
2/29

Network Information Source and Update

Timing

routing decisions usually based on knowledge ofnetwork (not always)

distributed routing

using local knowledge, info from adjacent nodes, info from all nodes on

a potential route central routing

collect info from all nodes

Issue of update timing when is network info held by nodes updated ?

fixed - never updated (Static)

adaptive - regular updates (Dynamic)

2OPM


3/29

General routing techniques

Shortest path Routing

Distance Vector Routing

Link State RoutingHierarchical Routing

Flooding

Broadcast routing

Multicast Routing

Routing for mobile Hosts

Routing in Ad Hoc Network

3OPM


4/29

Routing algorithm

It is part of the network layer software responsible for deciding

through which output linean incoming packet should be

transmitted.

4OPM


5/29

Routing algorithms

Adaptive (Dynamic):- Change their routing decisions to

reflect changes in the network traffic and topology

Non-adaptive (Static):- Do not change their routing decisionson measurements or estimates of the current traffic and

network topology.

5OPM


6/29

Least cost algorithm

All packet-switched and internet base networks take their

routing decision on some form of least-cost criterion.

Criterion may be minimum number of hops, minimum linkcost, time delay, and length of transmission line between

source and destination etc.

6OPM


7/29

Least cost algorithm

Link cost/value

Inversely proportional to the link capacity

Proportional to the current load on the link

Or some combination of above two

Link cost may be asymmetric in both direction

7OPM


8/29

Least cost algorithms

For packet switching networks and internet, mostly used least

cost routing algorithms:

Dijkstra algorithm

Bellman-Ford algorithm

8OPM


9/29

Example Packet Switched Network

9OPM


10/29

Dijkstras Algorithm Example

10OPM


11/29

DijkstrasAlgorithm

1. [Initialization]

T= {s}; where T and s are set of traveled nodes and source node

L(n) = w(s, n) for n is not equal to s; where L stores link cost from node s to

n

2. [Get Next Node] Choosing the next node which is having minimum link cost

Find x which is not a element of T such that L(x) =min L(j); for j is not a

element of T; where x (next node) would be the least cost node from s.

3. [Update Least-cost paths]

L(n)= min[L(n), L(x) + w(x, n)]; for all n not element of T

11OPM


12/29

Bellman-Ford Algorithm

This algorithm finds the shortest paths from a given source

node subject to the constraints that the paths contains at most

one link

Then find the shortest paths with a constraints of paths of at

most two links, and so on.

12OPM


13/29

Example of Bellman-Ford algorithm

13OPM


14/29

Bellman-Ford Algorithm

1. [initialization]

L0(n)= infinity, all values of n except s (source node)

Lh(s) = 0, for all h (hops)

2. [Update]

For each successive h>=0:

Lh+1(n) = min [Lh(j) + w(j, n)];

Where j represent all the nodes one hop less distant than n from s.

Where

h = hop count

s= source node

n = node (Router) in the network

Lh(n)= link cost of node n which is at h hop distance from s

j = node one hop less distance then node n

w(j, n)= link cost from node j to n

14OPM


15/29

Comparison

In Bellman-Ford algorithm, the calculation for node n

involves knowledge of the link cost of all neighboring nodes

plus the total link cost to each neighboring nodes from a

particular node /Source Node

In Dijkstra algorithm, each node must have complete

topological information about the network

15OPM


16/29

Flooding

Static algorithm Every incoming packet is sent out on every out going line except

the one it arrived on.

It is used to find No. of hopsbetween source and destination

Consumes too much bandwidth

To reduce duplicate packets in the network Packet Sequence No.

can be used to discard the duplicate packets arrived at a node.

Selective flooding reduces bandwidth wastage by allowing

flooding in the direction/ path which leads to destination device.

16OPM


17/29

Distance vector/Bellman Ford Routing

Dynamic routing

Each router maintains a routing table indexed by, and containingone entry for each router in the subnet.

The entry contains link cost/hop count and out going path

It is good for the network having uniform data rate links but notfor non-uniform data rate links.

It encounters Infinite count problem (good news spreads quicklyand bad news spreads slowly)

17OPM


18/29

Link State Routing

Dynamic routing

Distance routing algorithm was replaced by Link state routing

Removes the count to infinity problem andconsiders non-

uniform data rate links

18OPM


19/29

Link State Routing

Steps of Link state routing

Discover neighbors and their Addresses

Measure the cost to each of its neighbors

Construct a packet telling all it has just learned Send this packet to all other nodes

Compute the shortest path to every other router (Dijkstra

algorithm can be used)

19OPM


20/29

Hierarchical routing

As network grows in size, routing tables grow, which results inmore memory consumption, CPU processing and bandwidth

are needed to send status report.

Routers are divided into Regions

Each router contains information about its region only

For big network, two or more level hierarchy may be possible

as Regions=> Clusters=> Zones=> Groups and so on

20OPM


21/29

Broadcast Routing

Hosts send message to all other nodes

This method is wasteful of bandwidth

Consumes too much bandwidth

21OPM


22/29

Multicast routing

Sending a message to a group in network is called multicasting,

and its routing algorithm is called multicast routing

Multicasting requires group management

Either host must inform their routers about changes in group

membership, or router must query their host periodically

22OPM


23/29

Routing for Mobile hosts

Hosts are mobile and routers are fixed

The portable devices supports mobility which increases

complexity in routing.

Type of Hosts

Stationary Host

Migratory Host

Mobile Host

23OPM


24/29


A big communication network is divided into small units such

as area (LAN, WLAN,MAN and PAN).

Foreign agents:- Keeps track of all mobile hosts visiting the

area.

Home agent:- Keeps track of hosts Whose home is in same

area

24OPM


25/29


Visiting host registration process

Periodically, each foreign agent broadcasts a packet announcing its

existence, vice versa also possible.

Mobile host registers with the foreign agent, giving its home address

& security information.

The foreign agent contacts the mobile hosts home agent.

The home agent examines the security information and permits

foreign agent to proceed.

The foreign agent gets the acknowledgement from the home agent, it

makes an entry and informs the mobile host that it is now registered

25OPM


26/29

Routing for Mobile hosts in GSM

Examples:

Host moves one cell to another cell

Host move one MSC to another MSC (Roaming)

International host

26OPM


27/29


Called MANETS (Mobile AdHoc NETworks)

Host and Router both are mobile

Each node works as independent entity . These nodes are self

organized and self governed.

No need of base station or facilitator

Each node works as a host and router

27OPM


28/29


Examples:

Military vehicles in battle fields with no existing/fixed

infrastructure

Groups of ships in sea

Wireless Network by the Laptops without WiFi Access Point

Bluetooth

28OPM


29/29


Topology may be changing all the time, due to which routing

in Adhoc network is different

AODV (AdHoc On-demand Distance Vector) routing

algorithm is a distant relative of Bellman-Ford algorithm

This Algorithm is adapted to work in a mobile environment

and takes into account the limited bandwidth and low battery

life

It is a on-demand algorithm

29OPM

6 network layer routing

Documents