Re: EIGRP Doubt

I don't know that it make any difference whether the topology is small
or large.

FC simply requires that a neighbor's RD be less than the locally-
computed FD. No doubt there's plenty of good discussion of this
concept in TCP/IP Vol I. Basically, if a neighbor's RD is less than
the locally-computed FD, then the path in question must not loop
through this local router; such a path would include this local
router's metric *plus* the neighbor's cost to reach this local
router. So, an example from the below might be R2's and R3's
potential routes to the R1-R5 subnet (never mind split horizon or any
of that for the purposes of FC discussion in such a simplified
example). From R2's perspective, any path that loops from R3 via R2
itself would include R2's cost to the R1-R5 subnet *plus* R3's cost to
reach R2 via the R2-R3 subnet and thus couldn't possibly meet the FC.
However, again from R2's perspective, it's safe to assume that if R3
advertises a route to the R1-R5 subnet whose RD is less than the
current R2 FD, then it must be via R6 (or possibly even R4->R6,
depending on where those lines are meant to connect) and R2 can safely
mark that as an FS for the destination.

Just because the FC for a given potential path doesn't meet the FC
doesn't necessarily mean that the route loops; it just means that it
>could< and thus some querying would have to take place to sort it
all out. Again, I think Vol I is probably a great source to step
through this concept. But again briefly, if the R3-R6 subnet
(assuming the lines were meant to depict one) and the R5-R6 subnet
were very high cost links, then probably the FC will not be met from
R2's perspective, meaning any path following these links cannot safely
be flagged as FS and immediately failed over to if necessary.
However, if R2's current route to the R1-R5 subnet is via R1 and R1's
interface to that subnet fails but R5's does not (it's an Ethernet
segment and there's a switch in between the two), R2 can still query
R3 and ultimately find that those high-cost links constitute a valid,
loop-free path to the destination. It just takes longer than
immediately switching over to an FS.

On Jan 5, 2011, at 11:25 , Routing Freak wrote:

> Feasibility condition in EIGRP is used to avoid routing loops
> Can anyone explain me in how it avoids loops wid a small topology
>
>
> R1----------------------R2------------------------R3----------------R4
> |
> | |
> |
> | |
> |
> | |
> R5----------------------
> R6---------------------------------------------|
>
>
> Blogs and organic groups at http://www.ccie.net
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html

Blogs and organic groups at http://www.ccie.net
Received on Wed Jan 05 2011 - 12:29:00 ART