The Attitude (was: the Internet Backbone)

Curtis Villamizar curtis at ans.net
Wed Apr 17 17:40:59 UTC 1996


In message <199604170236.WAA10645 at merit.edu>, EDS at RHQVM21.VNET.IBM.COM writes:
> > >> and it might even be possible to enforce some interesting
> > >> policies in that regard in the route servers (e.g. if you have more
> > >> than N routing or BGP peer transitions per time period, the route
> > >> server will refuse to peer with you for 48 hours - think of it as the
> > >> hold-down or damper from Hell).
> > >Interesting concept!
> 
> > Exponential backoff works well, too bad the code base that does
> > route flap damping isn't really stable enough for production
> > use in large parts of the Internet.
> > Best results are give, when everyone runs it.
> 
> Flap dampening works on most routers- just have to be aware of what
> the effects are, particularly with multiple BGP Peers.
> You have to decide whether to dampen outbound and/
> or inbound, and if short duration events like temporary
> recalculations of the IGP are considered a "flap".
> 
> Ed Segal  EDS at RHQVM21.VNET.IBM.COM   914 684-3976


Ed,

Route flap dampenning should only be applied to EBGP learned routes or
EBGP announcements.  It can be applied to inbound, outbound, or both,
but as defined is applied to inbound only.  It should never be applied
to IBGP or you will get routing loops.  It can only be applied to a
link adjacency by the router at the link so you keep a consistent IGP,
otherwise you get routing loops.

Please see:

  http://engr.ans.net/route-dampen

If you have any corrections to the usage, please let me know.  One of
these days I'll fix this thing up and submit it as a draft again.

Curtis



More information about the NANOG mailing list