OSPF route selection in detail

Hello there.

First of all, i should say that i'm just preparing to ROUTE exam, and could find any other place for posting my question, so if i made a mistake with a forum branch, administrator, make things right and place it where it should be.

Well, the question is about OSPF route selection. We have got some bunch of routers, they formed neighborships, elected DR/BDR (let's imagine that it's a 'broadcast' type of network) and exchanged the topology information. So they all have similar LSDBs and they are agree on whole inter-area topology. Then each router should calculate all paths to all destinations and place paths with the lowest cost into it's routing table. In Odom's ROUTE of.cert.guide it's desribed this way:

"1. Analyze the LSDB to find all possible routes to reach the subnet.

2. For each possible route, add the OSPF interface cost for all outgoing interfaces in

that route.

3. Pick the route with the lowest total cost."

 

I just wondered how this process occurs in detail and i found only one book where it's described: Routing TCP/IP by Doyle+Carroll. This process of OSPF route selection after exchanging LSDBs is described as building SPT with a router that builds it placed as root of this SPT. And the process of building the tree is very complex, using some candidate-tables and so on, but the main principle is "to follow the lowest cost interfaces". So, here is the question: following this principle we won't calculate all the paths and then select lowest cost ones (as Odom ensures), we just follow lowest-cost interfaces without taking into consideration all others. I just get lost it this process of calculation each route's cost... Can anybody tell a book or explain in detail how it is implemented?

 

Thanks a lot. I'm really sorry for the brutal explanation, english is not my native language, i'm trying to do my best :)

Comments

Sign In or Register to comment.