05-04-2005 03:43 AM - edited 03-03-2019 09:28 AM
I understand how the feasible successor is selected what I don't understand is why....
Say you have 3 paths between A and B
Path AD FD
1 20 30
2 30 40
3 200 210
With all links up path 1 is the successor and path 2 is the feasible successor.
Now if path 1 fails, path 2 being the feasible successor becomes the successor and is placed in the routing table.
But because Path 3's AD is greater than the current routes FD it does not become a feasible successor.
What I don't understand is why????
Surely having a feasible successor is better than having to go to the network and request routes.
Solved! Go to Solution.
05-04-2005 04:31 AM
Matt,
This is a scheme to avoid loops. What we are saying is that the next hop router for path 3 reckons he is 200 units away from the destination. We reckon that we are 40 units away from the destination. There is a risk that he learned his route from us - EIGRP is DistanceVector, so there is no way of telling. So to avoid the risk of a loop, we will ignore his route.
Does that make sense?
Kevin Dorrell
Luxembourg
05-04-2005 05:16 AM
I think that the explanation that Kevin gave is pretty good. The most important aspect of feasible successor is that the feasibility test guarantees that the feasible successor is a loop free path to the destination.
So to go back to the original question: yes having a feasible successor is better than not having one. But being sure that it is loop free is even more important. So if 1 fails 2 can be immediately put into the routing table because we know that it is loop free. If 2 fails we can not immediately put 3 into the table. 3 may in fact be loop free and may eventually be put into the routing table but we must querry the network first to find out if it is the best loop free path.
So the main difference between potential replacement route and feasible successor is that the feasible successor is guaranteed to be loop free. And 3 will not become feasible successor because based on what the router knows it can not guarantee that 3 is loop free.
HTH
Rick
05-04-2005 04:31 AM
Matt,
This is a scheme to avoid loops. What we are saying is that the next hop router for path 3 reckons he is 200 units away from the destination. We reckon that we are 40 units away from the destination. There is a risk that he learned his route from us - EIGRP is DistanceVector, so there is no way of telling. So to avoid the risk of a loop, we will ignore his route.
Does that make sense?
Kevin Dorrell
Luxembourg
05-04-2005 05:16 AM
I think that the explanation that Kevin gave is pretty good. The most important aspect of feasible successor is that the feasibility test guarantees that the feasible successor is a loop free path to the destination.
So to go back to the original question: yes having a feasible successor is better than not having one. But being sure that it is loop free is even more important. So if 1 fails 2 can be immediately put into the routing table because we know that it is loop free. If 2 fails we can not immediately put 3 into the table. 3 may in fact be loop free and may eventually be put into the routing table but we must querry the network first to find out if it is the best loop free path.
So the main difference between potential replacement route and feasible successor is that the feasible successor is guaranteed to be loop free. And 3 will not become feasible successor because based on what the router knows it can not guarantee that 3 is loop free.
HTH
Rick
05-04-2005 05:26 AM
Thanks you both that has been very helpful
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide