3 Replies Latest reply: Jun 4, 2011 3:55 PM by Sey

# When will EIGRP Feasible Successor be chosen?

Hello, I have been thinking about EIGRP and convergence and how it works because of feasible successor being present but am confused as to when this applies in real world scenarios where nowdays most links have equal delay and bw. For eg. in this simple diagram assume all links are 1Gb and delay is 10. When RT A advertises the route 10.0.0.0/24 to RT B (Metric 10) and RT C (Metric 10) and they in turn advertise to RT D and RT E respectively. RT D will install the route learnt via RT B with metric (20 / 10) and RT E will install the route learnt via RT E with metric (20 / 10).

Now RT D and RT E will also advertise their best route to each other but this route will never be chosen as a Feasible successor because the metric will be equal. RT D will see the route from RT E with metric ( 30/20 ) and RT E will see the route from RT D with metric (30 / 20). Since Advertised distance of the route from the other side is equal to the successor route's metric ( 20 = 20). So now what is the advantage of having this redundant route. There will be reconvergence and additional delay. Hence in this case would another routing protocol be better choice?

Is my understanding correct and does EIGRP rely on design to achieve goal?

Thx

• ###### 1. Re: When will EIGRP Feasible Successor be chosen?

EIGRP does not look at routes in terms of feasible successors from the perspective of their own routing table metric, they look at the RD - reported distance - from the neighbor who sourced the update.

In the case where the reported distance is less than or equal to the successor(there is some debate as to whether it is merely less than, or less than or equal to), it meets the feasibility condition and is chosen as the feasible successor.

You have to remember it is using this as a loop prevention technique as if the metric is less than or equal to the successor route to the destination prefix/network, it is understood that the route did not  already pass through the local router before coming back to it.  This is done because EIGRP is not a true link state protocol but instead is very much a distance vector protocol at heart, with several key improvements(like the successor/feasible successor concept).

Hope that helps, please let us know if you still have doubts.

• ###### 2. Re: When will EIGRP Feasible Successor be chosen?

Now RT D and RT E will also advertise their best route to each other but this route will never be chosen as a Feasible successor because the metric will be equal.

Advertised distance of the route from the other side is equal to the successor route's metric ( 20 = 20). So now what is the advantage of having this redundant route.

You are right, the second route will not become a Feasible Successor because it will not meet the feasibility condition. Feasibility condition (FC) says that a route can become a feasible successor on a router X only if the advertising router Y is closer to the destination network than the router X itself (through the successor route of course). The purpose of FC is to make sure there is no chance for a routing loop.

Let's assume that router A has a metric of 1 to the network 10.0.0.0/24. Then both routers D and E will have a successor with metric 10+10+1=21. Now router D hears an advertisement from router E about 10.0.0.0/24. This route would have a metric of 21+10=31. But it won't be considered at all as router E is the same distance (21) from 10.0.0.0/24 as router D itself, so the route does not meet the FC.

An example of equal metrics (this time I mean feasible distance, FD) is any network segment from the viewpoint of the respective farthermost router. Let's take the segment between A and C. From router D's point of view we can get to it by two equal paths - either through routers B and A, or through routers E and C. The metrics are equal in both cases: 10+10+10=30. This time we don't get a feasible successor because both paths are equal, so they both are Successors and as such will be both installed in the routing table for load balancing (provided there are no other routing protocols with lower administrative distance, in which case no eigrp routes will be installed).

P.S. I really like your diagram, the equilateral pentagon is ideal for learning successors, and feasible successors, and their metrics!

• ###### 3. Re: When will EIGRP Feasible Successor be chosen?

Hi Travis,

In the case where the reported distance is less than or equal to the successor(there is some debate as to whether it is merely less than, or less than or equal to), it meets the feasibility condition and is chosen as the feasible successor.

I think we agreed on the strict "less than". It might be otherwise long time ago, but now it is so.

https://learningnetwork.cisco.com/message/120371#120371