EBGP vs. IBGP

B

There must be a full mesh of iBGP sessions, in other words: each BGP router within an AS must have iBGP sessions with all other BGP routers in the AS. By requiring that all information in iBGP is learned directly from the router that learned the information over eBGP, there can't be any loops in iBGP. The full mesh requirement can be solved using either route reflectors or confederations.

BGP (Border Gateway Protocol) is a protocol used for routing data between different autonomous systems (ASes) in the Internet. BGP has two types of peering: EBGP (External BGP) and IBGP (Internal BGP).

EBGP is used between BGP speakers in different autonomous systems, while IBGP is used between BGP speakers in the same autonomous system. The main difference between EBGP and IBGP is the way they handle the routing information.

EBGP peering requires a full mesh of BGP speakers because EBGP speakers cannot readvertise EBGP-learned routes to other EBGP speakers. This means that each EBGP speaker must be directly connected to all other EBGP speakers in the network to ensure that all routes are propagated throughout the network.

On the other hand, IBGP peering also requires a full mesh of BGP speakers because IBGP speakers cannot readvertise IBGP-learned routes to other IBGP speakers. However, unlike EBGP, IBGP speakers can use a route reflector or confederation to avoid the full-mesh requirement.

A route reflector is a BGP speaker that reflects BGP routes to other IBGP speakers in the same autonomous system. The route reflector receives the routes from the IBGP speakers and reflects them to other IBGP speakers that are not directly connected to each other, thus reducing the number of IBGP peering sessions required.

A confederation is a way of dividing a large autonomous system into smaller sub-autonomous systems. Each sub-autonomous system has its own BGP speaker, and the BGP speakers in each sub-autonomous system communicate with the BGP speakers in other sub-autonomous systems using EBGP peering. This reduces the number of IBGP peering sessions required within each sub-autonomous system.

In summary, EBGP and IBGP are two types of BGP peering used for routing data between different and same autonomous systems, respectively. EBGP requires a full mesh of BGP speakers, while IBGP also requires a full mesh but can use a route reflector or confederation to avoid the full-mesh requirement.