Module 7, SMD123

1

• Module 7, SMD123
• Routing Part 1

2
Outline

• Quick flashback
• What is routing?
• Graph theory
• Routing algorithms
• Hierarchical routing

3
The Fellowship of the Routers
router
workstation
server
mobile
4
Forwarding vs. Routing

• Forwarding Looking at address and sending packet
  to correct place (either next-hop or direct)
• Uses forwarding tables
• Routing The process by which forwarding tables
  are built

5
What is Routing? - a human(e) network example

• Demonstrate 3 things
• Routers do not need to have global knowledge!
• Networks can repair themselves!
• Incorrect information can lead to serious errors!

6
A Human(e) Network cont.
(Rules for Routers)
Forwarding Tables
C
A
H2
S
B
H1
Routers
Point-to-Point links
7
A Human(e) Network Handling Failures
Forwarding Tables
C
A
H2
S
A
B
B
1
Right
1
Right
2
Left
2
-
Right
Left
H1
Routers
Point-to-Point links
8
A Human(e) Network Loops
D
Forwarding Tables
C
A
H2
S
A
B
B
1
Right
1
Right
Left
Right
2
Left
2
-
H1
Routers
Point-to-Point links
9
Types of routing

• Routing
• Can be Static or Dynamic
• Can be Distributed or Centralized
• Can be Global

Routing
Dynamic
Static/Manual
Distributed
Centralized
True Distrib.
Global
10
Outline

• Quick flashback
• What is routing?
• Graph theory
• Routing algorithms
• Hierarchical routing

11
Routing Foundations

• Graph theory used to build routing algorithms
• Allows us to prove properties about algorithms
• A little graph theory will last us a long way
• Nodes
• Links
• Link costs
• Minimum cost distance

12
A Little Something About Graphs

• Graphs are like maps only without geometry!
• Cares only about connections
• Consist of
• Nodes
• Links, which may have costs
• Can be
• Connected, contain loops
• Special Case A Tree
• A connected graph that does not contain loops!

13
Transforming a Map to a Graph
Topology Description of Graph
2
1
2
2
1.5
14
Making the point
Same Topology!
15
Routing Graph Theory

• Switches Hosts are Nodes (A,B,C,)
• Links between nodes are Edges (no names)
• Each edge has a Cost associated with it.
• C(B,C) 2
• C(C,B) 2
• Main Problem - Calculating the path with the
  lowest cost, from A to F. ? D(A,F) 8

1
1
2
3
9
6
7
4
16
Shortest Path Tree
1
2
1
1
C
2
3
9
4
6
7
4
10
6
17
Outline

• Quick flashback
• What is routing?
• Graph theory
• Routing algorithms
• Hierarchical routing

18
What Are We Looking For?

• Speed to convergence
• Memory Usage
• Amount of routing messages
• Scalability

19
Distance Vector Algorithm

• Outline of the Algorithm
• Based on Bellman-Ford algorithm
• An example of the Algorithm
• Problems Solutions

20
Outline of the Algorithm

• The Algorithm Basics
• Each node knows the cost of the link of each its
  directly connected neighbors initial
  forwarding table.
• Periodically each node transmits their table to
  all their neighbors.
• When a node receives such a table it calculates
  the distance to the nodes on the received table.
• If the node discovers a shorter path to a
  destination node, its own table is updated. If
  the table was updated the new table is
  immediately sent to the neighbors, otherwise the
  node waits for a timeout or a new message.

21
From the View of a Node

• A Nodes Routing Table consists of triples in the
  form ltDestination,Cost,Nexthopgt
• The Messages sent are lists containing tuples in
  the form ltDestination, Costgt

Routing Table for B
22
A Distance Vector Example
(A,1), (C,1)
A
Node
cost
Nexthop
B
1
B
C
1
C
(A,1), (B,1),(D,1)
E
1
E
F
1
F
D
2
C
G
2
F
(A,1), (G,1)
B
C
F
Node
cost
Nexthop
Node
cost
Nexthop
Node
cost
Nexthop
A
1
A
B
1
B
A
1
A
(A,1), (C,1)
(A,1), (B,1),(D,1)
(A,1), (G,1)
C
1
C
A
1
A
G
1
G
D
1
D
23
Final Information at Nodes
24
Book Stylie
1
7
8
2
1
2
D
c(E,D) min D (C,w)

w

22 4
D
c(E,D) min D (A,w)

w

23 5
loop!
B
c(E,B) min D (A,w)

w

86 14
loop!
25
Distance table gives routing table
Outgoing link to use, cost
A B C D
A,1 D,5 D,4 D,4
destination
Routing table
Distance table
26
Problems Solutions

• Node/link failures may give rise to unstable
  conditions - Count to Infinity
• Long convergence times on large networks
• Limit the size of Network
• Distance vector tables always transmitted (even
  if contents have not changed)
• Two partial solutions are
• Split Horizon
• Poison Reverse

27
A Count-to-Infinity Example
(E, 4),
(E, 3),
A
(E, 2),
(E, 4),
(E, 3),
Node
cost
Nexthop
B
1
B
(E, ?),
(E, 2),
(E, 3),
(E, 4),
C
1
C
-
?
4
5
B
E
1
E
B
(E,?),
F
1
F
D
2
C
G
2
F
B
C
Node
cost
Nexthop
Node
cost
Nexthop
-
-
?
?
C
B
C
B
E
2
A
E
2
A
C
B
4
4
5
5
3
3
28
Split Horizon
(E, 3),
A
(E, 2),
(E, 3),
Node
cost
Nexthop
B
1
B
(E, ?),
(E, 2),
(E, 3),
C
1
C
-
?
B
E
1
E
B
(E,?),
F
1
F
D
2
C
G
2
F
B
C
Node
cost
Nexthop
Node
cost
Nexthop
-
-
?
E
2
A
E
2
A
C
B
3
3
29
Count-to-infinity

• Only real solution is to set a low infinity
• E.g., 16
• Split Horizon Poisoned Reverse only breaks
  loops with 2 links

30
Link-State Algorithm

• Outline of the Algorithm
• (Dijkstras Shortest Path Algorithm)
• Reliable Flooding
• Route Calculation - Forward Search Alg.
• Properties

31
Outline of the Algorithm (1)

• The Algorithm Basics
• Assumption Each node knows the cost of the link
  of each its directly connected neighbors. Same as
  for the Dist. Vector Alg.
• Basic Idea Every node knows how to reach its
  neighbors. If this information is spread to every
  node, then every node will have enough
  information to determine the correct path to any
  node in the network. In fact every node will be
  able to build a complete map of the network. This
  means that every node will eventually have access
  to the same information, as every other node.

32
Outline of the Algorithm (2)

• Thus Link-State protocols rely on 2 things
• Reliable spreading of link-state information
  (Reliable Flooding)
• Calculation of paths from the sum of all the
  accumulated link-state knowledge

33
Purpose of Reliable Flooding
34
Route Calculation (1)

• Question How is shortest path calculated from
  complete map?
• Is most often performed by a realization of
  Dijkstras algorithm, called Forward Search or
  Shortest Path First.
• It contains two lists a Tentative a Confirmed
  one
• Each list contains entries in the form of
  ltDestination, Cost, NextHopgt

35
Route Calculation (2)

• The Basic Idea
• Initialize Confirmed list with myself
• Call the most recently added node next
• For each neighbor of next, calculate the cost
  to reach the neighbor from me via next
• a) If neighbor is on neither of the lists, then
  add ltneighbor, Cost, Nexthopgt to the Tentative
  list.
• b) If neighbor is on the Tentative list but the
  cost using the new path is less. Then replace the
  current entry.
• Pick the entry with the smallest cost and add it
  to the Confirmed list and return to step 2).

36
A Forward Search Example
D
Step
Confirmed
Tentative
1
(D,0)
(B,11) (C,2)
E
2
(D,0) (C,2)
(B,5) (A,12) (E,11)
9
B
5
3
3
(A,10) (E,11)
(D,0) (C,2) (B,5)
A
C
10
4
(D,0) (C,2) (B,5) (A,10)
(E,11)
11
2
D
5
--
(D,0) (C,2) (B,5) (A,10) (E,11)
37
Properties

• Self-stabilizes quickly
• Does not generate much traffic
• Responds rapidly to topology changes
• - Large amount of information (an entire LSP)
  needs to be stored at every node.
• A drawback in scalability

38
Distance Vector vs Link-State

• The difference between the two algorithms can be
  summarized
• D.V. Each node only talks to its directly
  connected neighbors, but sends its entire
  forwarding table.
• L-S Each node talks to all other nodes, but only
  tells them what it knows for sure (i.e. only the
  state of its directly connected links)

39
Outline

• Quick flashback
• What is routing?
• Graph theory
• Routing algorithms
• Hierarchical routing

40
Hierarchical Routing

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

• scale with 200 million destinations
• cant store all dests in routing tables!
• routing table exchange would swamp links!

• administrative autonomy
• internet network of networks
• each network admin may want to control routing in
  its own network

41
Hierarchical Routing

• aggregate routers into regions, autonomous
  systems (AS)
• routers in same AS run same routing protocol
• intra-AS routing protocol
• Interior Gateway Protocol - IGP
• 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
• also responsible for routing to destinations
  outside AS
• run inter-AS routing protocol with other gateway
  routers
• Exterior Gateway Protocol - EGP

42
Intra-AS and Inter-AS routing
Host h2
Intra-AS routing within AS B
Intra-AS routing within AS A

• Well examine specific inter-AS and intra-AS
  Internet routing protocols shortly