection is ALIVE exchangeroute UPDATE messages

BGP session

Routing 4-52

Why different Intra- and Inter-AS routing ?

Policy: Inter-AS: admin wants control over how its traffic

routed, who routes through its net. Intra-AS: single admin, so no policy decisions

needed

Scale: hierarchical routing saves table size, reduced

update trafficPerformance: Intra-AS: can focus on performance Inter-AS: policy may dominate over performance

Routing 4-53

Routing: summary

Next and final stop: the Data

link layer!

What we’ve covered: network layer services routing principles: link state

and distance vector hierarchical routing Internet routing protocols RIP,

OSPF, BGP