An interesting addendum is that this behavior is important to consider
when attempting to use the variance command to influence path selection.
I was configuring a variance scenario once and it wasn't working
despite all my efforts. It wasn't working because I didn't understand
this issue well enough.
Variance only works on paths through feasible successors. In the first
example below, Router C is a feasible successor so its path would be
available for consideration when using variance. However, if Router C's
AD is higher than the current FD, it will not be considered when using
variance.
In my case, I had tweaked the metrics to create just this situation and
it was driving me batty. The router in question was only considering
one path and would completely ignore the other. After posting this to
the list, someone quite correctly pointed out the problem.
John
>>> "[EMAIL PROTECTED] (Timothy Ouellette)"
4/19/02 12:16:28 PM >>>
Excellent description John. Reinforced my understanding of FD and AD
Tim
On 19 Apr 2002 12:07:22 -0400, [EMAIL PROTECTED] ("John
Neiberger") wrote:
>The key words here are Feasible Distance and Advertised Distance, or
in
>this case, reported distance. The FD is the metric for the current
>path. If we have two neighbors who are reporting that they can reach
>that network, both will be advertising what *their* FD is, which from
>our perspective is the Advertised Distance.
>
>
>[RA]-------(10)--------[RB]-----(50)-----Network Z
> \ /
> \ /
> \------(20)--------[RC]-----(50)------/
>
>
>Imagine that somewhere beyond Routers B and C is a Network, Z. They
>each have a metric of 50 to that network, which is their FD. Router
A
>will see two available routes to Z but the one through B has a lower
>metric and it will be installed into the routing table.
>
>However, because RC's Advertised Distance to Z (50) is less than
Router
>A's current FD (60), it will be installed as a feasible successor.
If
>the metric from Router C to Network Z was 60 or over, it would not be
a
>feasible successor. In that case, if the link from A to B were to go
>away, A would not immediately begin using RC as the next hop to Z.
>Instead it would send queries to all of its EIGRP neighbors and it
would
>start forwarding to C after C answers that it can reach Z.
>
>I hope that makes sense. I have a cold and am fairly medicated right
>now. ;-)
>
>John
>
>
>>>> "Sean Wolfe" 4/19/02 9:34:03 AM >>>
>EIGRP question:
>
>According to Cisco's website: "Feasible distance is the best metric
>along a
>path to a destination network, including the metric to the neighbor
>advertising that path. A feasible successor is a path whose reported
>distance is less than the feasible distance."
>
>But wouldn't a route with a distance less than the feasible distance
be
>in
>the routing table already, since it had a better metric?
>
>It makes more sense to me that the feasible successor is a route with
>a
>slightly larger metric than the current route. That way if the
current
>route
>dies, the next-best path is promoted.
>
>But that's not what I'm reading at
>http://www.cisco.com/warp/public/103/eigrp1.html#6
>
>So . . . whaddya say?
>
>Thanks, -Sean.
Message Posted at:
http://www.groupstudy.com/form/read.php?f=7&i=41996&t=41957
--------------------------------------------------
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
