Route Flaps at NAP's
huddle at mci.net
Wed Sep 17 16:10:56 UTC 1997
Merit has done studies on this in the past. Their data is on
That said, why are route flaps important? In the past, its
been said that certain length prefixes were the cause of lots of
routing instability on the Internet, and much effort was
spent on developing elaborate dampening algorithms and policy
efforts were spent on determining how to set the parameters
on such a system.
As it turns out, these aren't important ... prefixes flap pretty
much independent of length. And while dampening will help prevent
lots of up/down cyclic behavior and resulting churn, the whole issue,
I think, got a lot more intellectual curiosity that it deserved,
not to dissuade you ;)
Other than it being something that one can count, I'm not sure why
route flaps, in general, are interesting.
At 11:38 AM 9/17/97 -0400, Ramesh Kalathur wrote:
>I have couple of questions on Route Flaps at NAP's.
>1)How often do they happen? (per hour, per day etc)
>2)When route flaps occur, on an average how many routes are flapped?
>Any pointers indicating the above statistics will be helpful.
More information about the NANOG