RECAP – Distance Vector Routing Distance Vectors shared Initialization – directly connected routers Sharing – Share the first two columns of the routing table Updating – Better cost replaces old entry When to share? – Periodic and triggered Problems Eg: RIP
LINK STATE ROUTING Goal : To find the shortest path to the destination. Difference from DV : In DV each router knows only the next hop. But in LS each router has the entire topology of the domain.
Steps: Initial State : All routers know their own link state. Creation of LSPs Reliable Flooding of LSPs Formation of Shortest Path Tree Routing Table Calculation
Creating LSP Link State Packets (LSP) carry large amount of information. Also called Hello Packets. Important Fields : Router ID, Link state of the router, Sequence number, Time to Live. Created on two occasions: When there is change in the topology. Periodic basis.
Reliable Flooding ALL nodes broadcast their LSPs to their adjacent nodes. Depending upon the sequence number and the link state values, the receiving node decides whether to discard or store the LSP.
Formation of Shortest Path tree After the LSP are shared among all nodes, each and every node has the entire topology of the domain. Then Dijkstra’s algorithm is run over this information to create a shortest path tree from the graph obtained. Permanent List and Tentative list – An eg. Problem Demo - http://www.ifors.ms.unimelb.edu.au/tutorial/dijkstra /island.html
Creation of routing table From the generated shortest paths, the routing table is generated. In both DV and LS, the final routing table is the same. The difference lies only in the way the routing table is prepared.
OSPF – Open Shortest Path First The Autonomous System is divided into Areas. LS routing protocol runs inside an Area. The summary of the routing information is given to the Backbone area through ABR.