Routing Algorithms and Routing

in the Internet

Network Layer 4-1

Interplay between routing and
forwarding
routing algorithm

local forwarding table

header value output link
0100 3
0101 2
0111 2
1001 1

value in arriving
packet’s header
0111 1

3 2

Network Layer 4-2

Graph abstraction
5
3
v w 5
2
u 2 1 z
3
1 2
Graph: G = (N,E)
x 1
y

N = set of routers = { u, v, w, x, y, z }

E = set of links ={ (u,v), (u,x), (v,x), (v,w), (x,w), (x,y), (w,y), (w,z), (y,z) }

Remark: Graph abstraction is useful in other network contexts

Example: P2P, where N is set of peers and E is set of TCP connections

Network Layer 4-3

Graph abstraction: costs
5
• c(x,x’) = cost of link (x,x’)
3
v w 5
2 - e.g., c(w,z) = 5
u 2 1 z
3 • cost could always be 1, or
1 2 inversely related to bandwidth,
x 1
y
or inversely related to
congestion

Cost of path (x1, x2, x3,…, xp) = c(x1,x2) + c(x2,x3) + … + c(xp-1,xp)

Question: What’s the least-cost path between u and z ?

Routing algorithm: algorithm that finds least-cost path

Network Layer 4-4

Routing Algorithm classification
Global or decentralized Static or dynamic?
information?
Static:
Global:
 all routers have complete  routes change slowly
topology, link cost info over time
 “link state” algorithms
Dynamic:
Decentralized:
 routes change more
 router knows physically-
connected neighbors, link quickly
costs to neighbors  periodic update
 iterative process of
 in response to link
computation, exchange of info
with neighbors cost changes
 “distance vector” algorithms

Network Layer 4-5

A Link-State Routing Algorithm

Dijkstra’s algorithm Notation:

 net topology, link costs  c(x,y): link cost from node
known to all nodes x to y; = ∞ if not direct
 accomplished via “link
neighbors
state broadcast”  D(v): current value of cost
 all nodes have same info
of path from source to
 computes least cost paths
dest. v
from one node (‘source”) to
 p(v): predecessor node
all other nodes
 gives forwarding table along path from source to v
for that node  N': set of nodes whose
 iterative: after k iterations, least cost path definitively
know least cost path to k known
dest.’s

Network Layer 4-6

Dijsktra’s Algorithm
1 Initialization:
2 N' = {u}
3 for all nodes v
4 if v adjacent to u
5 then D(v) = c(u,v)
6 else D(v) = ∞
7
8 Loop
9 find w not in N' such that D(w) is a minimum
10 add w to N'
11 update D(v) for all v adjacent to w and not in N' :
12 D(v) = min( D(v), D(w) + c(w,v) )
13 /* new cost to v is either old cost to v or known
14 shortest path cost to w plus cost from w to v */
15 until all nodes in N'

Network Layer 4-7

 Dijkstra’s algorithm: example
Step N' D(v),p(v) D(w),p(w) D(x),p(x) D(y),p(y) D(z),p(z)
0 u 2,u 5,u 1,u ∞ ∞
1 ux 2,u 4,x 2,x ∞
2 uxy 2,u 3,y 4,y
3 uxyv 3,y 4,y
4 uxyvw 4,y
5 uxyvwz

5
3
v w 5
2
u 2 1 z
3
1 2
x 1
y

Network Layer 4-8

 Dijkstra’s algorithm, discussion
Algorithm complexity: n nodes
 each iteration: need to check all nodes, w, not in N
 n(n+1)/2 comparisons: O(n2)
 more efficient implementations possible: O(nlogn)

Oscillations possible:
 e.g., link cost = amount of carried traffic

1 A A A A
1+e 2+e 0 0 2+e 2+e 0
D 0 0 B D 1+e 1 B D B D 1+e 1 B
0 0
0 e 0 0 1 1+e 0 e
1
C C C C
1
e
… recompute … recompute … recompute
initially
routing
Network Layer 4-9
Distance Vector Algorithm (1)
Bellman-Ford Equation (dynamic programming)
Define
dx(y) := cost of least-cost path from x to y

Then

dx(y) = min {c(x,v) + dv(y) }

where min is taken over all neighbors of x

Network Layer 4-10
Bellman-Ford example (2)
5
3
Clearly, dv(z) = 5, dx(z) = 3, dw(z) = 3
v w 5
2
u 2 1 z B-F equation says:
3
1 du(z) = min { c(u,v) + dv(z),
2
1
xy
c(u,x) + dx(z),
c(u,w) + dw(z) }
= min {2 + 5,
1 + 3,
5 + 3} = 4
Node that achieves minimum is next
hop in shortest path ➜ forwarding table
Network Layer 4-11
Distance Vector Algorithm (3)
 Dx(y) = estimate of least cost from x to y
 Distance vector: Dx = [Dx(y): y є N ]
 Node x knows cost to each neighbor v:
c(x,v)
 Node x maintains Dx = [Dx(y): y є N ]
 Node x also maintains its neighbors’
distance vectors
 For each neighbor v, x maintains
Dv = [Dv(y): y є N ]

Network Layer 4-12

 Distance vector algorithm (4)
Basic idea:
 Each node periodically sends its own distance
vector estimate to neighbors
 When node a node x receives new DV estimate
from neighbor, it updates its own DV using B-F
equation:
Dx(y) ← minv{c(x,v) + Dv(y)} for each node y ∊ N

 Under minor, natural conditions, the estimate Dx(y)

converge the actual least cost dx(y)

Network Layer 4-13

Distance Vector Algorithm (5)
Iterative, asynchronous: Each node:
each local iteration caused
by:
 local link cost change wait for (change in local link
 DV update message from cost of msg from neighbor)
neighbor
Distributed: recompute estimates
 each node notifies
neighbors only when its DV
changes if DV to any dest has
 neighbors then notify changed, notify neighbors
their neighbors if
necessary

Network Layer 4-14

 Dx(y) = min{c(x,y) + Dy(y), c(x,z) + Dz(y)} Dx(z) = min{c(x,y) +
= min{2+0 , 7+1} = 2 Dy(z), c(x,z) + Dz(z)}
node x table = min{2+1 , 7+0} = 3
cost to cost to cost to
x y z x y z x y z
x 0 2 7 x 0 2 3 x 0 2 3
from

from

from
y ∞∞ ∞ y 2 0 1 y 2 0 1
z ∞∞ ∞ z 7 1 0 z 3 1 0
node y table
cost to cost to cost to
x y z x y z x y z y
2 1
x ∞ ∞ ∞ x 0 2 7 x 0 2 3 x z
from
from

from
y 2 0 1 y 2 0 1 y 2 0 1 7
z ∞∞ ∞ z 7 1 0 z 3 1 0
node z table
cost to cost to cost to
x y z x y z x y z

x ∞∞ ∞ x 0 2 7 x 0 2 3
from

from
from

y ∞∞ ∞ y 2 0 1 y 2 0 1
z 71 0 z 3 1 0 z 3 1 0
time
Network Layer 4-15
 Distance Vector: link cost changes
Link cost changes: 1
 node detects local link cost change y
4 1
 updates routing info, recalculates
x z
distance vector 50
 if DV changes, notify neighbors

At time t0, y detects the link-cost change, updates its DV,

and informs its neighbors.
“good
news At time t1, z receives the update from y and updates its table.
travels It computes a new least cost to x and sends its neighbors its DV.

fast” At time t2, y receives z’s update and updates its distance table.
y’s least costs do not change and hence y does not send any
message to z.

Network Layer 4-16

Distance Vector: link cost changes
Link cost changes:
 good news travels fast
 bad news travels slow - “count to infinity” problem! 60
 44 iterations before algorithm stabilizes: see text
y
Poissoned reverse:
 If Z routes through Y to get to X :
4 1
 Z tells Y its (Z’s) distance to X is infinite (so Y won’t route to X via Z)
x z
50
 will this completely solve count to infinity problem?

Network Layer 4-17

Comparison of LS and DV algorithms
Message complexity Robustness: what happens if
 LS: with n nodes, E links, O(nE) router malfunctions?
msgs sent LS:
 DV: exchange between  node can advertise incorrect
neighbors only link cost
 convergence time varies  each node computes only its
own table
Speed of Convergence
 LS: O(n2) algorithm requires DV:
O(nE) msgs
 DV node can advertise
 may have oscillations
incorrect path cost
 DV: convergence time varies
 each node’s table used by
others
 may be routing loops
• error propagate thru
 count-to-infinity problem network

Network Layer 4-18

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

scale: with 200 million administrative autonomy

destinations:  internet = network of
 can’t store all dest’s in networks
routing tables!  each network admin may
 routing table exchange want to control routing in its
would swamp links! own network

Network Layer 4-19

Hierarchical Routing
 aggregate routers into
Gateway router
regions, “autonomous  Direct link to router in
systems” (AS)
another AS
 routers in same AS run
same routing protocol
 “intra-AS” routing
protocol
 routers in different AS
can run different intra-
AS routing protocol

Network Layer 4-20

Interconnected ASes

3c
3a 2c
3b 2a
AS3 2b
1c AS2
1a 1b
1d AS1  Forwarding table is
configured by both
intra- and inter-AS
routing algorithm
Intra-AS Inter-AS
Routing Routing
algorithm algorithm
 Intra-AS sets entries
Forwarding for internal dests
table
 Inter-AS & Intra-As
sets entries for
external dests
Network Layer 4-21
Inter-AS tasks AS1 needs:
 Suppose router in AS1 1. to learn which dests
receives datagram for are reachable through
which dest is outside AS2 and which
of AS1 through AS3
 Router should forward 2. to propagate this
packet towards on of
reachability info to all
the gateway routers,
but which one? routers in AS1
Job of inter-AS routing!

3c
3a 2c
3b 2a
AS3 2b
1c AS2
1a 1b
1d AS1
Network Layer 4-22
Example: Setting forwarding table
in router 1d
 Suppose AS1 learns from the inter-AS
protocol that subnet x is reachable from
AS3 (gateway 1c) but not from AS2.
 Inter-AS protocol propagates reachability
info to all internal routers.
 Router 1d determines from intra-AS
routing info that its interface I is on the
least cost path to 1c.
 Puts in forwarding table entry (x,I).

Network Layer 4-23

Example: Choosing among multiple ASes
 Now suppose AS1 learns from the inter-AS protocol
that subnet x is reachable from AS3 and from AS2.
 To configure forwarding table, router 1d must
determine towards which gateway it should forward
packets for dest x.
 This is also the job on inter-AS routing protocol!
 Hot potato routing: send packet towards closest of
two routers.

Use routing info Determine from

Learn from inter-AS Hot potato routing: forwarding table the
from intra-AS
protocol that subnet Choose the interface I that leads
protocol to
x is reachable via gateway to least-cost gateway.
determine
multiple gateways that has the Enter (x,I) in
costs of least-cost
smallest least cost forwarding table
paths to each
of the gateways

Network Layer 4-24

Intra-AS Routing
 Also known as Interior Gateway Protocols (IGP)
 Most common Intra-AS routing protocols:

 RIP: Routing Information Protocol

 OSPF: Open Shortest Path First

 IGRP: Interior Gateway Routing Protocol (Cisco

proprietary)

Network Layer 4-25

RIP ( Routing Information Protocol)
 Distance vector algorithm
 Included in BSD-UNIX Distribution in 1982
 Distance metric: # of hops (max = 15 hops)

u destination hops
v
u 1
A B w v 2
w 2
x 3
x y 3
z C D z 2
y

Network Layer 4-26

RIP advertisements
 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

Network Layer 4-27

RIP: Example
z
w x y
A D B

C
Destination Network Next Router Num. of hops to dest.
w A 2
y B 2
z B 7
x -- 1
…. …. ....
Routing table in D

Network Layer 4-28

 RIP: Example
Dest Next hops
w - - Advertisement
x - - from A to D
z C 4
…. … ...
z
w x y
A D B

C
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 Network Layer 4-29
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
(infinite distance = 16 hops)

Network Layer 4-30

RIP Table processing
 RIP routing tables managed by application-level
process called route-d (daemon)
 advertisements sent in UDP packets, periodically
repeated
routed routed

Transprt Transprt
(UDP) (UDP)
network forwarding forwarding network
(IP) table table (IP)
link link
physical physical

Network Layer 4-31

OSPF (Open Shortest Path First)
 “open”: publicly available
 Uses Link State algorithm
 LS packet dissemination
 Topology map at each node
 Route computation using Dijkstra’s algorithm

 OSPF advertisement carries one entry per neighbor

router
 Advertisements disseminated to entire AS (via
flooding)
 Carried in OSPF messages directly over IP (rather than TCP
or UDP

Network Layer 4-32

OSPF “advanced” features (not in RIP)

 Security: all OSPF messages authenticated (to

prevent malicious intrusion)
 Multiple same-cost paths allowed (only one path in
RIP)
 For each link, multiple cost metrics for different
TOS (e.g., satellite link cost set “low” for best
effort; high for real time)
 Integrated uni- and multicast support:
 Multicast OSPF (MOSPF) uses same topology data
base as OSPF
 Hierarchical OSPF in large domains.
Network Layer 4-33
Hierarchical OSPF

Network Layer 4-34

Hierarchical OSPF
 Two-level hierarchy: local area, backbone.
 Link-state advertisements only in area
 each nodes has detailed area topology; only know
direction (shortest path) to nets in other areas.
 Area border routers: “summarize” distances to nets
in own area, advertise to other Area Border routers.
 Backbone routers: run OSPF routing limited to
backbone.
 Boundary routers: connect to other AS’s.

Network Layer 4-35

Internet inter-AS routing: BGP
 BGP (Border Gateway Protocol): the de
facto standard
 BGP provides each AS a means to:
1. Obtain subnet reachability information from
neighboring ASs.
2. Propagate the reachability information to all
routers internal to the AS.
3. Determine “good” routes to subnets based on
reachability information and policy.
 Allows a subnet to advertise its existence
to rest of the Internet: “I am here”

Network Layer 4-36

 BGP basics
 Pairs of routers (BGP peers) exchange routing info over semi-
permanent TCP conctns: BGP sessions
 Note that BGP sessions do not correspond to physical links.
 When AS2 advertises a prefix to AS1, AS2 is promising it will
forward any datagrams destined to that prefix towards the
prefix.
 AS2 can aggregate prefixes in its advertisement

3c
3a 2c
3b 2a
AS3 2b
1c AS2
1a 1b
AS1 1d
eBGP session
iBGP session
Network Layer 4-37
Distributing reachability info
 With eBGP session between 3a and 1c, AS3 sends prefix
reachability info to AS1.
 1c can then use iBGP do distribute this new prefix reach info
to all routers in AS1
 1b can then re-advertise the new reach info to AS2 over the
1b-to-2a eBGP session
 When router learns about a new prefix, it creates an entry
for the prefix in its forwarding table.

3c
3a 2c
3b 2a
AS3 2b
1c AS2
1a 1b
AS1 1d
eBGP session
iBGP session
Network Layer 4-38
Path attributes & BGP routes
 When advertising a prefix, advert includes BGP
attributes.
 prefix + attributes = “route”
 Two important attributes:
 AS-PATH: contains the ASs through which the advert
for the prefix passed: AS 67 AS 17
 NEXT-HOP: Indicates the specific internal-AS router to
next-hop AS. (There may be multiple links from current
AS to next-hop-AS.)
 When gateway router receives route advert, uses
import policy to accept/decline.

Network Layer 4-39

BGP route selection
 Router may learn about more than 1 route
to some prefix. Router must select route.
 Elimination rules:
1. Local preference value attribute: policy
decision
2. Shortest AS-PATH
3. Closest NEXT-HOP router: hot potato routing
4. Additional criteria

Network Layer 4-40

BGP messages
 BGP messages exchanged using TCP.
 BGP messages:
 OPEN: opens TCP connection to peer and
authenticates sender
 UPDATE: advertises new path (or withdraws old)
 KEEPALIVE keeps connection alive in absence of
UPDATES; also ACKs OPEN request
 NOTIFICATION: reports errors in previous msg;
also used to close connection

Network Layer 4-41

 BGP routing policy
legend: provider
B network
X
W A
customer
C network:

Figure 4.5-BGPnew: a simple BGP scenario

 A,B,C are provider networks
 X,W,Y are customer (of provider networks)
 X is dual-homed: attached to two networks
X does not want to route from B via X to C
 .. so X will not advertise to B a route to C

Network Layer 4-42

 BGP routing policy (2)
legend: provider
B network
X
W A
customer
C network:

Figure 4.5-BGPnew: a simple BGP scenario

 A advertises to B the path AW
 B advertises to X the path BAW
 Should B advertise to C the path BAW?
 No way! B gets no “revenue” for routing CBAW since neither
W nor C are B’s customers
 B wants to force C to route to w via A
 B wants to route only to/from its customers!

Network Layer 4-43

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
traffic
Performance:
 Intra-AS: can focus on performance
 Inter-AS: policy may dominate over performance

Network Layer 4-44