In a traditional Layer 3 network, an IP address should be assigned to each interface to provide connectivity from one link endpoint to another. But in the data center network, with a large number of spine and leaf devices and the number of interfaces interconnecting them could be very large. This leads to a sharp increase in the number of IP addresses required. Such large networks can consume a lot of IP addresses, where each peer requires a separate IP address.
BGP unnumbered is very useful in this kind of scenarios. BGP unnumbered interface is a BGP interface that does not need to specify a unique IP address, it uses the IPv6 link-local address, which is assigned automatically to each interface. BGP unnumbered works by using the Extended Next Hop Encoding (ENHE) as defined in RFC 5549, which provides a way to advertise an IPv4 route with an IPv6 next-hop. The IPv6 next-hop is resolved by using the IPv6 Neighbor Discovery (ND) process. Prior to RFC 5549, an IPv4 route could only be advertised with an IPv4 next-hop.
For BGP unnumbered, the next-hop address for each prefix is an IPv6 link-local address, which is assigned automatically to each interface. Using the IPv6 link-local address as a next-hop instead of an IPv4 unicast address, BGP unnumbered saves you from having to configure IPv4 addresses on each interface.
NOTEs: BGP can only be used on two switches with point-to-point connection. |