Manav,

please look at draft-pillay-esnault-ospf-flooding-03.txt

Peter

Manav Bhatia wrote:
>
> Hi,
> If the LS age field of one of the router's self-originated LSAs reaches the
> value LSRefreshTime (30 mins), a new instance of the LSA is originated, even
> if the contents of the LSA (apart from the LSA header) are same. So if a
> router has some 2000 LSAs whcih is a very nominal figure then it will
> require refreshing every 1 sec with no other activity and nothing
> interesting happening on the network. This periodic nature of OSPF routing
> traffic requires demand circuit's underlying data-link connection to be
> constantly open resulting in unwanted user charges. To avoid this "do not
> age" LSAs were introduced which in theory we can press into service and are
> never aged. These were introduced in rfc 1793 [OSPF over Demand Circuits].
> Why is this option not being explored when dealing with scalability issues
> in OSPF .. Imagine the avg refresh time if the number of LSAs go beyond
> 50,000 .. It would run into milliseconds ! Why are these not being
> considered as a part of OSPFv3 ?
> Is it because these are not backward compatible and that we need to change
> the entire network before deploying this ?
>
> Thanks and regards,
> Manav

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]