BGP

BGP

1. Design Principles

   • Provide loop-free routing through the use of path vectors
   • Minimize routing information volume via incremental updates
   • Reliable exchange of route information using a reliable transport – TCP
   • Enable incremental addition of new features by encoding BGP information as <type, length, value> triplet attributes

2. eBGP

   • exchanging route information between AS

3. iBGP

   • distributing route information to BGP routers within an AS
     • does not advertise routes learned from other iBGP peers
     • does not change attributes and next-hop of routes learned from other iBGP peers

4. Features

   • Neighbor must be reachable by IP before BGP can run
   • Advertise new Network Layer Reachability Information (NLRI) and/or withdraw a previously advertised NLRI
   • Each update message contains one set of prefixes with the same attributes

5. Route Information Volume versus Route Optimality

   • Default Route Only