9.5 Redundancy, Symmetry, and Load Balancing
9.5.5 Load balancing
Load balancing is the capability to divide data traffic over multiple connections. When a BGP speaker learns two identical EBGP paths for a prefix from a neighboring AS, it will choose the path with the lowest route ID as the best path. This best path is installed in the IP routing table. If BGP multipath support is enabled and the EBGP paths are learned from the same neighboring AS, instead of picking one best path, multiple paths are installed in the IP routing table.

To enable BGP load balancing over equal-cost paths, use the maximum-paths command, which has the following syntax:

Router(config-router)# maximum-paths number

BGP supports a maximum of six paths per destination, but only if they are sourced from the same AS. By default, BGP will install only one path to the IP routing table.

The figure illustrates how inbound and outbound traffic behaves. The path for outbound traffic to reach NetA depends on where NetA is learned. Because NetA is received from both San Francisco and New York, outbound traffic toward NetA can go via San Francisco or New York.

On the other hand, the path for inbound traffic to reach the local networks, NetB and NetC, depends on how these networks are advertised. If you advertise NetC over the New York link only, then incoming traffic toward NetC will take the New York link. Similarly, if you advertise NetB over the San Francisco link only, traffic toward NetB will take the San Francisco link.