Re: OSPF

If you go back to the original post he asked if there was a way to modify
the next hop, not if there was a way to keep from doing so. Based on that
it sounded like NBMA or something. Still wouldn't a change of network
type be sufficient here?

From:
Dale Shaw <dale.shaw_at_gmail.com>
To:
Keegan.Holley_at_sungard.com
Cc:
MDevarajan_at_inautix.co.in, ccielab_at_groupstudy.com
Date:
11/22/2009 07:15 PM
Subject:
Re: OSPF
Sent by:
<nobody_at_groupstudy.com>

Hi,

On Mon, Nov 23, 2009 at 10:07 AM, <Keegan.Holley_at_sungard.com> wrote:
>
> I think you're getting your features crossed. The point of bgp next hop
self is that your route to the next hop may not be advertised to all of
your bgp routers since it is not meant to be part of the bgp topology.
OSPF has no such complications so most of the time the next hop problem is
based on media alone. Hence you just change the next hop value or map the
next hop to a usable layer-2 address.

Umm, who mentioned BGP? :-)

The question, paraphrased, was "is there an OSPF equivalent to the
_EIGRP_ command "no ip next-hop-self eigrp".

This command is useful in DMVPN networks, where despite being a NBMA
interface, only a single neighbour adjacency is maintained from the
spoke's perspective and you need to allow for direct spoke-to-spoke
communication without hair-pinning all traffic through the hub. With
the above command, the hub advertises routes to the spokes without
modifying the next-hop IP address. The result is that 'spoke A' learns
a route to a network attached to 'spoke B' directly from the hub, but
the next-hop IP is set to 'spoke B'.

cheers,
Dale

Blogs and organic groups at http://www.ccie.net
Received on Sun Nov 22 2009 - 20:07:53 ART