Extensions to the Path Computation Element Communication Protocol (PCEP) for Point-to-Multipoint Traffic Engineering Label Switched Paths

Quintin Zhao, Daniel King, Fabien Verhaeghe, Tomonori Takeda, Zafar Ali, Julien MeuricRFC 2010 · DOI

TL;DR

This paper introduces crucial extensions to the Path Computation Element Communication Protocol (PCEP) to enable the computation and management of Point-to-Multipoint (P2MP) Traffic Engineering Label Switched Paths (TE LSPs). These extensions are vital for efficiently establishing and optimizing complex network topologies where a single source needs to reach multiple destinations.

What Problem Does This Paper Address?

The existing Path Computation Element (PCE) and its communication protocol (PCEP) were primarily designed for point-to-point (P2P) connections. However, modern networks increasingly require the establishment of Point-to-Multipoint (P2MP) TE LSPs, where a single traffic flow from a source needs to be distributed to multiple destinations. The challenge was that the standard PCEP lacked the necessary mechanisms to signal, request, and manage these P2MP LSPs, leaving a significant gap in network management capabilities for complex, multi-destination traffic engineering.

Methods

The paper proposes several key extensions to PCEP:

  • P2MP Capability Advertisement: New mechanisms are introduced for PCEs to advertise their P2MP computation capabilities, both through existing IGP discovery protocols (OSPF, IS-IS) and during the PCEP session establishment (Open Message Extension).
  • P2MP Path Computation Request/Reply Message Extensions:
    • RP Object Extension: New flags are added to the Request Parameters (RP) object to indicate P2MP requests, fragmentation status, and ERO compression.
  • New END-POINTS Object: This object is crucial for specifying multiple destinations (leaves) in a P2MP request, supporting different leaf types (new, old-modifiable, old-unchanged) to efficiently manage P2MP trees.
  • Efficient Presentation of P2MP LSPs: The use of RSVP-TE Explicit Route Objects (EROs) and Secondary Explicit Route Objects (SEROs) is specified for encoding P2MP LSP paths.
  • P2MP Objective Functions and Metric Types: Two new objective functions (Shortest Path Tree - SPT, Minimum Cost Tree - MCT) are defined for P2MP path computation, along with three new metric types (P2MP IGP, P2MP TE, P2MP hop count).
  • Handling Non-Support and Errors: Mechanisms are defined for PCEs to indicate non-support of P2MP computation, including specific PCEP-ERROR objects and types for various scenarios like policy violations, capability issues, and fragmentation failures.
  • P2MP Path Reoptimization and Management: The document details how to request reoptimization of existing P2MP paths, add/prune leaves, and discover branch nodes.
  • Synchronization and Diversity: Extensions are introduced for synchronizing P2MP LSP computations and requesting path diversity (partial and link direction diverse).
  • Fragmentation: Procedures for fragmenting large P2MP requests and responses across multiple messages are defined to overcome PCEP message size limitations.
  • UNREACH-DESTINATION Object: This object allows PCEs to report unreachable destinations in P2MP path computation requests.

Datasets

Not discussed in the paper.

Key Insights and Ideas

  • P2MP is a distinct problem: The paper highlights that P2MP path computation is not simply an extension of P2P but requires fundamentally different signaling and management constructs due to the multi-destination nature.
  • Efficient leaf management is key: The introduction of the new END-POINTS object with distinct leaf types is a clever way to manage the complexity of P2MP trees and reduce message overhead.
  • Backward compatibility is preserved: The extensions are designed to work with existing PCEP mechanisms, allowing for gradual adoption and interoperability with older implementations.
  • Comprehensive error handling: The detailed definition of PCEP-ERROR objects and types for P2MP-specific issues is crucial for robust network operation.

Key Results

The paper defines the protocol extensions and message formats necessary for PCEP to support P2MP TE LSPs. While specific quantitative results from implementation or testing are not presented, the document establishes the framework for:

  • PCEs to advertise and PCCs to discover P2MP capabilities.
  • PCCs to request P2MP path computations with multiple destinations.
  • PCEs to compute and return P2MP paths.
  • Managing and reoptimizing existing P2MP LSPs.
  • Handling various error conditions and non-support scenarios.

Limitations and Future Work

  • Implementation and Testing: The paper focuses on protocol definition. Future work would involve implementing these extensions and conducting thorough testing to validate their performance and interoperability in real-world scenarios.
  • MIB Definitions: The paper notes that a new document will be required to define MIB objects for PCEP control and monitoring of P2MP computations, which is a natural next step for manageability.
  • Performance Analysis: While the paper discusses potential increases in PCE state and load during network failures, a detailed performance analysis of the proposed extensions under various conditions would be beneficial.

Summary updated

BibTeX
@article{zhao2010extensions,
  author = {Quintin Zhao and Daniel King and Fabien Verhaeghe and Tomonori Takeda and Zafar Ali and Julien Meuric},
  title = {{Extensions to the Path Computation Element Communication Protocol (PCEP) for Point-to-Multipoint Traffic Engineering Label Switched Paths}},
  journal = {RFC},
  year = {2010},
  doi = {10.17487/rfc6006},
}

Loading related content...

Figure