Cisco IOS XR software uses a neighbor-based model for BGP configuration. All BGP settings are managed at the individual neighbor level.

Key aspects of BGP configuration in Cisco IOS XR include:

• Neighbor-specific configuration: All settings for a particular neighbor must be defined under the neighbor configuration. You must apply configuration statements to each BGP neighbor individually.

• Configuration and update groups: IOS XR BGP does not support traditional peer groups. Instead, it uses configuration groups and update groups to simplify management and improve performance.

  • Configuration groups allow you to create templates with reusable configuration statements that can be applied to multiple neighbors. This approach reduces configuration errors and ensures consistency across multiple BGP neighbor configurations.

  • Update groups are automatically generated by the system based on common outbound policies among BGP neighbors. Update groups enable efficient sharing of update messages among neighbors that have the same outbound policies, improving scalability and network performance.

BGP configurations are organized into multiple command modes, with each mode enabling you to configure specific aspects of BGP operation. You can enter each mode at the CLI prompt, and within that mode, use the ? command for a list of available configuration commands.

These are common BGP configuration modes with examples demonstrating how to access each:

Router configuration mode

Use this mode to configure global BGP settings for an autonomous system.

Router# configuration
Router(config)# router bgp 140
Router(config-bgp)#

Router address family configuration mode

Use this mode to configure address family–specific BGP settings.

Router(config)# router bgp 112
Router(config-bgp)# address-family ipv4 unicast
Router(config-bgp-af)#

Neighbor configuration mode

Use this mode to configure settings for a specific BGP neighbor.

Router(config)# router bgp 140
Router(config-bgp)# neighbor 10.0.0.1
Router(config-bgp-nbr)#

VRF configuration mode

Use this mode to configure BGP for a specific Virtual Routing and Forwarding (VRF) instance.

Router(config)# router bgp 140
Router(config-bgp)# vrf vrf_A
Router(config-bgp-vrf)#

VRF neighbor configuration mode

Use this mode to configure a BGP neighbor within a specific VRF.

Router(config)# router bgp 140
Router(config-bgp)# vrf vrf_A
Router(config-bgp-vrf)# neighbor 11.0.1.2
Router(config-bgp-vrf-nbr)#

VRF neighbor address family configuration mode

Use this mode to configure address family settings for a BGP neighbor in a VRF.

RP/0/RP0/CPU0:router(config)# router bgp 112
RP/0/RP0/CPU0:router(config-bgp)# vrf vrf_A
RP/0/RP0/CPU0:router(config-bgp-vrf)# neighbor 10.0.1.2
RP/0/RP0/CPU0:router(config-bgp-vrf-nbr)# address-family ipv4 unicast
RP/0/RP0/CPU0:router(config-bgp-vrf-nbr-af)#

VPNv6 address family configuration mode

Use this mode to configure VPNv6-specific BGP address family settings.

Router(config)# router bgp 150
Router(config-bgp)# address-family vpnv6 unicast
Router(config-bgp-af)#

L2VPN address family configuration mode

Use this mode to configure Layer 2 VPN-specific BGP address family settings.

Router(config)# router bgp 100
Router(config-bgp)# address-family l2vpn vpls-vpws
Router(config-bgp-af)#

BGP neighbor configuration in Cisco IOS XR uses submodes to simplify and organize neighbor and address-family specific settings. Each submode determines the command scope and reduces repetitive command syntax for efficient configuration.

Neighbor submode

Neighbor submode allows you to configure attributes for a specific BGP neighbor without prefixing each command with the neighbor keyword.

You can enter neighbor submode by using the neighbor ip-address command in BGP configuration mode.

Benefit:

Streamlines configuration by targeting a single neighbor, reducing command repetition.

  Router(config-bgp)# neighbor 192.23.1.2
  Router(config-bgp-nbr)# remote-as 2002
  Router(config-bgp-nbr)# address-family ipv4 unicast

Address-family submode

Address-family submode, accessed within neighbor submode, enables you to configure address-family-specific settings, such as IPv4 or IPv6 unicast, for a selected neighbor.

You can enter address-family submode for a specific neighbor by using the address-family <afi> <safi> command within neighbor submode.

Benefit:

Groups configuration by address family, minimizes command repetition, and clarifies parameter application.

  Router(config-bgp)# neighbor 2002::2
  Router(config-bgp-nbr)# remote-as 2023
  Router(config-bgp-nbr)# address-family ipv6 unicast
  Router(config-bgp-nbr-af)# next-hop-self
  Router(config-bgp-nbr-af)# route-policy one in

Note	Commands entered in a submode automatically apply to the appropriate neighbor and address-family context. Leaving a submode returns you to the higher-level mode, and changes only apply within the current submode context.

BGP configuration templates in Cisco IOS XR ensure consistent, scalable, and efficient BGP configuration, especially in large-scale deployments. Templates allow you to define common configurations once and apply them to multiple BGP neighbors, reducing configuration errors, and simplifying ongoing maintenance.

Types of BGP configuration templates

There are three primary types of BGP configuration templates, each designed for specific configuration requirements:

• af-group: This template type is used to define address family–specific settings for groups of neighbors. It allows you to configure commands that are dependent on the address family, such as IPv4 or IPv6. Neighbors can inherit these settings by referencing the af-group with the use command.

• session-group: The session-group template is intended for address family–independent (global) settings, which apply to groups of neighbors regardless of address family. Typical commands set at this level include global BGP options. Neighbors inherit these configuration commands from the session-group template by using the use command.

• neighbor-group: This template brings together both address family–independent and address family–dependent configurations. It enables you to apply a comprehensive set of commands, both global and specific to a group of neighbors. Neighbors assigned to a neighbor-group inherit all its configuration settings using the use command. Additionally, a neighbor-group can reference both af-group and session-group templates, allowing for hierarchical inheritance.

Inheritance behavior

• When you assign a neighbor to a group using the use command, the neighbor inherits all settings from the group unless you explicitly override a setting at the neighbor level.

• neighbor-group templates can include references to both session-groups and af-groups, allowing hierarchical inheritance.

• If a setting is configured directly under a neighbor, it overrides inherited values from any template.

• Some inherited configuration settings may be hidden if multiple group types are layered.

Key effects and attributes

• Commands entered at the session-group level define global, address family–independent options. These match the commands available in neighbor submode.

• Commands entered at the af-group level define address family–specific options, as found in the neighbor-address-family configuration submode.

• Commands at the neighbor-group level include both global and address family–dependent options. These commands may reference both af-group and session-group templates through the use command.

Configuration examples

Router(config)# router bgp 140
Router(config-bgp)# af-group afmcast1 address-family ipv4 unicast
Router(config-bgp-afgrp)#
    

session-group:

  Router# router bgp 140
  Router(config-bgp)# session-group session1
  Router(config-bgp-sngrp)# 
   

neighbor-group:

  Router(config)# router bgp 123
  Router(config-bgp)# neighbor-group nbrgroup1
  Router(config-bgp-nbrgrp)#
    

neighbor-group with address family:

 Router(config)# router bgp 140
 Router(config-bgp)# neighbor-group nbrgroup1
 Router(config-bgp-nbrgrp)# address-family ipv4 unicast
 Router(config-bgp-nbrgrp-af)#
    

Note	Directly configured settings always take precedence over inherited template settings.

When multiple configuration templates are applied to a BGP neighbor, the final settings are determined based on a specific order of precedence. Understanding this order is essential to troubleshooting and verifying BGP behavior.

Order of inheritance precedence

When multiple configuration sources apply to a BGP neighbor or group, the effective value is determined in the following order (from highest to lowest precedence):

1. Direct neighbor configuration: Values set directly on the neighbor override all inherited values.

2. Session group or address family group: If not set directly, values from a session group or an address family group override those from a neighbor group.

3. Neighbor group: If not set in the neighbor or higher-level groups, values from a neighbor group (including any values it inherits) are applied.

4. System default: If a value is not configured at any level, the Cisco IOS XR system default is used.

Inheritance rules summary

• Neighbors can inherit configuration from both session groups and neighbor groups, as well as address family groups.

• Neighbor groups can inherit from session groups, address family groups, and other neighbor groups.

• Session groups can inherit from other session groups.

• Address family groups can inherit from other address family groups.

Address family inheritance

• Inheritance rules apply to both address family-independent and address family-dependent configurations.

• If a property can be set in both an address family-dependent and independent context, the most specific value takes precedence, following the same order.

Examples

Example 1: Direct neighbor configuration takes highest precedence

If you set advertisement-interval both on a neighbor group and directly on a neighbor, the neighbor setting is applied.

Router(config)# router bgp 140
Router(config-bgp)# neighbor-group AS_1
Router(config-bgp-nbrgrp)# advertisement-interval 15
Router(config-bgp-nbrgrp)# exit
Router(config-bgp)# neighbor 10.1.1.1
Router(config-bgp-nbr)# remote-as 1
Router(config-bgp-nbr)# use neighbor-group AS_1
Router(config-bgp-nbr)# advertisement-interval 20

This sample output from the show bgp neighbors command shows that the advertisement interval used is 20 seconds:

  Router# show bgp neighbors 10.1.1.1
  
  BGP neighbor is 10.1.1.1, remote AS 1, local AS 140, external link
   Remote router ID 0.0.0.0
    BGP state = Idle
    Last read 00:00:00, hold time is 180, keepalive interval is 60 seconds
    Received 0 messages, 0 notifications, 0 in queue
    Sent 0 messages, 0 notifications, 0 in queue
    Minimum time between advertisement runs is 20 seconds
  
   For Address Family: IPv4 Unicast
    BGP neighbor version 0
    Update group: 0.1
    eBGP neighbor with no inbound or outbound policy; defaults to 'drop'
    Route refresh request: received 0, sent 0
    0 accepted prefixes
    Prefix advertised 0, suppressed 0, withdrawn 0, maximum limit 524288
    Threshold for warning message 75%
  
    Connections established 0; dropped 0
    Last reset 00:00:14, due to BGP neighbor initialized
    External BGP neighbor not directly connected.
  

Example 2: Session-group or address family group overrides neighbor group

If a value is set both in a session-group (or address family group) and in a neighbor group, but not directly on the neighbor, the session group or address family group value is used.

  Router(config)# router bgp 140
  Router(config-bgp)# session-group AS_2
  Router(config-bgp-sngrp)# advertisement-interval 15
  Router(config-bgp-sngrp)# exit
  Router(config-bgp)# neighbor-group AS_1
  Router(config-bgp-nbrgrp)# advertisement-interval 20
  Router(config-bgp-nbrgrp)# exit
  Router(config-bgp)# neighbor 192.168.0.1
  Router(config-bgp-nbr)# remote-as 1
  Router(config-bgp-nbr)# use session-group AS_2
  Router(config-bgp-nbr)# use neighbor-group AS_1
 

This sample output from the show bgp neighbors command shows that the advertisement interval used is 15 seconds:

  Router# show bgp neighbors 192.168.0.1
  
  BGP neighbor is 192.168.0.1, remote AS 1, local AS 140, external link
   Remote router ID 0.0.0.0
    BGP state = Idle
    Last read 00:00:00, hold time is 180, keepalive interval is 60 seconds
    Received 0 messages, 0 notifications, 0 in queue
    Sent 0 messages, 0 notifications, 0 in queue
    Minimum time between advertisement runs is 15 seconds
   For Address Family: IPv4 Unicast
    BGP neighbor version 0
    Update group: 0.1
    eBGP neighbor with no inbound or outbound policy; defaults to 'drop'
    Route refresh request: received 0, sent 0
    0 accepted prefixes
    Prefix advertised 0, suppressed 0, withdrawn 0, maximum limit 524288
    Threshold for warning message 75%
    Connections established 0; dropped 0
    Last reset 00:03:23, due to BGP neighbor initialized
    External BGP neighbor not directly connected.
  

Example 3: Neighbor group inherited value is used when higher precedence is absent

If the neighbor uses only a neighbor group (not a session group or address family group), and no value is set directly on the neighbor, the neighbor group value is used, either directly or through its own inheritance.

  Router(config)# router bgp 150
  Router(config-bgp)# session-group AS_2
  Router(config-bgp-sngrp)# advertisement-interval 20
  Router(config-bgp-sngrp)# exit
  Router(config-bgp)# neighbor-group AS_1
  Router(config-bgp-nbrgrp)# advertisement-interval 15
  Router(config-bgp-nbrgrp)# exit
  Router(config-bgp)# neighbor 192.168.1.1
  Router(config-bgp-nbr)# remote-as 1
  Router(config-bgp-nbr)# use neighbor-group AS_1
    

This sample output from the show bgp neighbors command shows that the advertisement interval used is 15 seconds:

   Router# show bgp neighbors 192.168.1.1
  
  BGP neighbor is 192.168.2.2, remote AS 1, local AS 140, external link
   Remote router ID 0.0.0.0
    BGP state = Idle
    Last read 00:00:00, hold time is 180, keepalive interval is 60 seconds
    Received 0 messages, 0 notifications, 0 in queue
    Sent 0 messages, 0 notifications, 0 in queue
    Minimum time between advertisement runs is 15 seconds
  
   For Address Family: IPv4 Unicast
    BGP neighbor version 0
    Update group: 0.1
    eBGP neighbor with no outbound policy; defaults to 'drop'
    Route refresh request: received 0, sent 0
    Inbound path policy configured
    Policy for incoming advertisements is POLICY_1
    0 accepted prefixes
    Prefix advertised 0, suppressed 0, withdrawn 0, maximum limit 524288
    Threshold for warning message 75%
  
    Connections established 0; dropped 0
    Last reset 00:01:14, due to BGP neighbor initialized
    External BGP neighbor not directly connected.
  

Example 4: Default value is applied if not configured or inherited

If the advertisement-interval is not configured directly or inherited, the system default (30 seconds) applies.

Additional notes

• The same inheritance and precedence rules apply when groups inherit values from other groups.

• For any inheritable property, always refer to the effective configuration displayed by the show bgp neighbors command to verify which value is ultimately used.

Result

These rules ensure that BGP neighbor and group configurations in Cisco IOS XR software are inherited and applied in a predictable order, allowing flexible and scalable BGP policy management.

After applying configuration templates, it's important to verify which settings are active and inherited, using dedicated show commands such as show bgp neighbors configuration, show bgp af-group, and others. These sections provide sample outputs and tips for interpreting results.

BGP can support multiple address families, such as IPv4 and IPv6, each of which must be explicitly enabled and configured. Proper address family configuration ensures correct routing behavior for various protocol families.

No default address family in BGP

• You must explicitly configure each required address family at the BGP router level for it to be active.

• To activate a BGP session for a neighbor under a specific address family, you must configure that address family under both the router and the neighbor configuration.

• You can configure a neighbor without any address family, but cannot configure an address family under a neighbor unless that address family is present at the global BGP router level.

Key Point:

No address family is enabled by default in BGP. Every address family must be explicitly defined both at the global router level and under the neighbor if required.

Neighbor address family combinations

• Default VRF

  In the default VRF, you can configure both:

  • IPv4 unicast

  • IPv4 labeled-unicast

  Both address families are supported and can be enabled under the same neighbor.

• Non-default VRF

  In non-default VRFs, configuring both IPv4 unicast and IPv4 labeled-unicast address families under the same neighbor is not supported.

  The router allows the configuration but displays this error:

  Router: %ROUTING-BGP-4-INCOMPATIBLE_AFI : IPv4 Unicast and IPv4 Labeled-unicast Address families together are not supported under the same neighbor.

  If both address families are present in a BGP session for a neighbor in a non-default VRF, behavior becomes unpredictable, which can result in:

  • Incorrect advertisement of prefixes

  • Reachability issues

Tip	To avoid reachability issues, explicitly configure a route policy to advertise prefixes only through either IPv4 unicast or IPv4 labeled-unicast address families, not both.

Default address family for show commands

Many BGP operational (show) commands allow specifying the Address Family Identifier (AFI) and Subsequent Address Family Identifier (SAFI).

For protocol definitions, see RFC 1700 and RFC 2858.

Setting default AFI and SAFI

In Cisco IOS XR, you can define default AFI and SAFI values for parser commands, eliminating the need to specify them repeatedly.

Parser commands to set defaults:

• set default-afi { ipv4 | ipv6 | all }

• set default-safi { unicast | multicast | all }

Default values:

• AFI: ipv4

• SAFI: unicast

  To change these defaults, you must use the parser commands.

Note	If you specify an AFI or SAFI in a show command, those supplied values override any parser defaults.

Checking current AFI or SAFI defaults

Use the show default-afi-safi-vrf command to view the current default AFI and SAFI settings.

In very large networks, managing a full iBGP mesh can be challenging. BGP confederations address this by partitioning an AS into sub-ASs, improving scalability and simplifying management.