Intradomain Routing EECS 122: Lecture 10

1
Intradomain RoutingEECS 122 Lecture 10

• Department of Electrical Engineering and Computer
  Sciences
• University of California
• Berkeley

2
What is Routing?

• Routing is the core function of a network
• It ensures that
• information accepted for transfer
• at a source node
• is delivered to the correct
• set of destination nodes,
• at reasonable levels of performance.

3
Datagram v/s Virtual Circuit

• Datagram routing
• Each packet to be forwarded independently
• Virtual Circuit
• Each packet from same o-d uses same route
• More state (pick the right granularity)
• QoS sensitive networks use VCs and signaling
• Find a route that has the resources available for
  the connection.
• Reserve the resources before sending data
  packets

4
Many Kinds of Routing

• Driven by
• Destination set
• IP point-to-point
• Multicast
• Physical Characteristics
• Optical
• mobile wireless
• Diffusion (sensor networks)
• Interconnection network routing
• Geographic (wireless)
• Network Function
• P2P
• Content Distribution Networks

5
A Graph Model

• Nodes are Routers/Hosts
• Edges are undirected

6
Walks
A Walk from 1 to 11
Cycle 4-8-7-5-4
7
5
4
8
6
11
2
10
3
1
13
12
7
Paths
Routes are Paths
A Path is a Walk with no cycles
7
5
4
There are 24 paths from 1 to 11
8
6
11
2
10
3
1
13
12
8
Routes
Routes are Paths
A Path is a Walk with no cycles
Edges have weights
There are 24 paths from 1 to 11
9
Routing
Routes are Paths
A Path is a Walk with no cycles
7
5
Edges have weights
4
There are 24 paths from 1 to 11
8
Select the shortest path route Many ways to do
this
6
11
2
10
3
1
13
12
10
Routing affects Flow Control

• Any function that paces the flow of bits into
  or within the network is flow control
• Example All links have capacity 1

1
1
A
B
Nothing admitted
E
E
C
F
F
D
11
What is going on?

• Routing and flow intimately related
• Link congestion metrics for routing depends on
  flow control
• Flow control feedback delay depends on routing
• Optimizing them jointly is nice in theory but
  intractable in practice.
• Separating flow control and routing makes both
  extremely difficult to implement with high
  performance

12
The internet has many Administrative Domains
B
7
5
4
8
6
11
2
10
3
1
13
12
C
A
13
Border Routers
B
7
5
4
RIP
4
8
6
6
11
2
2
10
OSPF
BGP
3
13
3
13
1
12
C
IGRP
A
14
Hierarchical Routing
4
4
B
6
BGP
6
B
7
5
IntraDomain
2
InterDomain
2
4
RIP
8
6
3
13
3
13
11
2
10
OSPF
IntraDomain
IntraDomain
3
1
13
12
IGRP
C
A
15
Routing Sub-Functions

• Topology Update Characterize and maintain
  connectivity
• Discover neighbors
• Measure distance (one or more metric)
• Disseminate
• Route Computation
• Kind of path Multicast, Unicast
• Centralized or Distributed Algorithm
• Policy
• Hierarchy
• Switching Forward the packets at each node

16
Routing Sub-Functions

• Topology Update Characterize and maintain
  connectivity
• Discover neighbors
• Measure distance (one or more metric)
• Disseminate
• Route Computation
• Kind of path Multicast, Unicast
• Centralized or Distributed Algorithm
• Policy
• Hierarchy
• Switching Forward the packets at each node

17
Three types of Routing Protocols

• Topology changes can be detected by nearby nodes
• These changes must be reflected in the routes
• Mechanisms for disseminating information
• Link State Communicate the names and costs of
  neighbors. Each node maintains the entire
  topology. E.g. used in OSPF
• Distance Vector Communicate current distance
  estimates of node to every other node. E.g. used
  in RIP
• Path Vector Communicate current estimates of
  preferred paths from node to every other node.
  E.g. used in BGP

18
Link State Protocols

• Every node learns the topology of the network
• Flooding of Link State Packets (LSP)
• An efficient shortest path algorithm computes
  routes to every other node
• Node updates Forwarding Table

19
Flooding Link State Information

• Every router sends Link State Packets (LSPs) to
  all of its neighbors
• LSPs arrive and wait in buffers to be accepted
• If node j receives a LSP from node k it compares
  the sequence numbers. If this is the most recent
  one from k, send to N(j)-k.
• This way each router can send its LSP to all
  other routers
• Age starts out at 7. At any router, value is
  decremented every 8 seconds. At 0 discard.
• As long as sequence dont wrap this works
• Otherwise things can get ugly

20
Example Network
7
5
4
8
6
11
2
10
3
1
13
12
21
Example Network
7
5
4
8
6
11
2
10
3
1
13
12
22
Example Network
7
5
4
8
6
11
2
10
3
1
13
12
23
Example Network
7
5
4
8
6
11
2
10
3
1
13
12
24
Real story

• Sequence numbers from some router, s, wrapped
  around
• A lt B lt C lt A
• Router, t, has a buffer with LSPs from s of all
  three values in order A, B, C
• Store and flood A
• Replace A with B and flood B
• Replace B with C and flood C
• Router u receives the LSPs in order ABC and goes
  through the same cycle and sends to v
• The entire Arpanet was sending these LSPs and
  crashed
• LSPs did not wait in buffers long enough to age

25
Some Issues

• What happens if some routers are much faster at
  transmitting LSPs?
• What happens if sequence numbers wrap?
• What happens when a partitioned network is
  reconstituted?
• What about security?
• Etc., etc.
• Many lines of code

26
Route Computation Dijkstra

• Every node knows the graph
• All link weights are gt 0
• Goal at node 1 Find the shortest paths from 1 to
  all the other nodes.
• Each node computes the same shortest paths so
  they all agree on the routes
• Strategy at node 1 Find the shortest paths in
  order of increasing path length

3
2
3
1
2
1
4
1
1
4
4
6
5
1
1
27
Dijkstra Shortest Path
Notation c(i,j) gt0 cost of link from (I,j)
D(1,i) Shortest path from 1 to i. D(1,x,i)
Shortest path from 1 to i via x Let P(k) be the
set of nodes k-closest to 1

• IDEA Given P(k) we can find P(k1) efficiently
• To get P(k1), observe that
• This node cannot be in P(k)
• It must be one hop away from some node in P(k)
• Suppose 2 were false. We picked i
• Node i has no edge into P(k)
• There must be a node x, not in P(k) such that
• x is one hop away from P(k) and
• D(1,i)D(1,x)D(x,i)
• But then, D(1,x) lt D(1,i) and we would have
  picked x instead.
• Pick node(s) that is one hop away from P(k) that
  is closest to 1.
• Keep iterating until all nodes are in P

P(2)1,2
3
D(1,5)2 D(1,6,5)5
2
3
1
2
1
4
1
1
4
4
6
5
1
1
28
Dijkstra Shortest Path
29
Dijkstra Forwarding Table
At node 5
Outgoing Cost
1 2 2
2 2 1
3 3 1
4 3 3
6 6 1
3
2
3
1
2
1
4
1
1
4
4
6
5
1
1
30
Distance Vector Protocols

• Communicate current distance estimates of node to
  every other node
• This is called its distance vector Di
  (D(i,1),D(i,2),,D(i,n))
• Initially, assume that all distance estimates are
  c(i,j)
• The nodes do not need to learn the entire
  topology
• Just the distance estimates (vectors) of their
  neighbors
• Periodically each node sends its distance vector
  to all of its neighbors

3
2
3
1
2
1
4
1
1
4
4
6
5
1
1
At node 1 D1 (0,1,8,8, 8,4) Neighbor
estimates 2 (1,0,8,8,8,8) 6 (4,8,8,8,8,0)
At node 6 D6 (4,8,8,8,1,0) Neighbor
estimates 1 (0,8,8,8,8,4) 5 (8,8,8,8,0,1)
31
Distance Vector Protocols

• Upon receiving a more recent distance vector from
  its neighbors, a node, i, stores it and revises
  Di
• New D(i,d)
• The total cost to send it via neighbor j is the
  sum of
• The link cost c(i,j)
• The stored estimate to reach d from j
• Pick the lowest sum over all the neighbors
• D(i,d) minjeN(i) c(i,j) D(j,d)

3
2
3
1
2
1
4
1
1
4
4
6
5
1
1
Node 6 receives (0,1,8, 8,8,8,4) from
1 (8,1,1,4,0,1) from 5
Old Info DV (4,8,8,8,1,0) Neighbor
estimates 1 (0, 8,8,8,8,4) 5 (8,8,8,8,0,1)
Revised DV (4,2,2,5,1,0) Neighbor estimates 1
(0,1, 8,8, 8,4) 5 (8,1,1,4,0,1)
Send all packets to 2 via 1.
32
Distance Vector Protocols

• Forwarding Table at 6

3
2
3
1
2
Estimates DV (3,2,2,4,1,0) Neighbor
estimates 1 (0,1,3,5,2,3) 5 (2,1,1,3,0,1)
1
4
1
1
4
4
6
5
1
1
Node Cost
1 5 3
2 5 2
3 5 2
4 5 4
5 5 1
33
Why does this compute shortest paths?

• Suppose in every tick each node sends its
  distance vector.
• Assume that initial distances are 8
• At time h, node i has as an estimate of the
  shortest path to node j that has lt h1 hops!
• Dh1(i,j) minkeN(i) Dh(k) c(i,k)

1
1
1
1
3
3
4
2
2
2
2
6
3
3
3
1
1
1
4
1
4
5
6
5
6
6
6
5
5
4
2
3
2
3
2
4
34
Counting to Infinity
All links cost 1
A
B
C
3
4
0
A
B
C
5
6
0
Ping-Pong to Eternity
35
Bad News Travels Slowly
1
4
3
1
1
1
2
M
1
D(2,1)2, D(3,1)1, D(4,1)2
36
Bad News Travels Slowly
1
4
3
1
Node 2 takes about M Iterations to figure out
that D(2,1)M
1
1
2
M
1

• Fundamental Cause After a network change, think
  of the network
• protocol running from time 0. The initial
  conditions are arbitrary
• Tricks exist to get around these problems but not
  fool proof

37
Asynchronous Bellman Ford

• In general, nodes are using different and
  possibly inconsistent estimates
• If no link changes after some time t, the
  algorithm will eventually converge to the
  shortest path
• No synchronization required at all

38
Oscillations

• Link costs must reflect link speed AND congestion
• Under both LSP and DV routing occurs over a tree
• The costs of the links of this tree will increase
• The other links will not be congested
• Their costs will drop
• Routing protocol will shift traffic and create a
  new tree
• This process of shifting and reshifting can be
  severe
• Way out Change congestion costs slowly
  (exponential averaging) Route dampening

39
Link State vs. Distance VectorNo clear winner

• LS is robust since it each node computes its own
  routes independently
• Suffers from the weaknesses of the topology
  update protocol. Inconsistency etc.
• Excellent choice for a well engineered network
  within one administrative domain
• E. g. OSPF
• DV works well when the network is large since it
  requires no synchronization and has a trivial
  topology update algorithm
• Suffers from convergence delays
• Very simple to implement at each node
• Excellent choice for large networks
• E.g. RIP