cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
962
Views
1
Helpful
2
Comments
Meddane
VIP
VIP

By design, OSPF requires link-state advertisements (LSAs) to be refreshed as they expire after 3600 sec.

When OSPF network is stable. The OSPF Flooding Reduction feature is used to reducing unnecessary refreshing and flooding of already known and unchanged information. To achieve this reduction, the LSAs are now flooded with the higher bit set called DNA bit, thus making them DoNotAge (DNA) LSAs.

To suppress the unnecessary flooding of link-state advertisements (LSAs) in a stable topology we can use the following commands:

-With OSPFv2 use the ip ospf flood-reduction command.

-With OSPFv3, use the ipv6 flood-reduction command.

-With OSPFv3 AF, use the ospfv3 flood-reduction command.

RFC 2328 says on page 216 that OSPF LSAs age out after 60 minutes :

B. Architectural Constants

   LSRefreshTime

       The maximum time between distinct originations of any particular

       LSA. If the LS age field of one of the router's self-originated

       LSAs reaches the value LSRefreshTime, a new instance of the LSA

       is originated, even though the contents of the LSA (apart from

       the LSA header) will be the same. The value of LSRefreshTime is

       set to 30 minutes.

   MaxAge

       The maximum age that an LSA can attain. When an LSA's LS age

       field reaches MaxAge, it is reflooded in an attempt to flush the

       LSA from the routing domain (See Section 14). LSAs of age MaxAge

       are not used in the routing table calculation. The value of

       MaxAge is set to 1 hour.

Therefore routers must refresh (flood) their LSAs every 30 minutes. This can create a fair amount of unnecessary traffic if the LSDB (Link State Database) has a large number of LSAs and is relatively stable. To mitigate the this behavior use the flood reduction feature.

The RFC 4136 explains the flood reduction feature, it allows the routers to flood their self-originated LSAs with the DoNotAge (DNA) bit set:

2. Changes in the Existing Implementation

  This enhancement relies on the implementation of the DoNotAge bit and

  the Indication-LSA. The details of the implementation of the

  DoNotAge bit and the Indication-LSA are specified in "Extending OSPF

  to Support Demand Circuits" [2].

  Flooding-reduction-capable routers will continue to send hellos to

  their neighbors and keep aging their self-originated LSAs in their

  database. However, these routers will flood their self-originated

  LSAs with the DoNotAge bit set. Thus, self-originated LSAs do not

  have to be re-flooded every 30 minutes and the re-flooding interval

  can be extended to the configured forced-flooding interval. As in

  normal OSPF operation, any change in the contents of the LSA will

  cause a reoriginated LSA to be flooded with the DoNotAge bit set.

  This will reduce protocol traffic overhead while allowing changes to

  be flooded immediately.

  Flooding-reduction-capable routers will flood received non-self-

  originated LSAs with the DoNotAge bit set on all normal or flooding-

  reduction-only interfaces within the LSA's flooding scope. If an

  interface is configured as both flooding-reduction-capable and

  Demand-Circuit, then the flooding is done if and only if the contents

  of the LSA have changed. This allows LSA flooding for unchanged LSAs

  to be periodically forced by the originating router.

Another way to reduce flooding of unnecessary LSAs is the OSPF demand circuit options.

The OSPF flooding reduction and demand circuit features us ethe same mechanism to eliminate the need for the periodic LSA refresh in other words they set the DNA bit.

The main difference is that the OSPF demand circuit suppresses hellos while the OSPF flood reduction does not.

2 Comments
Martin L
VIP
VIP

Interesting.... thanks for sharing!

Joseph W. Doherty
Hall of Fame
Hall of Fame

An interesting feature, but any guidelines when it's really beneficial to use?  I.e. the tipping point?

I have in mind both that modern networks have much more link capacity then they often had years ago, and OSPF routers (or L3 switches) also generally have more control plane processing capacity.  So, when does LSA flooding really become impactful?

I also have in mind, the Cisco OSPF implementation where they provide several LSA pacing enhancements (group packet-pacing, flood packet-pacing and retransmission packet-pacing), so there's not one huge flood of them at any one time.

I also have in mind, classical OSPF design techniques in using stub areas and/or summarization on ABRs, both of which can much limit the volume of LSAs being flooded within a multi-area OSPF topology. 

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: