An MPLS service provider is offering a standard EoMPLS-based VPLS service to CustomerA providing Layer 2 connectivity between a central site and approximately 100 remote sites. CustomerA wants to…

EoMPLS-based VPLS is not inherently scalable for multicast because it floods multicast frames to all participating PE nodes by default, meaning all 20 streams would be sent to all 100 sites and consume significant unnecessary bandwidth.

A full mesh of GRE tunnels across 100 sites requires nearly 5,000 tunnel endpoints, adding enormous configuration complexity and overhead that makes it far less scalable than leveraging snooping within the existing VPLS fabric.

IGMP snooping on provider PE routers tracks multicast group membership across VPLS pseudowires and replicates multicast frames only toward PEs serving sites with active receivers, preventing the default flood-to-all behavior that would otherwise replicate all 20 video streams to every one of the 100 remote sites regardless of whether any receiver exists there.

While Layer 3 MVPN would provide scalable multicast delivery, it requires replacing the existing Layer 2 VPLS service entirely, which is a more disruptive change than enabling snooping as an optimization within the current service.