Usage Scenario
- For the VSI that adopts the static member discovery mechanism, you need to manually specify the peer address for the VSI.
- To configure a spoke PW on the HVPLS network, you can configure upe in the peer command. If the specified peer is a UPE, the peer is on the user side, which does not comply with split horizon.
In a split horizon scenario, the data packets received from the PW at the PSN side are not forwarded to other PWs. Instead, they are forwarded to the private network. On a VPLS network, full mesh and split horizon are used to prevent loops. HVPLS partitions the network, and the devices of different levels forward data to each other without complying with the split horizon principle.
- To configure a VPWS accessing VPLS network, you can adopt one of the following configurations to create a PW:
- Configuring the VSI ID to be the same as the VC ID of the peer
- Configuring the value of negotiation-vc-id vc-id to be the same as the VC ID of the peer
If the value of negotiation-vc-id vc-id is different from the value of vsi-id specified in the vsi-id command, you have to configure negotiation-vc-id vc-id when running the undo peer command to delete the VSI peer. Otherwise, the VSI peer cannot be deleted.
- On a network where PW redundancy is configured, when the primary PW fails, traffic is switched to the secondary PW. If the secondary PW is in the backup state, traffic cannot be forwarded, resulting in packet loss. If ignore-standby-state is configured for the PWs on the dual-homed PEs, the secondary PW will ignore the secondary state indicated by the peer and be always in the forwarding state, preventing packet loss during a primary/secondary PW switchover.
- On a PBB VPLS network where PW redundancy is configured, multiple I-VSIs are bound to a B-VSI, and ignore-standby-state is not used during PW configuration, the primary/secondary PW switchover relies on protocol convergence when the primary PW fails. Only the NetEngine 8000 F1A supports PBB VPLS.
- The secondary PW can be created to protect the primary PW. When the primary PW fails, traffic can be immediately switched to the secondary PW, preventing traffic interruption.
Note the following when you specify the secondary PW:
- The primary PW of the same VSI must exist and be unique, and the primary PW is not configured on a static U-PE.
- After secondary is configured in the peer command, the number of PWs that can be configured for the VSI is restricted. If the secondary PW is configured for a VSI, that VSI can have only two PWs, the primary PW and the secondary PW.
- To change the primary or secondary state of a PW, you need to delete the PW and then reconfigure it. In addition, the state parameter specified in the undo peer command must be consistent with the configured state of the PW.
- Before deleting the primary PW, delete the secondary PW of the same VSI.
- In a network on which PWs and an EVPN anycast VTEP interwork and the interworking points (PE2 and PE3) work in active-active mode, PE1 is dual-homed to PE2 and PE3 through a VPLS network. PE2 and PE3 work in master/backup mode, and assume that PW1 is the primary PW and PW2 is the secondary PW. PE2 and PE3 work in active-active mode when PE4 is dual-homed to them. PE2 and PE3 use a virtual anycast VTEP address to establish an anycast VXLAN tunnel with PE4. A bypass VXLAN tunnel is established between PE2 and PE3 to forward traffic in case of a link fault.
The bypass VXLAN tunnel works in Hub PW mode, whereas the anycast VXLAN tunnel works in Spoke PW mode. PW1 and PW2 are common VPLS PWs, and they are Hub PWs by default. In this case, PW1 does not process traffic forwarded by the bypass VXLAN tunnel. If PW1 is configured as a Spoke PW using the peer peer-address upe command, PW1 does not process traffic forwarded by the anycast VXLAN tunnel. To enable PW1 to process traffic sent by both anycast VXLAN tunnel and bypass VXLAN tunnel, run the peer peer-address ac-mode command to set PW1 to AC mode.
Prerequisites
- LDP sessions have been set up between all the PEs to fully mesh PEs on a VPLS network through PWs.
- The VSI ID has been set in the VSI-LDP view.
- The peer peer-address [ negotiation-vc-id vc-id ] pw pw-name command can be used to create the PW view only after the peer peer-address [ negotiation-vc-id vc-id ] [ tnl-policy policy-name ] command has been run to specify the IP address of the peer.
Configuration Impact
- Specifying the peer address means manually specifying the members in a VPLS domain.
- The execution of the undo peer command will interrupt traffic of the peer. Exercise caution when using this command.
Precautions
A VPLS network is a point-to-multipoint L2VPN and therefore multiple peers can be specified for one VSI.
If a BD has bound a VSI and an EVPN instance bound, the upe parameter cannot be specified in the
peer command to specify a UPE as a peer for the VSI.