From: Joe Chang (changjoe@earthlink.net)
Date: Sat Dec 28 2002 - 17:08:00 GMT-3
True, without route filtering the network in Doyle's example will not suffer
any sub-optimal routing because of internal EIGRP's lower cost than ISIS,
and ISIS's lower cost than external EIGRP. However by leaving the feedback
routes in the routing table of C and D you risk temporary routing loops if
networks dissappear in the EIGRP network. Say the ethernet on router A goes
down, router A will query its EIGRP neigbors for another route to that
ethernet. D will report that it has a route from F, and C will also report
an available route from E. A will then place two new routes in its table for
an ethernet that is quite obviously down. This loop will clear up because
ISIS will eventually discover the loss of the original route, but the table
at A will look pretty funny for a while.
----- Original Message -----
From: "Jorge Epifbnio" <jorge.epifanio@contacto5.icep.pt>
To: <ccielab@groupstudy.com>
Sent: Saturday, December 28, 2002 3:28 PM
Subject: Redistribution EIGRP/ISIS
> When redistributing EIGRP/ISIS (on routers C and D) is it really
> necessary to prevent loops and route feedback?
>
> An EIGRP route on a border router will have AD=90 and AD=115 (feedback
> from EIGRP route) so it will always choose the internal path through the
> EIGRP domain.
>
> An ISIS route on a border router will have AD=115 and AD=170 (feedback
> from ISIS route) so it will always choose the internal path through the
> ISIS domain.
>
>
>
>
>
> A---B
>
> | |
>
> C D
>
> | |
>
> E---F
>
>
>
> (Doyle page 799 Exercise 5)
> .
.
This archive was generated by hypermail 2.1.4 : Fri Jan 17 2003 - 17:21:54 GMT-3