BGP Address Family Configuration¶
The TNSR BGP implementation is capable of handling routing information for IPv4
and IPv6 independently, among other network layer protocols. The
address-family <family> <type>
command defines BGP behavior for each
specific supported case. The most common address families are ipv4 unicast
and ipv6 unicast
. The other possible choices supported in this version are
ipv4 multicast
and ipv6 multicast
.
The address-family
command changes to BGP address family mode, which
contains settings specific to each address family. The prefix for this mode
varies depending on the address family command which entered the mode. For
example, when configuring settings for the IPv4 unicast address family, the
prompt indicates config-bgp-ip4uni
.
tnsr(config-bgp)# address-family ipv4 unicast
tnsr(config-bgp-ip4uni)#
Each resulting mode, such as config-bgp-ip4uni
or config-bgp-ip6uni
,
contains its own set of commands. As these may differ, they are split up in
multiple sections here.
IPv4 or IPv6 Unicast¶
The following commands are available in config-bgp-ip4uni
and
config-bgp-ip6uni
modes:
- aggregate-address <ip-prefix> [as-set] [summary-only]:
This command configures route aggregation using the specified prefix. More specific routes contained within the specified prefix will be aggregated into the larger prefix, minimizing the set of networks advertised to peers.
- as-set:
When present, routes for the specified prefix will include an AS set. An AS set is a collection of AS numbers for which routes have been aggregated. This allows peers to detect routing loops, duplicate routes, and so on.
- summary-only:
When present, aggregated routes for this prefix will not be announced, so peers only see the aggregate prefix and not the component networks.
- distance external <extern> internal <intern> local <local>:
Configures distance values which control how BGP will treat routes based on the length of their AS path.
- external <extern>:
The distance at which routes are considered external, from
1-255
.- internal <intern>:
The distance at which routes are considered internal, from
1-255
.- local <local>:
The distance at which routes are considered local, from
1-255
.
- distance administrative <dist> prefix <ip-prefix> [access-list <access-list-name>]:
This command manually configures the administrative distance for a given prefix, with the following required parameters:
- administrative <dist>:
The administrative distance for this prefix, from
1-255
.- prefix <ip-prefix>:
The IP prefix to which this distance will be applied.
- access-list <access-list-name>:
An access list which can be used to apply the distance to only a subset of the configured prefix.
- import vrf <route-table-name>:
Import (“leak”) routes from a different VRF with the given name.
Note
The other VRF must be configured in BGP or it will not be available for importing. For example, it must be defined in BGP with
server vrf <name>
, have an ASN, and be set to redistribute the desired routes.- import vrf route-map <route-map-name>:
Specify a route map used to modify or filter routes imported from another VRF.
- maximum-paths <non-ibgp-paths> [igbp <ibgp-paths> [equal-cluster-length]]:
Configures the maximum number of paths for multi-path eBGP forwarding. This is enabled by default with a value of
64
. This allows the router to utilize multiple equal identical paths via different routers.Paths for prefixes advertised by multiple eBGP peers in the same AS are considered equal cost and result in a multi-path route.
Note
As this feature is enabled by default, to disable this behavior, set the value to
1
which limits routes to only a single path.- igbp <ibgp-paths>:
Configures a value for multi-path forwarding in iBGP roles.
- equal-cluster-length:
Only consider paths as matching when cluster lengths are also equal.
- neighbor <existing-neighbor>:
Specifies an existing neighbor address or peer group to use with this address family, and enters BGP Address Family Neighbor Configuration mode.
Warning
This command cannot define a new neighbor. A neighbor or peer group must first be defined using the
neighbor
command from withinconfig-bgp
mode before it can be used here.- network <ip-prefix> [route-map <route-map>]:
Configures a prefix to be advertised to peers in this address family.
- route-map <route-map>:
Specifies a route map used to limit advertisements of this prefix.
- redistribute <route-source> [metric <val>|route-map <route-map-name>]:
Enables redistribution of routes from another source. Available route sources are listed in Dynamic Routing Protocol Lists.
- metric <val>:
A MED value to apply to redistributed routes.
- route-map <route-map-name>:
Specifies a route map used to limit redistributed route advertisements.
- redistribute ospf [metric <val>|route-map <route-map-name>]:
Configure redistribution of routes from OSPF.
- metric <val>:
A MED value to apply to redistributed routes.
- route-map <route-map-name>:
Specifies a route map used to limit redistributed route advertisements.
- table-map <route-map-name>:
Uses the specified route map to control how routes received from BGP peers are passed to the dynamic routing manager process, and thus, into routing tables.
IPv4 Multicast¶
The following commands are available in config-bgp-ip4multi
mode. See
IPv4 or IPv6 Unicast for descriptions of the commands and
parameters:
aggregate-address
distance external
distance administrative
neighbor
network
table-map
IPv6 Multicast¶
The following commands are available in config-bgp-ip6multi
mode See
IPv4 or IPv6 Unicast for descriptions of the commands and
parameters:
distance external
distance administrative
neighbor
network