Resolution of NLRI from (directly connected) eBGP Speakers

For IPv4 or IPv6 unicast NLRI received from eBGP, directly connected BGP sessions are resolved by only using connected routes, or system-connected, in the parlance of this feature. This feature does not change this behavior, nor will configuration of a non-default resolution profile affect this behavior.

Address Family Profile Restrictions

_{* Tunnel domains refer to tunnel RIBs, e.g. system-colored-tunnel-rib, system-tunnel-rib, or user-defined tunnel RIBs.}

_{+ IP domains are either of system-unicast-rib or system-connected.}

EOS Release 4.x enhancements:

EOS Release 4.22.0 Enhancements:

EOS Release 4.22.1 Enhancements:

Added support for hitless rule updates. This enhancement ensures that persistent filtering rules remain active while other filtering rules update, for example, a BGP Peer publishes or withdraws rules.

EOS Release 4.23.1 Enhancements:

EOS Release 4.23.2 Enhancements:

EOS Release 4.24.0 Enhancements:

Added support for configuring BGP Flowspec on subinterfaces. To enable subinterface support, the TCAM profile of the flow-spec feature must include port qualifier size 3 bits (see Flowspec TCAM Profile and Flowspec Policer TCAM Profile below).

EOS Release 4.24.1 Enhancements:

EOS Release 4.24.2 Enhancements:

Removed EOS Release 4.23.1 limitation to best effort programming.

EOS Release 4.25.2 Enhancements:

• Added support for BGP Flowspec applied to SVI.
• BGP Flowspec releases TCAM banks when no longer needed to store matches. Previously, once TCAM banks allocated to BGP Flowspec, they never released.

EOS Release 4.30.0F Enhancements

• Added support for below and above police counters.

EOS Release 4.30.1F Enhancements

• Added support for the source VRF override the configuration and apply flowspec rules received in a different VRF.

EOS Release 4.30.2F Enhancements

• Added support for displaying per-match byte counters in addition to packet counters in the output of show flow-spec.

EOS Release 4.31.1F Enhancements

• Added support tor traffic-rate-packets policing action from RFC8955.

EOS Release 4.32.0F Enhancements

• Added support for Flowspec counter telemetry.

BGP BGP FlowSpec Functionality

flow-spec
counters poll interval 30-300 seconds

BGP Limitations

Platform Limitations

Scaling Limits

Resolution of NLRI from (directly connected) eBGP Speakers

For IPv4 or IPv6 unicast NLRI received from eBGP, directly connected BGP sessions are resolved by only using connected routes, or system-connected, in the parlance of this feature. This feature does not change this behavior, nor will configuration of a non-default resolution profile affect this behavior.

Address Family Profile Restrictions

_{* Tunnel domains refer to tunnel RIBs, e.g. system-colored-tunnel-rib, system-tunnel-rib, or user-defined tunnel RIBs.}

_{+ IP domains are either of system-unicast-rib or system-connected.}

Example

This command places the switch in router-BGP configuration mode. It also creates a BGP instance in AS 50 if an instance was not previously created.

switch(config)# router bgp 50
switch(config-router-bgp)#

When a BGP instance exists, the router bgp command must include its autonomous system. Any attempt to create a second instance results in an error message.

Example

This command attempts to open a BGP instance with a different AS number from that of the existing instance. The switch displays an error and stays in global configuration mode.

switch(config)# router bgp 100
% BGP is already running with AS number 50
switch(config)#

Examples

• These commands place the switch in BGP VRF configuration mode for VRF purple. Commands issued in this mode override the global BGP configuration for the specified VRF instance.

  switch(config)# router bgp 1
  switch(config-router-bgp)# vrf purple
  switch(config-router-bgp)#

• These commands activate IPv6 address-family support for the IPv6 neighbor 2001:0DB8:8c01::1 in VRF purple.

  switch(config-router-bgp-vrf-purple)# router-id 1.1.1.1
  switch(config-router-bgp-vrf-purple)# neighbor 2001:0DB8:8c01::1 remote-as 16
  switch(config-router-bgp-vrf-purple)# address-family ipv6
  switch(config-router-bgp-vrf-purple-af)# neighbor 2001:0DB8:8c01::1 activate
  switch(config-router-bgp-vrf-purple-af)#

• This command configures a route distinguisher for VRF purple.

  switch(config-router-bgp-vrf-purple)# rd 530:12
  switch(config-router-bgp-vrf-purple)#

Examples

• These commands configure the switch in RCF functions for IPv4 application.

  switch(config)# router bgp 64500
  switch(config-router-bgp)# address-family ipv4
  switch(config-router-bgp-af)# neighbor 192.168.0.1 rcf in INBOUND_POLICY()
  switch(config-router-bgp-af)# neighbor 192.168.0.1 rcf out OUTBOUND_POLICY()

• These commands configure the switch in RCF functions for IPv6 unicast application.

  switch(config)# router bgp 64500
  switch(config-router-bgp)# address-family ipv6 labeled-unicast
  switch(config-router-bgp-af-label)# neighbor 192.168.0.1 rcf in LU_INBOUND_POLICY()
  switch(config-router-bgp-af-label)# neighbor 192.168.0.1 rcf out LU_OUTBOUND_POLICY()

• These commands configure RCF function with the redistribute configuration statement for connected and static routes.

  switch(config)# router bgp 64500
  switch(config-router-bgp)# redistribute connected rcf CONNECTED_POLICY()
  switch(config-router-bgp)# redistribute static rcf STATIC_POLICY()

• These commands configure RCF function on routes redistributed into BGP from IS-IS. Level 1, level 2, or both IS-IS level routes can be specified for RCF application.

  switch(config)# router bgp 64500
  switch(config-router-bgp)# redistribute isis level-1 rcf ISIS_LEVEL_1_POLICY()
  switch(config-router-bgp)# redistribute isis level-2 rcf ISIS_LEVEL_2_POLICY()
  switch(config-router-bgp)# redistribute isis level-1-2 rcf ISIS_LEVEL_1_2_POLICY()

Static BGP Neighbors

The neighbor remote-as command connects the switch with a peer, establishing a static neighbor.

Once established, a static neighbor may be added to an existing peer group. Any configuration applied to the peer group then is inherited by the neighbor, unless a conflicting configuration has been entered for that peer. Settings applied to a member of the peer group override group settings.

Examples

• These commands establish an internal BGP connection with the peer at 10.1.1.14.

  switch(config)# router bgp 50
  switch(config-router-bgp)# neighbor 10.1.1.14 remote-as 50
  switch(config-router-bgp)#

• These commands establish an external BGP connection with the peer at 192.168.2.5.

  switch(config)# router bgp 50
  switch(config-router-bgp)# neighbor 192.168.2.5 remote-as 100
  switch(config-router-bgp)#

Dynamic BGP Neighbors

The bgp listen range command specifies a range of IPv4 addresses from which the switch will accept incoming dynamic BGP peering requests, and creates the named dynamic peer group to which those peers belong. Dynamic BGP neighbors are peers which have not been manually established, but are accepted into a dynamic peer group when the switch receives a peering request from them.

Dynamic peers cannot be configured individually, but inherit any configuration that is applied to the peer group to which they belong. Peering relationships with dynamic peers are terminated if the peer group is deleted.

Example

switch(config)# router bgp 50
switch(config-router-bgp)# bgp listen range 192.168.2.0/24 peer-group brazil remote-as 50
switch(config-router-bgp)#

Displaying Neighbor Connections

The show ip bgp summary and show ip bgp neighbors commands display neighbor connection status.

Example

switch# show ip bgp summary
BGP summary information for VRF default
BGP router identifier 192.168.104.2, local AS number 50
Neighbor Status Codes: m - Under maintenance
  Neighbor     V  AS  MsgRcvd  MsgSent  InQ OutQ  Up/Down State  PfxRcd PfxAcc
  192.168.2.5  4  100     198      281    0    0 03:11:31 Estab  12     12
switch#

Static BGP Peer Groups

A static BGP peer group is a collection of BGP neighbors which can be configured as a group. Once a static peer group is created, the group name can be used as a parameter in neighbor configuration commands, and the configuration will be applied to all members of the group. Neighbors added to the group will inherit any settings already created for the group. Static peer group members may also be configured individually, and the settings of an individual neighbor in the peer group override group settings for that neighbor.

When the default form of a BGP configuration command is entered for a member of a static peer group, the peer inherits that configuration from the peer group.

A static peer group is created with the neighbor peer group (create) command, or by using the bgp listen range command to accept dynamic peering requests. Once a static peer group has been created, static neighbors can be manually added to the group by using the neighbor peer group (neighbor assignment) command. The neighbor peer group (neighbor assignment) command removes a neighbor from a static peer group.

The neighbor peer group (create) command will delete a static peer group. When a peer group is deleted, the members of that group revert to their individual configurations, or to the system default for any attributes that have not been specifically configured for that peer.

Examples

• These commands create a peer group named akron.

  switch(config)# router bgp 50
  switch(config-router-bgp)# neighbor akron peer group
  switch(config-router-bgp)#

• This command adds the neighbors at 1.1.1.1 and 2.2.2.2 to peer group akron.

  switch(config-router-bgp)# neighbor 1.1.1.1 peer group akron
  switch(config-router-bgp)# neighbor 2.2.2.2 peer group akron
  switch(config-router-bgp)#

• These commands configure the members of peer group akron, but cause the neighbor at 1.1.1.1 to use the system default value for out-delay.

  switch(config-router-bgp)# neighbor akron remote-as 109
  switch(config-router-bgp)# neighbor akron out-delay 101
  switch(config-router-bgp)# neighbor akron maximum-routes 12000
  switch(config-router-bgp)# no neighbor 1.1.1.1 out-delay
  switch(config-router-bgp)#

Dynamic BGP Peer Groups

A dynamic BGP peer group is a collection of BGP neighbors in a specified address range which makes a peer request to the switch. Members of dynamic peer group are configured in groups and not as individuals. A dynamic peer group name is used as a parameter to apply the configuration across all the members in the group. Neighbors joining the group inherit any settings already created for the group.

The bgp listen range command is used to create a dynamic peer group. This command identifies the BGP peering request from a range of IP address, and names the dynamic peer group to which those peers belong to. The bgp listen range command can be configured to accept a peering request from a single AS number or to accept peer request from the range of AS numbers. To accept the request from the range of AS numbers use the peer filter option in the command as shown. If the peer filter referred by the bgp listen range command does not exist, or if the filter exists but has no match commands, it will accept any AS number.

To delete a dynamic peer group, use the no or default form of the bgp listen range command. All peering relationships with group members are terminated when the dynamic peer group is deleted.

Examples

• These commands create a dynamic peer group called brazil in a single AS, which accepts peering requests from the 192.0.2.0/24 subnet the single AS is 5.

  switch(config)# router bgp 1
  switch(config-router-bgp)# bgp listen range 192.0.2.0/24 peer-group brazil remote-as 5
  switch(config-router-bgp)#

• These commands create a dynamic peer group called brazil in a range of ASNs, which accepts peering requests from the 192.0.2.0/24 subnet. The range of AS numbers is defined by peer filter option.

  switch(config)# router bgp 1
  switch(config-router-bgp)# bgp listen range 192.0.2.0/24 peer-group brazil peer-filter group-1
  switch(config-router-bgp)#

• The show ip bgp peer-group command displays the source of a listen range’s remote AS number definition as shown.

  switch(config-router-bgp)# show ip bgp peer-group
  BGP peer-group is brazil
   BGP version 4
   Listen-range subnets:
  VRF default:
  192.0.2.0/24, remote AS 5
  192.0.2.0/24, peer filter group1
  switch(config-router-bgp)#

Examples

• These commands define a peer filter that accepts any AS number.

  switch(config)# peer-filter group1
  switch(config-peer-filter-group1)# 10 match as-range 1-4294967295 result accept
  switch(config-peer-filter-group1)#

• These commands define a peer filter that accepts any AS number between 65000 and 65100 (inclusive) except 65008 and 65009.

  switch(config)# peer-filter group2
  switch(config-peer-filter-group2)# 10 match as-range 65008-65009 result reject
  switch(config-peer-filter-group2)# 20 match as-range 65000-651000 result accept
  switch(config-peer-filter-group2)#

• These commands define a peer filter that accepts 3 specific remote AS numbers.

  switch(config)# peer-filter group3
  switch(config-peer-filter-group3)# 10 match as-range 65003 result accept
  switch(config-peer-filter-group3)# 20 match as-range 65007 result accept
  switch(config-peer-filter-group3)# 30 match as-range 65009 result accept
  switch(config-peer-filter-group3)#

• The show peer-filter command displays the peer filter definition.

  switch(config)# show ip bgp peer-group3
  peer-filter group3
     10 match as-range 65003 result accept
     20 match as-range 65007 result accept
  30 match as-range 65009 result accept
  switch(config)#

Activating IPv6 Neighbors

By default, the switch does not negotiate or advertise IPv6 BGP routes. In order to establish a session with an IPv6 neighbor, it must be made active in the IPv6 address family. The ipv6-unicast option of the bgp default command causes the switch to send IPv6 capability messages and all network advertisements with IPv6 prefixes to all BGP neighbors. The neighbor activate command issued in IPv6 address family configuration mode does the same for a single BGP neighbor.

Examples

• These commands make all BGP neighbors active in the IPv6 address family.

  switch(config)# router bgp 11
  switch(config)# address-family ipv6
  switch(config-router-bgp-af)# bgp default ipv6-unicast
  switch(config-router-bgp-af)# exit
  switch(config-router-bgp)#

• These commands make the BGP neighbor at 2001:0DB8:8c01::1 active in the IPv6 address family.

  switch(config)# router bgp 11
  switch(config)# address-family ipv6
  switch(config-router-bgp-af)# neighbor 2001:0DB8:8c01::1 activate
  switch(config-router-bgp-af)# exit
  switch(config-router-bgp)#

Sending IPv4 NLRIs over IPv6 Connections

The switch supports the exchange of IPv4 NLRIs with IPv6 neighbors. To enable this feature for all IPv6 neighbors, use the ipv4-unicast transport ipv6 option of the bgp default command in the IPv4 address family configuration mode. To enable it for a single IPv6 neighbor, use the neighbor activate command for that neighbor in the IPv4 address family configuration mode.

To send IPv4 NLRIs to IPv6 neighbors, the IPv4 next-hop address must also be communicated. To explicitly configure an IPv4 next hop to send to a specific IPv6 neighbor, use the neighbor local-v4-addr command. In some network configurations, the switch can also be configured to automatically determine the best IPv4 next-hop address for an individual IPv6 neighbor or for all neighbors in the VRF using the neighbor auto-local-addr command.

Examples

• These commands permit IPv4 NLRI transport over all IPv6 connections by making the IPv4 address family active on IPv6 BGP neighbors, then configure the switch to automatically select a local IPv4 address to be sent in NLRIs to the IPv6 neighbors in a peer group called indianapolis.

  switch(config)# router bgp 11
  switch(config-router-bgp)# address-family ipv4
  switch(config-router-bgp-af)# bgp default ipv4-unicast transport ipv6
  switch(config-router-bgp-af)# exit
  switch(config-router-bgp)# neighbor indianapolis auto-local-addr
  switch(config-router-bgp)#

• These commands permit IPv4 NLRI transport with the IPv6 neighbor at 2001:0DB8:8c01::1 using a local IPv4 address of 10.7.5.11.

  switch(config)# router bgp 11
  switch(config-router-bgp)# address-family ipv4
  switch(config-router-bgp-af)# neighbor 2001:0DB8:8c01::1 activate
  switch(config-router-bgp-af)# exit
  switch(config-router-bgp)# neighbor 2001:0DB8:8c01::1 local-v4-addr 10.7.5.11
  switch(config-router-bgp)#

Example

switch(config-router-bgp)# timers bgp 15 45
switch(config-router-bgp)#

The show ip bgp neighbors command displays the hold time.

Example

switch# show ip bgp neighbors 10.100.100.2
BGP neighbor is 10.100.100.2, remote AS 100
BGP version 4, remote router ID 192.168.100.13, VRF default
  Negotiated BGP version 4
  Last read 00:00:05, last write 00:00:05
  Hold time is 45, keepalive interval is 15 seconds         <= hold time
  Configured hold time is 45, keepalive interval is 15 seconds
  Connect timer is inactive
  Idle-restart timer is inactive
  BGP state is Established, up for 04:44:05
  Number of transitions to established: 11
  Last state was OpenConfirm
  Last event was RecvKeepAlive
  Last sent notification:Cease/administrative reset, Last time 04:44:09
  Last rcvd notification:Cease/peer de-configured, Last time    2d02h, First time 7d08h, Repeats 1
  Neighbor Capabilities:
  Multiprotocol IPv4 Unicast: advertised and received and negotiated
  Four Octet ASN: advertised and received
             <-------OUTPUT OMITTED FROM EXAMPLE------->
switch#

Maximum Routes

The neighbor maximum-routes command determines the number of BGP routes the switch accepts from a specified neighbor. The switch disables peering with the neighbor when this number is exceeded.

Example

switch(config-router-bgp)# neighbor 192.168.18.24 maximum-routes 15000
switch(config-router-bgp)#

Route Reflection

Participating BGP routers within an AS communicate eBGP-learned routes to all of their peers. They do not re-advertise iBGP-learned routes within the AS to prevent routing loops. Although a fully meshed network topology ensures that all AS members share routing information, this topology can result in high volumes of iBGP messages when scaled. Alternatively, one or more routers can be configured as route reflectors in larger networks.

A route reflector re-advertises routes learned through iBGP to a group of BGP neighbors within the AS, replacing the function of a fully meshed topology. The neighbor route-reflector-client command configures the switch to act as a route reflector and configures the specified neighbor as a client. The bgp client-to-client reflection command enables client-to-client reflection.

Cluster IDs When using route reflectors, an AS is divided into clusters. A cluster contains at least one route reflector and a group of clients to which they re-advertise route information. A cluster may contain multiple route reflectors to provide redundancy protection. Each reflector has a cluster ID. When the cluster has a single route reflector, the cluster ID is its router ID. When a cluster has multiple route reflectors, a 4-byte cluster ID is assigned to all route reflectors in the cluster, allowing them to recognize updates from other cluster reflectors. The command configures the cluster ID in a cluster with multiple route reflectors.

Attribute Preservation Outbound BGP policies can rewrite the BGP attributes (next-hop, local preference and metric) of routes advertised by a route reflector. To configure the route reflector to preserve these attributes regardless of policy (unless those policies are included in a route map), use the bgp route-reflector preserve-attributes command. To include route attributes at all times (even contrary to policies included in route maps), use the always option of the command.

Client-to-client Reflection Usually the clients of a route reflector are not interconnected, and any routes learned by a client are mirrored to other clients and re-advertised within the AS by the route reflector. If the clients of a route reflector are fully meshed, routes received from a client do not need to be mirrored to other clients. In this case, client-to-client reflection should be disabled using the bgp client-to-client reflection command.

Route Preference

The primary function of external peers is to distribute routes they learn from their peers. Internal peers receive route updates without distributing them. External peers receive route updates, then distribute them to internal and external peers.

Local preference is a metric that iBGP sessions use to select an external route. Preferred routes have the highest local preference value. UPDATE packets include this metric in the LOCAL_PREF field.

The neighbor export-localpref command specifies the LOCAL_PREF that the switch sends to an internal peer. The command overrides previously assigned preferences and has no effect on external peers.

Example

switch(config-router-bgp)# neighbor 10.1.1.45 export-localpref 200
switch(config-router-bgp)#

The neighbor import-localpref command assigns a local preference to routes received through UPDATE packets from an external peer. This command has no effect when the neighbor is an internal peer.

Graceful Restart

Graceful BGP restart allows a BGP speaker with separate control plane and data plane processing to continue forwarding traffic during a BGP restart. Its neighbors (receiving speakers) may retain routing information from the restarting speaker while a BGP session with it is being re-established, reducing route flapping.

Arista switches can act as helpers (receiving speakers) for graceful BGP restart with neighbors that advertise graceful restart capability.

Graceful restart helper mode is enabled by default, but can be turned off globally with the graceful-restart-helper command. Per-peer configuration takes precedence over the global configuration.

Peers with graceful restart capability advertise a restart time value as an estimate of the time it will take them to restart a BGP session. When a BGP session with a restarting speaker goes down, the switch (receiving speaker) marks routes from that peer as stale and starts the restart timer. If the session with the peer is not re-established before the restart time runs out, the switch deletes the stale routes from that peer. If the session is re-established within that time, the stale path timer is started. If the stale paths are not updated by the restarting speaker before the stale path time runs out, they are deleted. The maximum time these stale paths are retained after the BGP session is re-established is 300 seconds by default, but can be configured using the graceful-restart stalepath-time command.

Example

switch(config-router-bgp)# graceful-restart stalepath-time 500
switch(config-router-bgp)#

Filtering with Route Maps

BGP uses route maps to directly filter IPv4 unicast routes. The neighbor route-map (BGP) command applies a route map to inbound or outbound BGP routes. To display the route maps associated with a specific BGP neighbor, use the show ip bgp neighbors command.

The redistribution of BGP unicast routes into multicast address families allows the network to take a different path for the multicast traffic. It allows redistribution of IPv4 unicast routes into the IPv4 multicast address family and IPv6 unicast routes into the IPv6 multicast address family.

The following command configures the redistribution of IPv4 unicast routes into IPv4 multicast address family in both default and non-default VRF.

switch(config-router-bgp)# address-family ipv4 multicast
switch(config-router-bgp-af)# route input address-family ipv4 unicast rcf onePfx()

bgprtr1(config-router-multicast)# show bgp ipv4 unicast
BGP routing table information for VRF default
Router identifier 1.1.1.1, local AS number 100
Route status codes: s - suppressed, * - valid, > - active, E - ECMP head, e - ECMP
                    S - Stale, c - Contributing to ECMP, b - backup, L - labeled-unicast
                    % - Pending BGP convergence
Origin codes: i - IGP, e - EGP, ? - incomplete
RPKI Origin Validation codes: V - valid, I - invalid, U - unknown
AS Path Attributes: Or-ID - Originator ID, C-LST - Cluster List, LL Nexthop - Link Local Nexthop

          Network                Next Hop              Metric  AIGP       LocPref Weight  Path
 * >      10.10.10.1/32          1.1.1.2               0       -          100     0       200 i
 * >      10.10.20.1/32          1.1.1.2               0       -          100     0       200 i

The following command shows BGP IPv4 multicast output, when a RCF function filters 10.10.20.1/32.

bgprtr1# show bgp ipv4 multicast
BGP routing table information for VRF default
Router identifier 1.1.1.1, local AS number 100
Route status codes: s - suppressed, * - valid, > - active, E - ECMP head, e - ECMP
                    S - Stale, c - Contributing to ECMP, b - backup, L - labeled-unicast
                    % - Pending BGP convergence
Origin codes: i - IGP, e - EGP, ? - incomplete
AS Path Attributes: Or-ID - Originator ID, C-LST - Cluster List, LL Nexthop - Link Local Nexthop

          Network                Next Hop              Metric  AIGP       LocPref Weight  Path
 * >      10.10.20.1/32          1.1.1.2               -       -          -       0       ?

Filtering with BGP Communities

Example

• These commands assign two network subnets to a prefix list, assign a community number to the prefix list members, then utilize that community in an ip community-list command to permit the routes into the BGP domain.
  1. Compose the IP prefix list.

     switch(config)# ip prefix-list PL_1 permit 10.1.2.5/24
     switch(config)# ip prefix-list PL_1 permit 10.2.5.1/28
     switch(config)#

  2. Create a route map that matches the IP prefix list and sets the community value.

     switch(config)# route-map MAP_1 permit
     switch(config-route-map-MAP_1)# match ip address prefix-list PL_1
     switch(config-route-map-MAP_1)# set community 500
     switch(config-route-map-MAP_1)# exit
     switch(config)#

  3. Create a community list that references the community.

     switch(config)# ip community-list CL_1 permit 500
     switch(config)#

BGP extended communities identify routes for VRFs or for Link BandWidth (LBW). Extended community clauses utilize Route Target (RTt) and Site of Origin Options (SOO):

• route targets identify sites that may receive appropriately tagged routes.
• site of origin identifies the site where the router learned the route.

Filtering with AS Path Access Lists

An AS path access list is a named list of permit and deny statements which use regular expressions to filter BGP routes based on their AS path attribute. AS path access lists are created using the ip as-path access-list command, and are applied using a route map match clause with the name of the access list as a parameter.

Example

Aggregation Options

The aggregate-address command provides the following aggregate route options:

• AS_PATH attribute inclusion: the as-set option controls the aggregate route’s AS_PATH and ATOMIC_AGGREGATE attribute contents. AS_PATH identifies the autonomous systems through which UPDATE message routing information passes. ATOMIC_AGGREGATE indicates that the route is an aggregate or summary of more specific routes.

  When the command includes as-set, the aggregate route’s AS_SET attribute contains the AS numbers of contributor routes. This can help BGP neighbors to prevent loops by rejecting aggregate routes that include their AS number in the AS_SET.

  When the command does not include as-set, the aggregate route’s ATOMIC_AGGREGATE attribute is set and the AS_PATH attribute does not include AS numbers of contributing routes.

• Attribute assignment: the attribute-map option assigns attributes contained in set commands in a specified route map’s lowest sequence with any set command to the aggregated route, overriding the automatic determination of the aggregate route’s attributes by the switch.
• Route suppression: the summary-only option suppresses the advertisement of the contributor routes that comprise the aggregate.
• Contributor filtering: the match-map option uses a route map to filter out contributor routes that would otherwise be included in the aggregate.

Example

• These commands create an aggregate route (10.16.48.0/20) from four contributor routes (10.16.48.0/23, 10.16.50.0/23, 10.16.52.0/23, and 10.16.54.0/23). The aggregate route includes the AS_PATH information from the contributor routes.

  switch(config)# router bgp 1
  switch(config-router-bgp)# aggregate-address 10.16.48.0/20 as-set
  switch(config-router-bgp)# exit
  switch(config)#

• These commands create an aggregate route and use a route map to add a local-preference attribute to the route.

  switch(config)# route-map map1 permit 10
  switch(config-route-map-map1)# set local-preference 40
  switch(config-route-map-map1)# exit
  switch(config)# router bgp 1
  switch(config-router-bgp)# aggregate-address 10.16.48.0/20 attribute-map map1
  switch(config-router-bgp)# exit
  switch(config)#

• These commands create an aggregate route and use a route map to allow only those contributors that match a specified prefix list included in the aggregate route.

  switch(config)# route-map matchmap permit 10
  switch(config-route-map-matchmap)# match ip address prefix-list agglist
  switch(config-route-map-matchmap)# exit
  switch(config)# router bgp 1
  switch(config-router-bgp)# aggregate-address 1.1.0.0/16 summary-only match-map matchmap
  switch(config-router-bgp)#

The match-map parameter in the aggregate-address command advertises the route(s) denied in match-map.

Identifying BGP Aggregate Contributors Match in Outbound Policy

When configured, this feature introduces the ability to match on the following parameters:

The attributes that are currently supported for matching on BGP aggregate contributors are community, local-preference, prefix, next-hop,route-type.

Match Contributors to Any Aggregate

switch(config-route-map-test)# match aggregate-role contributor

The match aggregate-role contributor clause only works with outbound policies.

Example

switch(config)# ip community-list BLUE permit 65536:100 
switch(config)# route-map OUTBOUND_POLICY permit 10
switch(config-route-map-OUTBOUND_POLICY)# match aggregate-role contributor 
switch(config-route-map-OUTBOUND_POLICY)# set community community-list BLUE
!
switch(config)# route-map OUTBOUND_POLICY permit 20
switch(config-route-map-OUTBOUND_POLICY)# description “Permit the routes rejected by seq10"
!
switch(config)# router bgp 65536
switch(config-router-bgp)# aggregate-address 203.0.113.0/24
switch(config-router-bgp)# neighbor 192.0.2.1 route-map OUTBOUND_POLICY out 
switch(config-router-bgp)#

Match Contributors to Specific Aggregates

To match contributors which contribute only to a BGP aggregate with specific attributes (say communities) and set attributes (say communities again) on said contributor, add an outbound policy with the clause:

switch(config-route-map-test)# match aggregate-role contributor aggregate-attributes MATCH_AGG_COLOR

switch(config)# route-map MATCH_AGG_COLOR
switch(config-route-map-MATCH_AGG_COLOR)# match community RED

Add an aggregate definition to explicitly set the desired attributes on the aggregate of interest:

switch(config)# route-map AGG_SET_COLOR
switch(config-route-map-AGG_SET_COLOR)# set community community-list RED
!
switch(config)# router bgp 65536
switch(config-router-bgp)# aggregate-address 203.0.113.0/24
switch(config-router-bgp)# aggregate-address 203.0.113.0/24 attribute-map AGG_SET_COLOR

The route map referenced by the match aggregate-role contributor aggregate-attributes clause discards all set operations.

Invert-result Support

This match clause supports the invert-result modifier. When applied, invert-result inverts the result of the match clause to which it is applied.

switch(config-route-map-test)# match invert-result aggregate-role contributor aggregate-attributes MATCH_AGG_COLOR

1. match all routes that are not contributors to any aggregate.
2. match all routes that are contributors to aggregates where the aggregate doesn’t match MATCH_AGG_COLOR (provided MATCH_AGG_COLOR is configured).
3. not match all routes that are contributors to aggregates where the aggregate matches MATCH_AGG_COLOR (provided MATCH_AGG_COLOR is configured).
4. not match all routes that are contributors to aggregates if MATCH_AGG_COLOR is not configured.

Related Command

match (route-map)

Example

These commands configure the switch to replace its local ASN in OPEN messages sent to the peer at 10.13.64.1 with ASN 64500, and configure the peer to expect that ASN in messages received from the switch.

Switch Configuration

switch(config)# router bgp 64497
switch(config-router-bgp)# neighbor 10.13.64.1 local-as 64500 no-prepend
switch(config-router-bgp)#

Peer Configuration

peer(config)# router bgp 64502
peer(config-router-bgp)# neighbor 10.4.3.10 remote-as 64500
peer(config-router-bgp)#

Replacing Remote ASN in Outbound Route Announcements

To replace a remote ASN with the local ASN in BGP route announcements sent to a specified router, use the neighbor as-path remote-as replace out command.

Example

switch(config)# router bgp 1
switch(config-router-bgp)# neighbor 192.168.2.15 as-path remote-as replace out
switch(config-router-bgp)#

Ignoring Local ASN in Incoming Route Announcements

To accept BGP routes that include the local ASN in their AS-path attribute, use the neighbor allowas-in command.

Example

switch(config)# router bgp 1
switch(config-router-bgp)# neighbor 192.168.1.30 allowas-in
switch(config-router-bgp)#

Default Neighbor Activation

Activating Individual Neighbor Addresses

The switch sends the following announcements to addresses that are active in an address family:

• IPv4 address family: IPv4 capability and all network routes with IPv4 prefixes.
• IPv6 address family: IPv6 capability and all network routes with IPv6 prefixes.

The neighbor route-map (BGP) command applies a route map to inbound or outbound BGP routes. In address-family mode, the route map is applied to routes corresponding to the configuration-mode address family. When a route map is applied to outbound routes, the switch advertises only routes matching at least one section of the route map. One outbound and one inbound route map can be applied to a neighbor for each address family. Applying a route map to a route replaces the previous corresponding route map assignment.

Network Route Advertising in Address Families

The network (BGP) command specifies a network for advertisement through UPDATE packets to BGP peers. The command is available in Router-BGP and Router-BGP-Address-Family configuration modes; the mode in which the command is issued does not affect the command’s execution.

• Commands with an IPv4 address are advertised to peers that are IPv4 address family-active.
• Commands with an IPv6 address are advertised to peers that are IPv6 address family-active.

Examples

• These commands instantiate BGP, configure three neighbors, and configure two network routes.

  The default activity level for IPv4 and IPv6 address families is set to the default; all neighbor addresses are IPv4 address family active and IPv6 address family not active. IPv4 capability and network routes with IPv4 prefixes are advertised to all neighbor IPv4 addresses.

  switch(config)# router bgp 9
  switch(config-router-bgp)# neighbor 172.21.14.8 remote-as 15
  switch(config-router-bgp)# neighbor 172.23.18.6 remote-as 16
  switch(config-router-bgp)# neighbor 2001:0DB8:8c01::1 remote-as 16
  switch(config-router-bgp)# network 172.18.23.9/24
  switch(config-router-bgp)# network 2001:0DB8:de29::/64
  switch(config-router-bgp)#

• These commands instantiate BGP on the switch, set IPv4 default activity level (not active), set IPv6 default activity level (active), and configure three neighbor addresses and two network route prefixes.

  IPv6 capability and network routes with IPv6 prefixes are advertised to all neighbor addresses.

  switch(config)# router bgp 10
  switch(config-router-bgp)# bgp default ipv6-unicast
  switch(config-router-bgp)# no bgp default ipv4-unicast
  switch(config-router-bgp)# neighbor 172.21.14.8 remote-as 15
  switch(config-router-bgp)# neighbor 172.23.18.6 remote-as 16
  switch(config-router-bgp)# neighbor 2001:0DB8:8c01::1 remote-as 16
  switch(config-router-bgp)# network 172.18.23.9/24
  switch(config-router-bgp)# network 2001:0DB8:de29::/64
  switch(config-router-bgp)#

• These commands configure three neighbors, two network routes, and the default activity level for each address family (not active), and specify neighbor addresses for each address family that is active.

  switch(config)# router bgp 11
  switch(config-router-bgp)# neighbor 172.21.14.8 remote-as 15
  switch(config-router-bgp)# neighbor 172.23.18.6 remote-as 16
  switch(config-router-bgp)# neighbor 2001:0DB8:8c01::1 remote-as 16
  switch(config-router-bgp)# network 172.18.23.9/24
  switch(config-router-bgp)# network 2001:0DB8:de29::/64
  switch(config-router-bgp)# no bgp default ipv4-unicast
  switch(config-router-bgp)# no bgp default ipv6-unicast
  switch(config-router-bgp)# address-family ipv4
  switch(config-router-bgp-af)# neighbor 172.21.14.8 activate
  switch(config-router-bgp-af)# neighbor 172.23.18.6 activate
  switch(config-router-bgp-af)# exit
  switch(config-router-bgp)# address-family ipv6
  switch(config-router-bgp-af)# neighbor 2001:0DB8:8c01::1 activate
  switch(config-router-bgp-af)# exit
  switch(config-router-bgp)#

• These commands permit IPv4 NLRI transport over all IPv6 connections by making the IPv4 address family active on IPv6 BGP neighbors.

  switch(config)# router bgp 11
  switch(config)# address-family ipv4
  switch(config-router-bgp-af)# bgp default ipv4-unicast transport ipv6
  switch(config-router-bgp-af)# exit
  switch(config-router-bgp)#