(Optional) Configuring a Secondary VPWS PW

A secondary VPWS PW is mainly used in the following networking modes:

• CEs symmetrically dual-homed to PEs

  Two communication paths exist between the CEs on the two ends of the VC. One is the primary PW path, and the other is the secondary PW path.

• CEs asymmetrically connecting to PEs

  Figure 1 CEs asymmetrically connecting to PEs
  
  

  On the network shown in Figure 1, a pair of primary and secondary PWs need to be configured on PE1. Only one PW needs to be configured on PE2 and PE3. Therefore, the two PEs do not determine whether the PW is a primary or secondary PW.

  The CE on one end of the VC accesses a higher-reliability PE over a single reliable link. The CE on the other end is dual-homed to lower-reliability PEs. The two CEs can communicate over two paths. The higher-reliability path serves as the primary path, and the lower-reliability path serves as the secondary path.

  In the networking where CEs asymmetrically connect to PEs, you need to configure both primary and secondary PWs. The primary and secondary PWs must be of the same type.

  In the scenario where CEs asymmetrically connect to PEs, the interface used by the singled-homed CE to connect to the corresponding PE requires both primary and secondary IP addresses. When the primary path is available, the CE uses the primary IP address to communicate with the remote CE. When a fault occurs on the primary path, this CE uses the secondary IP address to communicate with the remote CE.

  

  On a network where CEs asymmetrically connect to PEs, the secondary PW does not transmit data when the primary and secondary paths both work properly. If the secondary PW AC interface borrows the IP address of the primary PW AC interface, the following situations arise:

  • Non-revertive switching cannot be configured.

  • The local CE has two equal-cost direct routes to the remote CE, with both routes having the same destination address and same next hop. The route that passes through the secondary PW is actually an invalid route.

  • If CEs exchange routing information using routing protocols, you need to change the cost or metric of the secondary path AC interface to be greater than that of the primary path AC interface. In this case, the local CE may fail to communicate with the remote CE, but can communicate with other user devices on the remote end.

  • If CEs use static routes to exchange routing information, ensure that the preference of the backup route is lower than that of the primary route (a larger value indicates a lower preference). To do so, run the ip route-static dest-ip-address mask out-interface preference preference-value command.

  Perform the following steps on the PE to which a CE is single-homed: