I'm sure many of us know that split-horizon is disabled in RIP when running
over a frame-relay physical interface. That is well understood. Now,
imagine you have a hub and spoke setup R1, R2, R3 where R1 is the hub and
R2/R3 are spokes. Everything runs RIP on frame-relay physical interfaces.
That means that R1 has split-horizon DISABLED by default and R2/R3 will see
each others routes.
NOW...I am trying to find any kind of documentation that states the hub
won't actually send RIP updates out the same interface it learned them on
out the SAME DLCI it learned them on. For example, with the above setup --
if split-horizon is indeed disabled what is stopping R1 from learning routes
from R2 and sending them right back out to R2?
In my lab test on this topic, I found that a "debug ip rip" on the hub shows
that R1 does indeed send the routes from R2 right back to R2. However,
"debug ip rip" on R2 never shows those same routes being received, and
nothing is added to the RIP database on R2.
Perhaps it is just because when R2 sees the route it originated coming back
in from the hub the metric is obviously higher than the one R2 has....but
I'm thinking I should at the very least see the update in "debug ip rip" and
I'm not quite sure why I don't.
Thoughts?
-- Regards, Joe Astorino CCIE #24347 Blog: http://astorinonetworks.com "He not busy being born is busy dying" - Dylan Blogs and organic groups at http://www.ccie.netReceived on Wed Jul 13 2011 - 19:51:01 ART
This archive was generated by hypermail 2.2.0 : Mon Aug 01 2011 - 06:30:05 ART