ospf link state

Unanswered Question
Apr 19th, 2010

OSPF cost is associated with the output side of each router  interface not the input so if i change bandwidth on a router, the cost will change on the local router interface and that will trigger spf and only affect the local router routing table not other routers with the area even though they will receive the type 1 link state update.

Any comment?


I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Tharak Abraham Mon, 04/19/2010 - 21:46

Why not ?

Whatever routers downstream would have their costs updated if it uses the link with the new bandwidth to learn routes.

This is applicable to intra as well as inter area routes.

Giuseppe Larosa Mon, 04/19/2010 - 22:26

Hello Francisco,

a new router LSA will be flooded within the area to which the modified link belongs to.

This can trigger an SPF recalculation on all routers within the area if the link is used for forwarding traffic (if there is an OSPF neighbor on it)

IF it is only a leaf (no OSPF neighbors on the modified link),  with recent IOS code that supports partial recalculation a simplified SPF recalculation is performed on the other nodes, that can be as simple as just accepting the link cost change without SPF changes.

Each node creates an SPF rooted at itself using  all the topology information provided by routers in the same area.

For each IP prefix that is present in the area, the local node has to calculate the best path.

In other words the path cost will change reflecting the change in the flooded router LSA.

So at least a partial recalculation is performed on other nodes in the same area

Hope to help


francisco_1 Tue, 04/20/2010 - 04:03


The reason for my posting is related to a small test i done in a lab (see attached), r1 has more than one path to get to r4 loopbacks, since r1 exit path (via r2/r3) are costed the same, loadbalacing is happening fine (i could see cef loadbalancing between exit interfaces on r1) . Now when i changed the bw and delay on the interface on router r2 towards r4, that affected the cost on r2 interface towards r4 and that triggered spf on r2. r1 saw the link state advertisement from r2 without making any spf change, r1 still using the path via r2 and cef still performing per-destination load balancing on r1.




This Discussion