CS558 Sylabus

Link State Routing Algorithm (Centralized)

Preliminaries to the Link State Routing Algorithm

Recall that in the link state type route computation algorithm, every node sends its link state information (information about the delays on its links to neighbor nodes) to all nodes in the network.
Examples of link state routing methods used in the Internet are:
- Open Shortest Path First (OSPF)
- OSI's IS-IS
You need to be aware of this Chicken and Egg dilemma in routing:
- The routing table is used to forward packets to a destination
- We are trying to construct the routing table for a node
- Therefore:
Recall that a node using the link state route construction algorithm needs to send link cost information to every node in the network

Let me make this more explicit:

A node using the link state route construction algorithm must send link cost information to every node in the network WITHOUT using routing tables !!!

(After all, we are constructing the routing tables... If you can use the information in the routing table, why do you want to construct something that you already have ???)

Sending information to ALL node in the network is caled broadcasting

So, a crucial step in implementing link state type route computation algorithm is:
(It's clear that this is a "chicken and egg" problem... You are developing a method to allow nodes to construct the routing tables. So... the routing tables have not been constructed, so you cannot use routing tables to send link state information to other nodes !)

Flooding
- Flooding is the basic algorithm to perform broadcasting without the use of routing tables
- The Flooding algorithm:
  1. A node select a Time To Live (TTL) value that is large enough to allow the packet to reach every node in the network (set to the maximum distance between 2 nodes in network)
  2. A node send the message to ALL its neighbors
  3. When a node send the FLOOD message, it decrements the value in the Time To Live (TTL) field.
  4. If TTL > 0 , the node forwards the FLOOD message to all its neighbors except the neighbor from it the message was received.
    If TTL = 0 , the FLOOD message is the discarded
  Example: Flooding
  - Message contains a TTL field.
  - Routers decrements TTL by 1 each times messaeg is forwarded.
  - An incoming message is forwarded on all links, except on the link where the message was received.
  - Message is not forwarded to next router if TTL becomes 0.
Reliable Flooding
- Reliable flooding:
  - Each message contains a TTL and a version number
  - Nodes maintains a last version number forwarded.
  - Node will only forward a newer version
  - The reason it is called reliable flooding is because the use of the sequence number allow nodes to identify the most recent copy of the information and prevent them from (re)transmitting old information.
- Recall that the information that are being broadcasted through flooding are the link state information.
  For example, node A will broadcast the following message to all nodes:
  Imagine that all nodes doing this. Then, at the end, every node will know the topology of the entire network. In other word, every node in the network will have the following information:
```
  Distance matrix

             To:
    From     A     B     C     D     E
    ------+---------------------------------
     A    |  *     5    10     *     *
     B    |  5     *     3    11     *
     C    | 10     3     *     2     *
     D    |  *    11     2     *     3
     E    |  *     *     *     3     *
```
  The matrix contains all the link state data from all nodes and the matrix is in fact another representation of the graph.
- Every node in the network will have the same information and each node will next run the well-known Dijkstra Algorithm on the graph to find the shortest paths.