Showing results for 
Search instead for 
Did you mean: 



Pls, refer to the document to answer the question.
R1 is in different AS, R2 is in different AS.
My router R has connections to both AS.
Imagine: (Determine route from R to AS)
1. Next hop is reachable
2. Local preference is the default value which is 100
3. AS path is also the same.
4. All routes have originated from iBGP, hence origin is also the same.
5. By default, MED can't be compared between two different AS. We can go for load balancing. However, I have placed 'always compare MED' in the configuration. Both MED is compared and unfortunately, it is the same.
6. Prefer the External path. Both paths are external.
7. Prefer the metric. Unfortunately, both are same which is 10
8. Prefer the route with lowest router ID. Both are same.
9. Prefer the lowest cluster ID. Here, also both are same.
10. Prefer the lowest peer ID. It is also the same.

My question is what happens when all the route attributes fail. What will happen next? How does the route propagate from R to any of the AS after 10th step?
Very eager to know about it.


Georg Pauwen
VIP Master VIP Master
VIP Master



there are three more selection criteria (see document attached). The last criteria is to prefer the path that comes from the lowest neighbor address (that is, the IP address configured with the 'neighbor' command).

What if the last selection criteria 'neighbor IP address' for both AS remains same?

What will happen next after all the criteria stated in the document are same for both paths?




in theory, one of the criteria should apply. If you have a test lab, you might want to test and see what happens when all 13 criteria are identical. I have never seen it, and the guys writing the BGP RFCs thought it through (I suppose), so in real life, there should never be a situation where all 13 are the same...

I might mention, to OP, if somehow this situation arises, whatever results you obtain probably could not be guaranteed to be the same on other platforms and/or on different IOS versions on the same platform, or even possibly different results on the same platform and IOS, during different tries, or even possibly an IOS fault.

In other words, Georg's suggestion to try this as an experiment in a test lab, is fine, but whatever the results, don't assume it will always be the same. When you do something undefined, results are generally inconsistent.

Awesome. That clears my doubts. Thank you so much. 

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: