Hi,
I will reffer to the following topology:
I am trying to identify scenarios in which loop or sub-optimal routing will occur due to different routing protocols and redistribution. I am presuming OSPF of EIGRP under IGP with standard AD.
So far, I have identified total of five (5) scenarios:
1. IGP A prefix gets redistributed from IGP B to BGP.
- Result: sup-optimal routing.
- Explanation: Site A prefixes will get routed over Site B
2. IGP B prefix gets redistributed from IGP A to BGP.
- Result: sub-optimal routing.
3. On of the routers has configured weight attribute pointing to one of eBGP peers (e.g. A2 towards B2).
- Result: sup-optimal routing.
- Explanation: A2 will route intra-site traffic to B2 due to IGP B to BGP redistribution on B2 since A2 will receive route with larger weight.
4. A1 and A2 are redistributing iBGP routes.
- Result: routing loop.
- Explanation: A1 then receives the IGP B prefix from A2 which gets installed int RIB due to lower AD of IGP A. Then, A2 looses its eBGP peering, removes its eBGP route from the RIB but due to crazy timing, receives iBGP route from A1. A2 will continue to redistribute iBGP route in OSPF and.
5. A1 and A2 (or B1, B2) have different AS numbers.
- Result: routing loop.
- Explanation: A1 and A2 will receive each other's prefixes for Site A and due to AD 20 the route will be installed and preferred in the RIB.
I would be very thankful if you can validate my thoughts, and maybe to identify a unique scenario that I inadvertently omitted.
Many thanks!
