In the PBB VPLS solution, a B-VPLS may be used as infrastructure for one or more I-VPLS instances. B-VPLS control plane (LDP Signaling or P-MSTP) replaces I-VPLS control plane throughout the core. This is raising an additional challenge related to blackhole avoidance in the I-VPLS domain as described in this section.
To address the PBB blackholing issue, assuming that the link between PE A1 and node 5 is active, the remote PEs participating in the orange VPN (for example, PE D) learn the C-MAC X associated with backbone MAC A1. Under failure of the link between node 5 and PE A1 and activation of link to PE A2, the remote PEs (for example, PE D) blackhole the traffic destined for customer MAC X to B-MAC A1 until the aging timer expires or a packet flows from X to Y through the PE A2. This may take a long time (default aging timer is 5 minutes) and may affect a large number of flows across multiple I-VPLSs.
A similar issue occurs in the case where node 5 is connected to A1 and A2 I-VPLS using active/standby pseudowires. For example, when node 5 changes the active pseudowire, the remote PBB PE keeps sending to the old PBB PE.
Another case is when the QinQ access network dual-homed to a PBB PE uses RSTP or M-VPLS with MSTP to provide loop avoidance at the interconnection between the PBB PEs and the QinQ SWs. In the case where the access topology changes, a TCN event is generated and propagated throughout the access network. Similarly, this change needs to be propagated to the remote PBB PEs to avoid blackholing.
A solution is required to propagate the I-VPLS events through the backbone infrastructure (B-VPLS) to flush the customer MAC to B-MAC entries in the remote PBB. As there are no IVPLS control plane exchanges across the PBB backbone, extensions to B-VPLS control plane are required to propagate the I-VPLS MAC flush events across the B-VPLS.