[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: delayed multihoming/mobility set-up
If we assume that we are using BGP for the EGP routing updates, we have
the Ahuja/Labovitz route cancellation effect. If I remember correctly,
their research showed that 40% of re-routing takes 2-4 minutes (I am
taking this out of my head).
That is a reason why global routing table should be shrinked.
This would IMHO give the reference of how
long a "rehoming event" could take. Anything lower than that is a bonus
It is not the reference for M6 as the proper timing is application
It, instead, is the reference for BGP++.