RE: draft-ietf-vrrp-unified-spec-00 issues

> What I think Steve is saying is: "We are seeing more and more 
> requests to support rates of failure detection this draft 
> addresses but it seems likely that in another decade, with 
> terabit speed links for example, that we will see demands for 
> failure detection 100 times faster than we currently address." 

And what I'm questioning is whether this is really true: we are
already into the range where the speed of the links are fast
enough so that the VRRP overhead at any interval would be a
negligible consumption of bandwidth. So the fact that link
rates are likely to keep going up is not particularly
interesting; the failover rate is already approaching the point
where the limiting factor is the infrastructure switchover
speeds, not the bandwidth. And that means the limit starts to
involve speed of light issues more and the limitations of
hardware designs less.

I don't claim to know the answer, I'm just saying that we're
very near the tipping point where we cannot extrapolate based
on past experience. So someone that actually understands the
hardware issues should consider what the limit will be of what
people will demand and what hardware will be able to support.
I don't really have any idea, so I have no reason to favor any
of the proposed units, including scalable.

Can anyone tell me what's going on in routing protocols in
failure recovery? In a "normal" configuration, returning
traffic will tend to go to the dead router until the routing
protocol detects the failure. How long does that take these
days?

-don

_______________________________________________
vrrp mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/vrrp