CS 640: Introduction to

Computer Networks
Aditya Akella

Lecture 10 -
Intra-Domain Routing

From previous lecture….

Three Types of Switching Fabrics

Switching Via a Memory

First generation routers
looked like PCs
• Packet copied by system’s (single) CPU
• Speed limited by memory bandwidth (2 bus crossings
per datagram)
Input Memory Output
Port Port

System Bus

Most modern routers switch via memory, but…

• Input port processor performs lookup, copy into
memory
• Cisco Catalyst 8500 3

1
 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

• 1 Gbps bus, Cisco 1900:

sufficient speed for access
and enterprise routers (not
regional or backbone)

Switching Via an Interconnection

Network
• Overcome bus and memory
bandwidth limitations

• Crossbar provides full NxN

interconnect
– Expensive
– Uses 2N buses

• Cisco 12000: switches Gbps

through the interconnection
network

Routing
• Past two lectures
– IP addresses are structured
– IP packet headers carry these
addresses
– When packet arrives at router
• Examine header for intended
destination Router
• Look up next hop in table
• Send packet out appropriate
port

• This lecture:
– How these forwarding tables
are built?
– Routing algorithms

2
 A Model of the Problem
• Network as a Graph:
– Represent each router as node
– Direct link between routers
represented by edge
– Symmetric links ⇒ undirected E 3 C
1
graph
1
2 F
• Edge “cost” c(x,y) denotes
6
measure of difficulty of using link
1
– delay, $ cost, or congestion level 3 D
A 4
• Task B
– Determine least cost path from
every node to every other node
• Path cost d(x,y) = sum of link costs

Ways to Compute Shortest Paths

• Centralized
– Collect graph structure in one place
– Use standard graph algorithm
– Disseminate routing tables

• Distributed
– Routers perform local computation
– Converge to a globally consistent routing state
– “Global”: Link-state
• Every node collects complete graph structure
• Each computes shortest paths from it
• Each generates own routing table
– Local: Distance-vector
• No one has copy of graph
• Nodes construct their own tables iteratively
• Each sends information about its table to neighbors
8

Distance-Vector Method
E 3 C
1
Initial Table for
A F 1
2
Dest Cost Nex 6
t 1
Hop 3 D
A 4
A 0 A B
B 4 B
C –
• Idea
∞
– At any time, have cost/next
D ∞ –
hop of best known path to
E 2 E destination
F 6 F – Use cost ∞ when no path known
• Initially
– Only have entries for directly 9
connected nodes

3
 Algorithm
Each node x stores:
– c(x,v) for each neighbor v
– Distance vector of node x: estimate of d(x,y) for all y
– Distance vectors heard from each neighbor

Initialization:
1. d(x,y) = c(x,y) for all y.
2. Send distance vector to each neighbor
Repeat:
Whenever link cost to neighbor changes or distance
vector received from neighbor
For every neighbor z
For every destination y
d(x,y) ← Update(x,y,z)
If d(x,y) changed for any y, send distance vector to all 10
neighbors

Distance-Vector Update
z
d(z,y)
c(x,z)

y
x
d(x,y)

Update(x,y,z)
d ← c(x,z) + d(z,y) /* Cost of path from x to y with first hop z */
if d < d(x,y)
/* Found better path */
return d,z /* Updated cost / next hop */
else
return d(x,y), nexthop(x,y) /* Existing cost / next hop */
11

Start
Optimum 1-hop paths
Table for A Table for B
E 3 C
Dst Cst Hop Dst Cst Hop 1
A 0 A A 4 A 1
2 F
B 4 B B 0 B
6
C ∞ – C ∞ – 1
D ∞ – D 3 D 3 D
A 4
E 2 E E ∞ –
B
F 6 F F 1 F
Table for C Table for D Table for E Table for F
Dst Cst Hop Dst Cst Hop Dst Cst Hop Dst Cst Hop
A ∞ – A ∞ – A 2 A A 6 A
B ∞ – B 3 B B ∞ – B 1 B
C 0 C C 1 C C ∞ – C 1 C
D 1 D D 0 D D ∞ – D ∞ –
E ∞ – E ∞ – E 0 E E 3 E
F 1 F F ∞ – F 3 F F 0 F 12

4
 Iteration #1
Optimum 2-hop paths
Table for A Table for B
E 3 C
Dst Cst Hop Dst Cst Hop 1
A 0 A A 4 A 1
2 F
B 4 B B 0 B
6
C 7 F C 2 F 1
D 7 B D 3 D 3 D
A 4
E 2 E E 4 F
B
F 5 E F 1 F
Table for C Table for D Table for E Table for F
Dst Cst Hop Dst Cst Hop Dst Cst Hop Dst Cst Hop
A 7 F A 7 B A 2 A A 5 B
B 2 F B 3 B B 4 F B 1 B
C 0 C C 1 C C 4 F C 1 C
D 1 D D 0 D D ∞ – D 2 C
E 4 F E ∞ – E 0 E E 3 E
F 1 F F 2 C F 3 F F 0 F 13

Iteration #2
Optimum 3-hop paths
Table for A Table for B
E 3 C
Dst Cst Hop Dst Cst Hop 1
A 0 A A 4 A 1
2 F
B 4 B B 0 B
6
C 6 E C 2 F 1
D 7 B D 3 D 3 D
A 4
E 2 E E 4 F
B
F 5 E F 1 F
Table for C Table for D Table for E Table for F
Dst Cst Hop Dst Cst Hop Dst Cst Hop Dst Cst Hop
A 6 F A 7 B A 2 A A 5 B
B 2 F B 3 B B 4 F B 1 B
C 0 C C 1 C C 4 F C 1 C
D 1 D D 0 D D 5 F D 2 C
E 4 F E 5 C E 0 E E 3 E
F 1 F F 2 C F 3 F F 0 F 14

Distance Vector: Link Cost Changes

Link cost changes:
• Bad news travels slow - 60
Y
“count to infinity” problem! 4 1
X Z
50

algorithm
continues
on!

15

5
 Distance Vector: Poison Reverse
If Z routes through Y to get to X : 60
• Z tells Y its (Z’s) distance to X is infinite (so Y Y
4 1
won’t route to X via Z)
• Will this completely solve count to infinity X Z
50
problem?

algorithm
terminates

16

Poison Reverse Failures

Table for A Table for B Table for D Table for F
Dst Cst Hop Dst Cst Hop Dst Cst Hop Dst Cst Hop
C 7 F C 8 A C 9 B C 1 C
1
∞
Table for A Table for F F C
Forced Forced 6
Dst Cst Hop Dst Cst Hop
Update Update
C ∞ – C ∞ – A
1
Table for A
Better 4
Dst Cst Hop 1
Route B
C 13 D
D
Table for B
Forced
Update Dst Cst Hop • Iterations don’t converge
C 14 A • “Count to infinity”
Table for D
Forced • Solution
Dst Cst Hop
Update – Make “infinity” smaller
C 15 B
Table for A – What is upper bound on
•
Dst Cst Hop Forced
• maximum path length?
Update 17
C 19 D •

Routing Information Protocol (RIP)

• Earliest IP routing protocol (1982 BSD)
– Current standard is version 2 (RFC 1723)

• Features
– Every link has cost 1
Hop count
– “Infinity” = 16
• Limits to networks where everything reachable within 15 hops

• Sending Updates
– Every router listens for updates on UDP port 520
– Triggered
• When every entry changes, send copy of entry to neighbors
– Except for one causing update (split horizon rule)
– Periodic
• Every 30 seconds, router sends copy of its table to each neighbor

18

6
 Link State Protocol Concept
• Every node gets complete copy of graph
– Every node “floods” network with data about its
outgoing links

• Every node computes routes to every other

node
– Using single-source, shortest-path algorithm

• Process performed whenever needed

– When interconnections die / reappear

19

Sending Link States by “Flooding”

• X wants to send X A X A
information
C B D C B D
– Sends on all outgoing
(a) (b)
links
X A X A
• When node Y receives
C B D C B D
information from Z
(c) (d)
– Resend on all links
other than Z

20

Dijkstra’s Algorithm
• Given
– Graph with source node s and edge costs
c(u,v)
– Determine least cost path from s to every
node v

• Single source shortest Path Algorithm

– Traverse graph in order of least cost from
source

21

7
 Dijkstra’s Algorithm
∞
2 E 3 C
5 1
Current Path Costs
F 2
2
6
1
Source 0 ∞
3 3 D
Node A 3
B
Done
Unseen
Horizon
•Node Sets
– Done • Label
• Already have least cost path to it – d(v) = path cost
– Horizon: • From s to v
• Reachable in 1 hop from node in
Done • Path
– Unseen: – Keep track of last link in
• Cannot reach directly from node path 22
in Done

Dijkstra’s Algorithm: Initially

∞
∞ E 3 C
∞ 1
Current Path Costs
F 2
2
6
1
Source 0 ∞
∞ 3 D
Node A 3
Horizon B
Done Unseen

• No nodes “done”
• Source in “horizon”

23

Dijkstra’s Algorithm: Initially

∞
2 E 3 C
6 1
Current Path Costs
F 2
2
6
1
Source 0 ∞
3 3 D
Node A 3
B
Done
Horizon Unseen

• d(v) to node A shown in red

– Only consider links from done nodes

24

8
 Dijkstra’s Algorithm
∞
2 E 3 C
5 1
Current Path Costs
F 2
2
6
1
Source 0 ∞
3 3 D
Node A 3
B
Done
Unseen
Horizon

• Select node v in horizon with minimum d(v)

• Add link used to add node to shortest path tree
• Update d(v) information
25

Dijkstra’s Algorithm
Horizon ∞
2 E 3 C
5 1
Current Path Costs
F 2
2
6
1
Source 0 ∞
3 3 D
Node A 3
B
Done
Unseen

• Repeat…

26

Dijkstra’s Algorithm
∞ Unseen
2 E 3 C
4 1
Current Path Costs
F 2
2
6
1
Source 0 6
3 3 D
Node A 3
B Horizon
Done

• Addition of node can add new nodes to horizon

27

9
 Dijkstra’s Algorithm
5
2 E 3 C
4 1

F 2
2
6
1
Source 0 6
3 3 D
Node A 3
B

• Final tree shown in green

28

Link State Characteristics

• With consistent
LSDBs*, all nodes
B
compute consistent 1 1
X
loop-free paths A 4
C

6 2
• Can still have D

transient loops Packet from C
A

may loop around BDC
if B knows about failure
and C & D do not
*Link State Data Base

29

OSPF Routing Protocol

• Open
– Open standard created by IETF

• More prevalent than RIP

30

10
 OSPF Messages
• Transmit link state advertisements
– Originating router
• Typically, IP address for router
– Link ID
• ID of router at other end of link
– Metric
• Cost of link
– Sequence number
• Incremented each time sending new link information

31

OSPF Flooding Operation

• Node X Receives LSA from Node Y
– With Sequence Number q
– Looks for entry with same origin/link ID

• Cases
– No entry present
• Add entry, propagate to all neighbors other than Y
– Entry present with sequence number p < q
• Update entry, propagate to all neighbors other than Y
– Entry present with sequence number p > q
• Send entry back to Y
• To tell Y that it has out-of-date information
– Entry present with sequence number p = q
• Ignore it
32

Flooding Issues
• When should it be performed
– Periodically
– When status of link changes
• Detected by connected node
• Congestion, lack of electric or optical signal

• What happens when router goes down & back up

– Sequence number reset to 0
• Other routers may have entries with higher sequence numbers
– Router will send out LSAs with number 0
– Will get back LSAs with last valid sequence number p
– Router sets sequence number to p+1 & resends

33

11
 Adoption of OSPF
• RIP viewed as outmoded
– Good when networks small and routers had
limited memory & computational power

• OSPF Advantages
– Fast convergence when configuration
changes
– Full topology map helps

34

Comparison of LS and DV Algorithms

Message complexity Speed of Convergence
• LS: with n nodes, v • LS: Complex computation
neighbors, O(nv) messages
per node – But…can forward before
computation
• DV: exchange between
neighbors only – may have oscillations
• DV: convergence time varies
– may be routing loops
– count-to-infinity problem
– (faster with triggered
updates)

35

Comparison of LS and DV Algorithms

Robustness: what happens if router malfunctions?
LS:
• node can advertise incorrect link cost
• each node computes only its own table
DV:
• DV node can advertise incorrect path cost
• each node’s table used by others
• errors propagate thru network
• Other tradeoffs
• Making LSP flood reliable difficult
• Prioritize routing packets?

36

12