BGP RR Cluster-ID's Benefits & Cons?
Imagine you had a topology of this:
http://postimg.org/image/8xlmcngyr/
Client 1 and 2 are part of RR1
Client 3 and 4 are part of RR2
Very Simple
The question now is whether to configure the same cluster-id or not. So I want to evaluate the pro's and cons of configuring the same cluster-id, but I can only see cons at the moment. This is why:
If we use the same cluster-id on both RR's then any routes reflected from RR1 towards RR2 (and visa versa) would be dropped because of cluster-id being the same. This means that there would be a latency and CPU performance overhead of RR1 for sending the prefixes over to RR2, and then a CPU performance overhead of RR2 dropping the routes.However if different cluster-id's were used it means that either RR could actually install these prefixes which would only add a bit more memory requirements to install the routes. That is it! So why would you want to use the same cluster-id?
Comments
In that design, if you use the same cluster-id on RR1 and RR2, the problem is that routes advertised by client1 and client2 to RR1 will never be received/learned by client3/client4 as RR2 will reject the updates; same goes for routes advertised by client3 and client4 which will not be learned/received by cient1/client2.
So my question is what is the benefits of using the same cluster-id?
INE - The Industry Leader in CCIE Preparation
http://www.INE.com
Subscription information may be found at:
http://www.ieoc.com/forums/ForumSubscriptions.aspx
Ok I will find it.
OK I watched it. AMAZING. They should have put this in the V5 videos because Brain explains it a lot better in this older video IMO. It's made everything so much more clear to understand now. Thanks.
INE - The Industry Leader in CCIE Preparation
http://www.INE.com
Subscription information may be found at:
http://www.ieoc.com/forums/ForumSubscriptions.aspx