Refer to the exhibit. This enterprise customer wants to stream one-way video from their head office to eight branch offices using multicast. Their current service provider provides a Layer 3 VPN…

The SP manages the CE routers, so the customer cannot configure a GRE tunnel between CE1 and CE2 without SP cooperation, which contradicts the requirement for a quick, customer-driven solution.

A GRE tunnel between nodes C2 and C4 creates an overlay that encapsulates multicast traffic as unicast GRE packets, allowing it to traverse the SP's non-multicast L3 VPN without requiring SP involvement or major network redesign. This is the fastest path to enabling multicast because it requires only tunnel interface configuration on customer-controlled nodes. The hub-and-spoke GRE model also scales to additional branches by extending the multicast tree over the existing tunnel infrastructure.

A GRE tunnel between C1 and C4 does not align with the hub-and-spoke distribution path needed to deliver one-way video from the head office to all eight branches efficiently.

Hub-and-spoke MPLS VPN over DMVPN (2547oDMVPN) is a scalable architecture but requires significant SP-side and customer-side configuration effort, making it a slower solution to deploy compared to a simple GRE overlay.

The Draft Rosen solution requires the SP to make changes on their network, which introduces dependency on the SP's timeline and does not meet the requirement of quickly enabling multicast on the customer's terms.