Link State Question

Hey everyone:

I had a question on the theory behind Link State Routing Protocols. I have been reading up on Distance Vector vs. Link State and have come to understand that when using a Link State protocol, such as OSPF, each router essentially creates LSAs which are flooded out to every other router. Each other router uses these LSAs to create its LSDB.

What I gather so far is that the collection of all these flooded LSAs allow each individual router to basically create and have a view of the entire topology of the network (which is stored as the LSDB). Is this correct? Does the construction of the LSDB give the router a "map" of the entire network?

Thanks

Find more posts tagged with

Free for TechExams community: Cybersecurity salary guide

Compare cert salaries and plan your next career move

Button

Comments

submit_purpose

That's exactly how it works for link state routing. Every node maintains a consistent view of the entire network and build shortest paths by assuming it as the root node of the constructed Dijkstra tree.

The subtle difference when it comes to link state routing protocol OSPF is the way the updates are propagated. Every node sends their LSA's only to the DR and BDR (multicast addr: 224.0.0.6), and from their they get relayed by the DR to all other nodes in the network (multicast addr: 224.0.0.5)

zerglings

submit_purpose wrote: »

The subtle difference when it comes to link state routing protocol OSPF is the way the updates are propagated. Every node sends their LSA's only to the DR and BDR (multicast addr: 224.0.0.6), and from their they get relayed by the DR to all other nodes in the network (multicast addr: 224.0.0.5)

It only happens when they're in a multi-access network, such as Ethernet.

typesh

Thanks!! I wanted to confirm I understood it correctly. At first it was a little strange to grasp the concept that each Router maintains a full topology-map of the entire network. It makes sense now. It's actually really cool how each Router considers itself the root node/starting point and runs the SPF Algorithm on the tree it constructed in order to calculate the best path. Thanks again.

For anyone else interested, here is a link I found that helps visualize it:

Distance vector versus link-state - Packet Life

Neiby

Keep in mind that OSPF is essentially a link-state protocol within an area, but distance vector between areas. An OSPF router only has a map for its own area.

typesh

submit_purpose wrote: »

That's exactly how it works for link state routing. Every node maintains a consistent view of the entire network and build shortest paths by assuming it as the root node of the constructed Dijkstra tree.

Does EIGRP allow routers to have the same full-topology view? I know the Dijkstra Algorithm doesn't apply here, but there is a Topology Exchange once the Neighbor relationship is formed.

Each EIGRP router receives the subnet number, subnet mask, Minimum Bandwidth, and Delay from it's sending neighbor. Based on this information, is the receiving router able to obtain a full map of the network? Or does it just know that SubnetA is in a certain direction and has a cost of x.

It doesn't appear that it is able to obtain a full topology view, but I could be wrong. It just seems that the receiving router just knows that a "certain subnet" can be reached through it's neighbor, where this neighbor provides some numbers (least-bandwidth and delay) for the receiving router to do some calculations on. The receiving router does not actually know through which subnets it must pass to get to it's destination, whereas in OSPF it would. I know that EIGRP is neither DV or Link State... and it appears that the DV properties are knowing that a certain subnet is in a given direction... and the Link State properties are knowing how to perform cost calculations for that direction. Correct?

Neiby

EIGRP routers are not aware of the entire topology like a router in OSPF would. EIGRP routers learn the best routes that their neighbors have available. They do not know of all possible routes like you would see with a link-state protocol.

typesh

Neiby wrote: »

EIGRP routers are not aware of the entire topology like a router in OSPF would. EIGRP routers learn the best routes that their neighbors have available. They do not know of all possible routes like you would see with a link-state protocol.

That's what it seemed like. Thanks for the reply!

Neiby

To clarify, I was only referring to OSPF routers that share an area. OSPF routers within an area have a complete picture of that area, but they do not have a complete picture of other areas. OSPF is link-state within an area, but distance vector between areas.