Route Maps

Route maps provide a means to both filter and/or apply actions to route, hence allowing policy to be applied to routes.

For a route reflector to apply a route-map to reflected routes, be sure to include bgp route-reflector allow-outbound-policy in router bgp mode.

Route maps are an ordered list of route map entries. Each entry may specify up to four distinct sets of clauses:

Matching Conditions
A route-map entry may, optionally, specify one or more conditions which must be matched if the entry is to be considered further, as governed by the Match Policy. If a route-map entry does not explicitly specify any matching conditions, then it always matches.
Set Actions
A route-map entry may, optionally, specify one or more Set Actions to set or modify attributes of the route.
Matching Policy

This specifies the policy implied if the Matching Conditions are met or not met, and which actions of the route-map are to be taken, if any. The two possibilities are:

  • permit: If the entry matches, then carry out the Set Actions. Then finish processing the route-map, permitting the route, unless an Exit Policy action indicates otherwise.
  • deny: If the entry matches, then finish processing the route-map and deny the route (return deny).

The Matching Policy is specified as part of the command which defines the ordered entry in the route-map. See below.

Call Action
Call to another route-map, after any Set Actions have been carried out. If the route-map called returns deny then processing of the route-map finishes and the route is denied, regardless of the Matching Policy or the Exit Policy. If the called route-map returns permit, then Matching Policy and Exit Policy govern further behaviour, as normal.
Exit Policy

An entry may, optionally, specify an alternative Exit Policy to take if the entry matched, rather than the normal policy of exiting the route-map and permitting the route. The two possibilities are:

  • next: Continue on with processing of the route-map entries.
  • goto N: Jump ahead to the first route-map entry whose order in the route-map is >= N. Jumping to a previous entry is not permitted.

The default action of a route-map, if no entries match, is to deny. I.e. a route-map essentially has as its last entry an empty deny entry, which matches all routes. To change this behaviour, one must specify an empty permit entry as the last entry in the route-map.

To summarise the above:

  Match No Match
Permit action cont
Deny deny cont
action
  • Apply set statements
  • If call is present, call given route-map. If that returns a deny, finish processing and return deny.
  • If Exit Policy is next, goto next route-map entry
  • If Exit Policy is goto, goto first entry whose order in the list is >= the given order.
  • Finish processing the route-map and permit the route.
deny
The route is denied by the route-map (return deny).
cont
goto next route-map entry
show route-map [WORD]

Display data about each daemons knowledge of individual route-maps. If WORD is supplied narrow choice to that particular route-map.

clear route-map counter [WORD]

Clear counters that are being stored about the route-map utilization so that subsuquent show commands will indicate since the last clear. If WORD is specified clear just that particular route-map’s counters.

Route Map Command

route-map ROUTE-MAP-NAME (permit|deny) ORDER

Configure the order’th entry in route-map-name with Match Policy of either permit or deny.

Route Map Match Command

match ip address ACCESS_LIST

Matches the specified access_list

match ip address prefix-list PREFIX_LIST

Matches the specified PREFIX_LIST

match ip address prefix-len 0-32

Matches the specified prefix-len. This is a Zebra specific command.

match ipv6 address ACCESS_LIST

Matches the specified access_list

match ipv6 address prefix-list PREFIX_LIST

Matches the specified PREFIX_LIST

match ipv6 address prefix-len 0-128

Matches the specified prefix-len. This is a Zebra specific command.

match ip next-hop address IPV4_ADDR

This is a BGP specific match command. Matches the specified ipv4_addr.

match ipv6 next-hop IPV6_ADDR

This is a BGP specific match command. Matches the specified ipv6_addr.

match as-path AS_PATH

Matches the specified as_path.

match metric METRIC

Matches the specified metric.

match tag TAG

Matches the specified tag value associated with the route. This tag value can be in the range of (1-4294967295).

match local-preference METRIC

Matches the specified local-preference.

match community COMMUNITY_LIST

Matches the specified community_list

match peer IPV4_ADDR

This is a BGP specific match command. Matches the peer ip address if the neighbor was specified in this manner.

match peer IPV6_ADDR

This is a BGP specific match command. Matches the peer ipv6 address if the neighbor was specified in this manner.

match peer INTERFACE_NAME

This is a BGP specific match command. Matches the peer interface name specified if the neighbor was specified in this manner.

match source-protocol PROTOCOL_NAME

This is a ZEBRA specific match command. Matches the originating protocol specified.

match source-instance NUMBER

This is a ZEBRA specific match command. The number is a range from (0-255). Matches the originating protocols instance specified.

Route Map Set Command

set tag TAG

Set a tag on the matched route. This tag value can be from (1-4294967295). Additionally if you have compiled with the --enable-realms configure option. Tag values from (1-255) are sent to the Linux kernel as a realm value. Then route policy can be applied. See the tc man page.

set ip next-hop IPV4_ADDRESS

Set the BGP nexthop address to the specified IPV4_ADDRESS. For both incoming and outgoing route-maps.

set ip next-hop peer-address

Set the BGP nexthop address to the address of the peer. For an incoming route-map this means the ip address of our peer is used. For an outgoing route-map this means the ip address of our self is used to establish the peering with our neighbor.

set ip next-hop unchanged

Set the route-map as unchanged. Pass the route-map through without changing it’s value.

set ipv6 next-hop peer-address

Set the BGP nexthop address to the address of the peer. For an incoming route-map this means the ipv6 address of our peer is used. For an outgoing route-map this means the ip address of our self is used to establish the peering with our neighbor.

set ipv6 next-hop prefer-global

For Incoming and Import Route-maps if we receive a v6 global and v6 LL address for the route, then prefer to use the global address as the nexthop.

set ipv6 next-hop global IPV6_ADDRESS

Set the next-hop to the specified IPV6_ADDRESS for both incoming and outgoing route-maps.

set local-preference LOCAL_PREF

Set the BGP local preference to local_pref.

set local-preference +LOCAL_PREF

Add the BGP local preference to an existing local_pref.

set local-preference -LOCAL_PREF

Subtract the BGP local preference from an existing local_pref.

[no] set distance DISTANCE

Set the Administrative distance to DISTANCE to use for the route. This is only locally significant and will not be dispersed to peers.

set weight WEIGHT

Set the route’s weight.

set metric METRIC

Set the BGP attribute MED.

set as-path prepend AS_PATH

Set the BGP AS path to prepend.

set community COMMUNITY

Set the BGP community attribute.

set ipv6 next-hop local IPV6_ADDRESS

Set the BGP-4+ link local IPv6 nexthop address.

set origin ORIGIN <egp|igp|incomplete>

Set BGP route origin.

set table (1-4294967295)

Set the BGP table to a given table identifier

Route Map Call Command

call NAME

Call route-map name. If it returns deny, deny the route and finish processing the route-map.

Route Map Exit Action Command

on-match next
continue

Proceed on to the next entry in the route-map.

on-match goto N
continue N

Proceed processing the route-map at the first entry whose order is >= N

Route Map Optimization Command

route-map optimization

Enable route-map processing optimization. The optimization is enabled by default. Instead of sequentially passing through all the route-map indexes until a match is found, the search for the best-match index will be based on a look-up in a prefix-tree. A per-route-map prefix-tree will be constructed for this purpose. The prefix-tree will compose of all the prefixes in all the prefix-lists that are included in the match rule of all the sequences of a route-map.

no route-map optimization

Disable the route-map processing optimization.

Route Map Examples

A simple example of a route-map:

route-map test permit 10
 match ip address 10
 set local-preference 200

This means that if a route matches ip access-list number 10 it’s local-preference value is set to 200.

See Miscellaneous Configuration Examples for examples of more sophisticated usage of route-maps, including of the call action.