Reference Guide

Multiprotocol BGP
Multiprotocol extensions for BGP (MBGP) is defined in IETF RFC 2858. MBGP allows different types of
address families to be distributed in parallel.
MBGP for IPv4 multicast is supported on the S4820T platform.
MBGP allows information about the topology of the IP multicast-capable routers to be exchanged
separately from the topology of normal IPv4 and IPv6 unicast routers. It allows a multicast routing
topology different from the unicast routing topology.
MBGP uses either an IPv4 address configured on the interface (which is used to establish the IPv6
session) or a stable IPv4 address that is available in the box as the next-hop address. As a result, while
advertising an IPv6 network, exchange of IPv4 routes does not lead to martian next-hop message logs.
NOTE: It is possible to configure BGP peers that exchange both unicast and multicast network layer
reachability information (NLRI), but you cannot connect multiprotocol BGP with BGP. Therefore,
you cannot redistribute multiprotocol BGP routes into BGP.
Implement BGP with Dell Networking OS
The following sections describe how to implement BGP on Dell Networking OS.
Additional Path (Add-Path) Support
BGP add-path is supported on the S4820T platform.
The add-path feature reduces convergence times by advertising multiple paths to its peers for the same
address prefix without replacing existing paths with new ones. By default, a BGP speaker advertises only
the best path to its peers for a given address prefix. If the best path becomes unavailable, the BGP speaker
withdraws its path from its local RIB and recalculates a new best path. This situation requires both IGP
and BGP convergence and can be a lengthy process. BGP add-path also helps switchover to the next
new best path when the current best path is unavailable.
Advertise IGP Cost as MED for Redistributed Routes
When using multipath connectivity to an external AS, you can advertise the MED value selectively to each
peer for redistributed routes. For some peers you can set the internal/IGP cost as the MED while setting
others to a constant pre-defined metric as MED value.
Dell Networking OS supports configuring the set metric-type internal command in a route-map
to advertise the IGP cost as the MED to outbound EBGP peers when redistributing routes. The configured
set metric value overwrites the default IGP cost.
By using the redistribute command with the route-map command, you can specify whether a peer
advertises the standard MED or uses the IGP cost as the MED.
When configuring this functionality:
• If the redistribute command does not have metric configured and the BGP peer outbound
route-map does have metric-type internal configured, BGP advertises the IGP cost as MED.
• If the redistribute command has metric configured (route-map set metric or
redistribute route-type metric) and the BGP peer outbound route-map has metric-type
194
Border Gateway Protocol IPv4 (BGPv4)