Abstract
This document provides functional requirements for network solutions that support multicast over Virtual Private LAN Service (VPLS). It specifies requirements both from the end user and service provider standpoints. It is intended that potential solutions will use these requirements as guidelines.
Introduction
RFC 5501, published in March 2009, serves as a foundational document for the evolution of Virtual Private LAN Services (VPLS). As bandwidth-intensive applications like IPTV and multi-party video conferencing became prevalent, the traditional "flood-and-learn" mechanism of VPLS proved inadequate for handling multicast traffic. This paper outlines the essential functional requirements for developing a more efficient "multicast VPLS" framework that optimizes core network bandwidth and improves scalability for service providers while maintaining high Quality of Service (QoS) for end-users.
The Problem: Inefficiencies of Traditional VPLS
Traditional VPLS emulates a large Ethernet switch across a wide-area network. While effective for unicast traffic, it handles multicast and broadcast traffic by replicating frames at the Provider Edge (PE) and sending them to every other PE in the VPN instance. This approach leads to two primary inefficiencies:
- Wasteful Replication: Traffic is delivered to sites that have no active receivers for a specific multicast group, wasting precious bandwidth.
- Redundant Core Traffic: Multiple copies of the same stream may traverse the same physical backbone links simply because they are destined for different PEs, creating bottlenecks in the provider core.
A Requirements-Driven Approach
As an Informational RFC, this document does not mandate a specific protocol or algorithm. Instead, it uses a requirements engineering approach to define the needs of two primary stakeholders: the Customer (end-user) and the Service Provider (SP). It establishes the "rules of the road" for designing solutions that solve the replication issues inherent in early VPLS architectures.
Key Technical Pillars
The authors define a multicast VPLS solution through several core functional concepts:
- Layer-3 Awareness: While VPLS is fundamentally a Layer-2 service, the document argues that PEs may need to "snoop" on IP multicast protocols (e.g., IGMP or PIM) to intelligently identify where receivers actually reside.
- Multicast Distribution Tunnels (MDTunnels): The document introduces the concept of MDTunnels—point-to-multipoint or multipoint-to-multipoint paths—to replace inefficient head-end replication in the provider core.
- Optimality vs. State Trade-off: A central engineering challenge identified is balancing the creation of perfect multicast trees for every group (high efficiency) against the aggregation of groups into fewer trees (lower state/memory usage).
Key Insights and Mandates
Fate-Sharing and Transparency
A critical requirement is "fate-sharing," where unicast and multicast traffic follow the same logical paths to ensure consistent application behavior during network failures. Furthermore, the solution must remain transparent to the end-user, requiring no changes to their existing Layer-3 multicast protocols; the provider network should adapt to the customer, not vice versa.
Scalability and OAM
Solutions must scale effectively relative to the number of PEs, VPLS instances, and multicast groups. Additionally, the RFC mandates robust Operations, Administration, and Maintenance (OAM) tools, allowing providers to test connectivity per multicast group and monitor performance metrics like jitter, loss, and delay specifically for multicast streams.
Security
The document treats security as a first-class citizen, identifying threats such as "join/leave storms" that could lead to Denial of Service (DoS) on PE routers. It requires solutions to implement policing and isolation mechanisms to protect the provider infrastructure from customer-driven traffic anomalies.
Conclusion and Legacy
RFC 5501 provided the necessary guidelines that eventually led to standardized protocols like BGP-based VPLS multicast (RFC 7117). By formalizing the trade-offs between efficiency and complexity, it allowed the industry to move toward scalable, high-performance multicast delivery in provider-provisioned VPNs without requiring "forklift upgrades" to existing infrastructure.
Summary updated