In Basic MVR Deployment, only simple topologies are considered, in which subscribers on different VLANs access a multicast VLAN. There are topologies where streams need to be forwarded onto another multicast VLAN, as shown in Inter-Multicast VLAN Forwarding. In this figure, a Multicast Service Provider (MSP) multicast VLAN is attached to ports 1:1-2 on both switches, SW1 and SW2. On the customer side, another multicast VLAN, delivers multicast streams to other switches around the ring.
In this topology, a multicast stream can be leaked into the customer multicast network through either switch SW1 or SW2. However, as described in MVR Forwarding, packets are not forwarded to router ports (ports 1:4 and 1:5 can be router ports if SW2 is an IGMP querier). To get around this, MVR needs to be configured on CustVlan either on SW1 or SW2. Since the forwarding rules apply only to non-MVR VLANs, traffic from one MVR VLAN is leaked into the router ports of another VLAN, if MVR is enabled on that.
In the topology above, the MSP multicast VLAN is carried on two switches that also carry the customer multicast VLAN. When multiple switches carry both multicast VLANs, it is imperative that MVR is configured on only one switch. Only that switch should be used as the transit point for multicast streams from one multicast ring into another. Otherwise, duplicate packets are forwarded. Also on the non-MVR switches, the ring ports should be configured as static router ports, so that ring ports are excluded from forwarding packets onto the customer ring. There is no mechanism to elect a designated MVR forwarder, so it must be configured correctly.