OSPF Flood Reduction

Let’s say you’re in The Lab and are asked to minimize OSPF traffic on a link. What does that mean? How can you do it? A quick question mark of OSPF options under an interface will tell you about flood-reduction:

Flood-reduction sounds like it will do the trick, but the description isn’t too descriptive. What does it really do? The answer has to do with OSPF’s need to refresh the LSAs in its database. RFC 2328 states on page 216 that OSPF LSAs age out after 60 minutes and routers must refresh (flood) their LSAs every 30 minutes. This can create a fair amount of unnecessary traffic if your network a) has a large number of LSAs and b) is relatively stable. This is somewhat mitigated by Cisco’s LSA Group Pacing, but if you really want to cut down on network traffic, you can stop LSA refreshes altogether. A standard OSPF database looks like this:

You can see that these particular LSAs are only 5-11 seconds old and will age out in an hour (3600 seconds) if they’re not refreshed (every 1800 seconds). To stop the unwanted flooding of LSAs, enable flood reduction on the interface:

As you can see, the neighbor relationship dies and restarts (to clear the database of all LSAs from R1) and then all LSAs received from R1 are now marked DNA (DoNotAge). The opposite is true on R1, all LSAs advertised by 3.3.3.3 are marked DNA.

To make sure these LSAs don’t hang around when the route is really gone, we shut down an interface on R1. R0 still gets an update from R1:

Tagged with:

Leave a Reply