multiple red points methodology - longish

From: Ozgur Rascall (polarccie@yahoo.co.uk)
Date: Fri May 09 2003 - 15:43:37 GMT-3


Hello everybody,
during mutual red in one router,
it is wise to select the routes to redistribute, since
the higher ad protocol might suffer, lose its
routes...
and it is enough to filter in the direction
lower ad protocol ---> higher ad protocol

with 2 red points case we face often,
i see that in scenarios, routes are still being
filtered,although as Doyle states "this causes a loss
of inherent redundancy of topology..."( "any" filter
will actually spoil redundancy)
First q , does that matter in the lab?

possible solutions of 2 red points are
*filter both ways...!!!
*set distance to y > max(ad1,ad2) on both protocols,
set the original ads from for routes learned from
neis...
??*set distance to z < min(ad1,ad2) on the higher ad
side for the routers learned from neis..no filter..

Does the last one make sense to you??
please comment...
Ozgur



This archive was generated by hypermail 2.1.4 : Mon Jun 02 2003 - 15:13:40 GMT-3