Information Technology Reference
In-Depth Information
BGP Address Family Configuration
BGP AF-style configuration is based on the concept of defining all the peering relationships
or neighbors in the main BGP router configuration. These neighbors are then activated
under the address family for each NLRI type that will be carried in this peering session. The
IPv4 NLRI is on by default for all BGP sessions. You can disable the default behavior of
carrying IPv4 NLRI on all BGP sessions using the
no bgp default ipv4-unicast
command.
Example 12-1 shows a pure IPv6 deployment.
Example 12-1
AF-Style Configuration for IPv6
router bgp 65000
no synchronization
bgp router-id 10.1.1.5
no bgp default ipv4-unicast
bgp log-neighbor-changes
neighbor 2001:400:0:1234::1 remote-as 65000
neighbor 2001:400:0:1234::1 update-source Loopback1
neighbor 2001:400:0:1234::2 remote-as 65000
neighbor 2001:400:0:1234::2 update-source Loopback1
no auto-summary
!
address-family ipv6
no synchronization
neighbor 2001:400:0:1234::1 activate
neighbor 2001:400:0:1234::2 activate
network 2001:400:0:ABCD::/64
exit-address-family
!
Injecting IPv6 Prefixes into BGP
The process of injecting IPv6 prefix information into BGP is the same as for IPv4; however,
it must be done under the IPv6 AF configuration. You can redistribute prefix information
from another routing protocol or inject prefixes from the routing table using the
network
command.
Prefix Filtering for IPv6
The two primary methods of matching and filtering prefix information are access control
lists (ACLs) and prefix lists.
ACLs are the most common form of prefix or packet filtering. When building ACLs for
IPv6 prefix information, you must use named ACLs, because numbered ACLs are not
supported. The initial implementation of ACLs for IPv6 supported only matching against
source and destination addresses. However, in Cisco IOS Releases 12.2(13)T and 12.0(23)S,
support was added to match against additional information. Table 12-3 shows the fields that
