|
Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com |
From: Manav Bhatia (manav
SAMSUNG.CO.KR)Date: Mon Sep 10 2001 - 06:55:38 CDT
Consider a setup wherein you can reach a network X via 2 ways. One, through
the routers within the area and the second through the routers lying in some
other areas. Also given that the route via the routers lying in other areas
is better (say a better less congested link) . What happens ?
OSPF would choose the path through the routers lying in its own area. Its
only when its not able to do so does it look at the routes in the other
areas. Now even though the route crossing the other areas is better we still
select the inferior route i.e. one within our own area. The same happens in
case of IS-IS, as long as the router is able to reach X through the L1
router it will never access the L2 routers.
A similar example can be extended for BGP which aggregates the routes it
learns from different autonomous systems.
Hope it helps.
Manav
----- Original Message -----
From: <swatirstogi
YAHOO.COM>
To: <OSPF
DISCUSS.MICROSOFT.COM>
Sent: Monday, September 10, 2001 4:30 PM
Subject: Optimal Routes
> Lets define an optimal route as a route that we would have computed in the
> absence of any areas/levels i.e. , an optimal route in a given network
that
> is composed of a single OSPF/IS-IS routing domain is a route that is
> computed if the network isn't partitioned into areas.
>
> I understand that partitioning a network into areas results in a loss of
> routing information. My question is that how does this lead to suboptimal
> routing wherein the computed routes will be different than the "optimal"
> routes ?
>
> This is important because a corollary of this is that it is impossible to
> gaurantee optimal routing in the presence of routing information
> aggregation/summarization.
>
> Regards,
> Swati
>
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]