Establishing an MP-EBGP Peer Relationship Between PEs

By introducing extended community attributes into BGP, MP-EBGP can advertise VPNv4 routes between PEs. PEs of different ASs are generally not directly connected. Therefore, to set up the EBGP connection between the PEs of different ASs, you need to configure the permitted maximum hops between PEs.

Procedure

Configure an autonomous system boundary router (ASBR) to send the loopback interface IP addresses of a PE used for peer relationship establishment to the ASBRs of other ASs and peer PEs. You can also configure a PE to advertise its loopback interface IP addresses used for peer relationship establishment to the peer PEs.

If you want to use inter-AS traffic engineering (TE) tunnels to transmit traffic in inter-AS Option C networking, perform the following steps on PEs so that the loopback interface IP addresses of the PEs used for peer relationship establishment can be advertised to peer PEs.
1. Run system-view
  The system view is displayed.
2. Run bgp { as-number-plain | as-number-dot }
  The BGP view is displayed.
3. Run ipv4-family labeled-unicast
  The bgp labeled unicast address family view is displayed.
4. Run network ip-address [ mask | mask-length ] [ route-policy route-policy-name ]
  The loopback address of the PE in the local AS is advertised to the remote ASBR.
5. Run commit
  The configuration is committed.
(Optional) Disable an ASBR from advertising BGP supernet labeled routes.
1. Run system-view
  The system view is displayed.
2. Run bgp { as-number-plain | as-number-dot }
  The BGP view is displayed.
3. Run ipv4-family labeled-unicast
  The bgp labeled unicast address family view is displayed.
4. Run supernet label-route advertise disable
  The ASBR is disabled from advertising BGP supernet labeled routes.
  
  After you disable the ASBR from advertising BGP supernet labeled routes, to advertise the loopback address route of a PE in the local AS to a PE in another AS, run the network command on the ASBR to advertise the BGP route to the loopback address of the PE in the same AS.
5. Run commit
  The configuration is committed.
Import the labeled routes to the public network's unicast address family routing table.
Perform the following steps on the PE.
1. Run system-view
  The system view is displayed.
2. Run bgp { as-number-plain | as-number-dot }
  The BGP view is displayed.
3. Run import-rib public labeled-unicast
  Imports the labeled routes to the public network's unicast address family routing table.
4. Run commit
  The configuration is committed.
Perform the following steps on the PE that is connected to a customer edge (CE).
1. Run system-view
  The system view is displayed.
2. Run bgp { as-number-plain | as-number-dot }
  The BGP view is displayed.
3. Run peer ipv4-address as-number { as-number-plain | as-number-dot }
  The peer PE is specified as the EBGP peer.
4. Run peer ipv4-address connect-interface loopback interface-number
  The source interface that sends BGP packets is specified.
5. Run peer ipv4-address ebgp-max-hop [ hop-count ]
  The maximum number of hops between PEs for which an EBGP peer relationship is to be configured is specified.
  
  PEs of different ASs are generally not directly connected. To set up the EBGP peer relationship between PEs of different ASs, configure the maximum number of hops between PEs and ensure that PEs are reachable.
6. Run ipv4-family vpnv4 [ unicast ]
  The BGP-VPNv4 address family is displayed.
7. Run peer ipv4-address enable
  The function to exchange VPNv4 routes with the peer PE is enabled.
8. Run commit
  The configuration is committed.
(Optional) Configure an RR.
To improve scalability, specify an RR in each AS and establish MP-EBGP peer relationships between the RRs in ASs to save all VPNv4 routes on the RRs. Then configure PEs in each AS as the RR's clients to exchange VPNv4 routing information with the RR. Perform the following steps on a device that needs to be configured as an RR:
1. Run system-view
  The system view is displayed.
2. Run bgp { as-number-plain | as-number-dot }
  The BGP view is displayed.
3. Run peer ipv4-address1 as-number { as-number-plain | as-number-dot }
  The RR in another AS is configured as its EBGP peer.
4. Run peer ipv4-address1 connect-interface loopback interface-number
  The source interface that sends BGP packets is specified.
5. Run peer ipv4-address1 ebgp-max-hop [ hop-count ]
  The maximum number of hops between RRs for which an EBGP peer relationship is to be configured is specified.
6. Run ipv4-family vpnv4 [ unicast ]
  The BGP-VPNv4 address family view is displayed.
7. Run peer ipv4-address1 enable
  The function to exchange VPNv4 routes with the peer RR is enabled.
8. Run quit
9. Run peer ipv4-address2 reflect-client
  The device is configured as an RR, and PEs are specified as its clients.
10. Run peer ipv4-address1 next-hop-invariable
  The next hop is not changed when the route is advertised to the EBGP peer.
11. Run peer ipv4-address2 next-hop-invariable
  The next hop is not changed when the route is advertised to the IBGP peer.
12. Run commit
  The configuration is committed.