From: Albert Lu (albert_ccie@xxxxxxxxx)
Date: Wed Nov 28 2001 - 05:46:17 GMT-3
Actually, sometimes route-maps just won't work and you require a
distribute-list. For example, IGRP will install its routes into the routing
table before you can filter it during the redistribution. So by using
distribute-list, you can prevent IGRP from ever learning the route from it's
next-hop and you're sweet!! =)
In regards to filtering routes being learnt from the next hop, you would
need to have knowledge of your current routing domain in order for you to
create the access-list required. What about if you have multiple subnets,
the access-list would be quite large. Plus, what about if you there is
another routing domain that is redistributed into this domain via another
router, then you have to make sure that those routes will not be fed back.
Albert
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
Hansang Bae
Sent: Wednesday, November 28, 2001 10:17 AM
To: ccielab@groupstudy.com
Subject: RE: flsm to vlsm
At 07:38 AM 11/28/2001 +1100, Albert Lu wrote:
>[snip]
> From my experience, it's best to use distribute-list for DV protocols and
>route-maps for LS protocols. What about the distribute-list 10 out ospf 10
>command, where you are applying the distribute-list to the routing protocol
>being redistributed in, would that be the same as using a route-map?
I've always been partial to RMs because to me they are more flexible. But
it's a personal choice. What you guys can practice is creating a RM that's
*THE* most efficient. Don't create a route-map that uses two access-lists
when it can be done with just one. You'll need to practice some 'not
logic' but it's doable.
hsb
This archive was generated by hypermail 2.1.4 : Fri Jun 21 2002 - 06:45:24 GMT-3