BGP Route Reflector
Within a BGP AS, all iBGP routers must create neighbor peerings in a full mesh arrangement. As the number of iBGP routers grows, so does the complexity of creating full mesh peerings.
To solve this problem, BGP introduced a route reflector (RR). Instead of having every router peer with every other router in an AS, each router peers only with the RR. Advertisements are sent to the route reflector, which are then "reflected" out to all other BGP speakers.
A route reflector will only reflect routes that are considered "best routes", and it is not allowed to change any reflected route attributes, including the next hop IP address.
In a sense, the RR plays a similar role to the DR in OSPF.
An RR will use the Cluster ID to prevent loops within the topology.
Links
Links to this page:
- BGP - RRs vs confederations
- BGP - iBGP full-mesh peering
- BGP - iBGP split horizon rule
- BGP Cluster ID
- BGP Originator ID
- BGP installing a BGP-learned route into the routing table
- BGP using next-hop-self with a route reflector
- Network Design - Choosing a technology for multiple datacenter topology
- Network Design Considerations Involving PE and P Routers
- SD-WAN - vSmart