Example: Next-Generation VPLS for Multicast with Multihoming

Step-by-Step Procedure

1. In BGP-based VPLS multihoming, it is recommended that you configure distinct route distinguishers for each multihomed router. Configuring distinct route distinguishers helps with faster convergence when the connection to a primary router goes down. It also requires the other backup PE routers to maintain additional state information for faster convergence.

   There are two levels of path selection:

   • The first is BGP: BGP uses a combination of route distinguisher, site ID, and VE block offset for BGP path selection.

   • The second is in VPLS: VPLS uses the site ID for VPLS path selection.

   By configuring unique route distinguishers, the prefixes for BGP path selection are all unique. Therefore, BGP path selection is skipped and VPLS path selection is used, which only looks at the site ID.

   On Router PE1, Router PE2, Router PE3, and Router PE4 configure a unique router distinguisher for the GOLD routing instance.

2. Configure site ID 1 on Routers PE1 and PE2 for Router CE1. Configure site ID 3 on Routers PE3 and PE4 for Router CE3.

3. Enable multihoming by including the multi-homing statement under the multihomed site configuration on Router PE1, Router PE2, Router PE3, and Router PE4.

4. Include the site-preference primary statement on Router PE1 and Router PE3, and include the site-preference backup statement on Router PE2 and Router PE4. The site-preference primary statement sets the local preference to the highest value (65535) and the site-preference backup statement sets the BGP local preference to 1. Since the site ID is the same, the routers select the highest local preference value as the designated forwarder.

Step-by-Step Procedure

This section presents show commands that you can use to verify the operation of the example configuration.

In this example the traffic patterns are:

• The source is connected to Router CE3 and sends 10,000 pps for the groups 203.0.113.1 to 203.0.113.10. Router CE3 is configured as a rendezvous point.

• Multicast receivers are connected to both Router CE1 and Router CE5. Protocol Independent Multicast (PIM) join messages are generated by the test equipment.

• The link between Router PE3 and Router CE3 and the link between Router PE1 and Router CE1 are configured as primaries for VPLS multihoming.

• All PE routers have a BGP session with the route reflector.

• All PE routers have a label-switched path (LSP) that is created to the route reflector so that the PE routers have a route to the route reflector in the inet.3 table for route resolution.

1. On Router PE1, use the show vpls connections command to verify that the VPLS connections are Up between Router PE1 and Router PE3 and between Router PE1 and PE5. Router PE1 is the primary link selected by the VPLS multihoming configuration.

2. On Router PE2, use the show vpls connections command to verify that the VPLS connections to Router PE3 and Router PE5 are in the LN state, meaning the local router is not the designated forwarder. Router PE2 is configured to be the backup link for Router CE1.

3. On Router PE3, use the show vpls connections command to verify that the VPLS connections to Router PE1 and Router PE5 are Up. Router PE3 is configured to be the primary link for Router CE3.

4. On Router PE4, use the show vpls connections command to verify that the VPLS connections are in the LN state, meaning the local site is not designated. Router PE4 is configured to be the backup link for Router CE3.

5. On Router PE1, use the show route advertising-protocol command to verify that Router PE1 (the multihoming primary router) is sending the BGP Layer 2 VPN route advertisement to the route reflector with the local preference value of 65535. The local preference is used by Router PE3 to select Router PE1 as the designated forwarder, rather than selecting Router PE2 that has a local preference of 1.

6. On Router PE2, use the show route advertising-protocol command to verify that Router PE2 is configured as the multihoming backup with a local preference of 1.

7. On Router PE3, use the show route receive-protocol command to verify that Router PE3 receives the Layer 2 VPN route from the route reflector for Router PE1 and Router PE2 with different local preference values.

   BGP route selection is based on the received l2vpn routes for the VPLS site connected to multihomed PE routers. Since the route distinguishers are different on Router PE1 and Router PE2, Router PE3 and Router PE4 consider the received routes from Router PE1 and Router PE2 as different routes. Router PE3 and Router PE4 run the BGP path selection algorithm and select Router PE1, the router advertising the route with the higher local preference value, as the designated forwarder.

8. On Router PE3, use the show route table command to verify that Router PE3 has selected the static point-to-multipoint LSP from Router PE3 to Router PE1 for forwarding.

   Notice that Router PE2 does not have any provider multicast service interface (PMSI) flags because PMSI attributes are not attached.

9. On Router PE3, use the show vpls connections command to verify that the VPLS connection is in the Up state.

   Notice the display also shows the local interface and the incoming and outgoing label values used.

Step-by-Step Procedure

After the control plane is verified using the previous steps, you can verify the data plane. The data plane operation in the VPLS multihoming scenario is the same as the regular next-generation VPLS operation. This section describes the show command outputs that you can use to validate the data plane.

1. On Router PE3, use the show mpls lsp command to verify the state of the static LSPs and sub-LSPs.

   Router PE2 is configured with static point-to-multipoint LSPs and sub-LSPs with link protection. Point to multipoint LSPs are not supported for single-hop LSPs. In the following output notice that the single-hop point-to-multipoint LSP from Router PE3 to Router PE4 is down.

2. On Router PE1, use the show mpls lsp command to verify the state of the dynamic LSPs.

   Router PE1 is using a dynamic point-to-multipoint LSP template configured with link protection. Notice that the LSP state is Up and that link protection is desired.

3. On Router PE3, use the monitor interface traffic command to verify the multicast replication behavior for the point-to-multipoint LSP on the designated forwarder Router PE3.

   The output shows that 10,000 pps are received on interface ge-1/0/0 from Router CE3. The traffic has been forwarded to the provider (P) Router P2 and Router PE4 through xe-0/0/0 and xe-0/1/0, respectively. Based on the output, you can determine that a single copy of the packet is being sent to Router P2 and Router PE4.

4. On Router P2, use the monitor interface traffic command to verify that the multicast packet replication happens close to the PE routers connected to the receivers.

   Router PE1 and Router PE5 are connected to receivers that have joined this multicast group. Notice that incoming multicast packets from Router PE3 on the ge-0/1/0 interface are replicated twice and sent out on the ge-1/1/0 interface.

5. On Router PE3, use the show vpls flood command to verify information about the flood next-hop route.

   Junos OS Release 9.0 and later identifies the flood next-hop route as a composite next hop. Notice that the interface is ge-1/0/0.1, the next-hop type is composite, and that the flood composition is flood-to-all. This means the traffic is flooded to all the PE routers.

6. On Router PE3, use the show vpls mac-table command to verify that the MAC address of the PE router at the remote end of the VPLS has been learned and added to the MAC address table.

   Notice that the MAC address is learned on the ge-1/0/0.1 interface.

7. On Router PE3, use the show route forwarding-table command to verify that the forwarding table has the required entries with two labels: one for the VPLS service and the other for the next-hop interface.