Re: Dial backup - purpose?

From: Kevin Baumgartner (kbaumgar@xxxxxxxxx)
Date: Wed Jan 24 2001 - 16:03:57 GMT-3


   
Well dial backup can be used if a physical interface goes down or is faulty
and then gives a backup routing path to get to the disconnected networks.
Dialer watch could be used if there is a disconnected network that is not
directly connected to the ISDN routers.

  When using dial backup it will only take the BRI interface out of standby
to active. It will not place a call until packets are sent across the
interface.
If you are running a routing protocol this shouldn't be a problem.

   By the way, in a much earlier ios version it was possible to use dial backup
and the call would be placed without packet across the BRI interface.

  Kevin

At 01:39 PM 1/24/01 -0500, Rob Webber wrote:
>I have been looking at dial quite a bit lately. I'm not sure how/why you
>would use dial backup. I understand the basic premise - that it places a BRI
>or dialer interface into standby and watches for the main interface to go
>down. However even the Cisco doc's indicate that actual dialing only occurs
>when there is a packet to be transmitted (via the dialer-list command).
>
>It seems to me (short of using dial watch) the only way you can get it to
>dial is by using floating static routes to point traffic to the BRI or
>dialer interface. However if you use floating statics, you don't really need
>the backup feature at all.
>
>For example, in Cisco's example:
>
>http://www.cisco.com/warp/public/793/access_dial/britobribackup.html
>
>IGRP will not run on the BRI0 because there is no "network 192.168.10.0"
>under router igrp 1. So routes will not be dynamically learned once Serial0
>goes down. So the only way traffic can route between the two routers in the
>example is to have a static default route pointing to the BRI0 with a high
>admin distance. Yet if that's the case, why use the backup command at all?
>
>Dial watch seems to be a better solution - it watches specific routes in the
>routing table and actually dials as soon as they disappear (at which point
>EIGRP or whatever can exchange routes). This also gets away from the "no
>static routes" limitation we so often see.
>
>Thoughts?
>Thanks - Rob.
>



This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:27:42 GMT-3