Hello Vishal,
but each router has the routing table which receives routing table from its neighbor. then how can we say that it doesnt have exact topology?
Because the word "topology" means much more than a plain listing of all existing destinations in a network and the nearest neighbors through which they can be reached. For example, using the routing table, you can not even know how many routers you have in your network, and you do not know how exactly they are interconnected. Knowing the topology means that you have a precise picture of the network - how many routers are there, how many stub networks, how many transit networks, how many interfaces each router has and what exactly are those interfaces connected into. This is knowing topology, and each single router that uses a link-state protocol (OSPF or IS-IS) knows this exact information about the entire network (or better said, its own area but let's keep things simple for now). In fact, if you understand the format of the link-state database, you can sit at a single router, let it display its link-state database, and you will be able to draw the entire exact diagram of your network simply by looking at its link-state database.
In distance-vector protocols, this is not possible. In information sciences, the word vector stands for array. Distance vector means an array of distances. This is what the distance vector routing protocols exchange in their update messages - arrays of distances, formatted in the ways network1:distance1, network2:distance2, ..., networkN:distanceN, where each router tells each other its own distances to all currently known networks. This tells you something about the reachable networks in your infrastructure but a single router will never know the entire network topology - and that is the point.
Best regards,
Peter
