Peter, Two other points ..
#1 - Imagine PE is performing control plane restart or ISSU. How ABR will be able to detect this and instead of keep the potential disruption local to the area ? Note that data plane of the PE is working all the time just fine. #2 - The PULSE expiration timer may be different between various implementations. It may even be configurable. So you will have a network where there are many different timers in place. Now further imagine that PE restarting (even without data plane being maintained) within IBGP holdtime ... less them 180 sec. So remote PEs never see a BGP path removal of the "affected" next hops. Does the proposal assume that after removing the path based on PULSE reception the remote PEs will all by itself recompute best path after PULSE timer expires (no other control plane event will be seen). Many thx, R. On Wed, Jan 5, 2022 at 12:22 PM Robert Raszuk <[email protected]> wrote: > Peter, > > Two other points .. > > #1 - Imagine PE >
_______________________________________________ Lsr mailing list [email protected] https://www.ietf.org/mailman/listinfo/lsr
