API Guide

Multiprotocol BGP
Multiprotocol BGP (MBGP) is an extension to BGP that supports multiple address families—IPv4 and IPv6. MBGP carries
multiple sets of unicast and multicast routes depending on the address family.
You can enable the MBGP feature on a per router, per template, and/or a per peer basis. The default is the IPv4 unicast routes.
BGP session supports multiple address family interface (AFI) and sub address family interface (SAFI) combinations, BGP uses
OPEN message to convey this information to the peers. As a result, the IPv6 routing information is exchanged over the IPv4
peers and vice versa.
BGP routers that support IPv6 can set up BGP sessions using IPv6 peers. If the existing BGP-v4 session is capable of
exchanging ipv6 prefixes, the same is used to carry ipv4 as well as ipv6 prefixes. If the BGP-v4 neighbor goes down, it also
impacts the IPv6 route exchange. If BGP-v6 session exists, it continues to operate independently from BGP-v4.
Multiprotocol BGPv6 supports many of the same features and functionality as BGPv4. IPv6 enhancements to MBGP include
support for an IPv6 address family and Network Layer Reachability Information (NLRI) and next hop attributes that use the IPv6
addresses.
Attributes
Routes learned using BGP have associated properties that are used to determine the best route to a destination when multiple
paths exist to a particular destination. These properties are called BGP attributes which influence route selection for designing
robust networks. There are no hard coded limits on the number of supported BGP attributes.
BGP attributes for route selection:
● Weight
● Local preference
● Multiexit discriminators
● Origin
● AS path
● Next-hop
Communities
BGP communities are sets of routes with one or more common attributes. Communities assign common attributes to multiple
routes simultaneously. Duplicate communities are not rejected.
Disable announcement of ASN values
Modify the AS_PATH attribute of the received routes.
● Disable prepending the local AS number in CONFIG-ROUTER-NEIGHBOR mode.
local-as as-number no-prepend
● Disable prepending the globally-configured AS number in CONFIG-ROUTER-NEIGHBOR mode.
local-as as-number no-prepend replace-as
Selection criteria
Best path selection criteria for BGP attributes:
1. Prefer the path with the largest WEIGHT attribute, and prefer the path with the largest LOCAL_PREF attribute.
2. Prefer the path that is locally originated using the network command, redistribute command, or aggregate-
address command. Routes originated using a network or redistribute command are preferred over routes that
originate with the aggregate-address command.
3. (Optional) If you configure the bgp bestpath as-path ignore command, skip this step because AS_PATH is not
considered. Prefer the path with the shortest AS_PATH:
Layer 3
575