Paper Review #12: The PIM Architecture for Wide-Area Multicast Routing

From: Yuhan Cai (yuhancai@cs.washington.edu)
Date: Tue Nov 09 2004 - 23:42:30 PST

  • Next message: Erika Rice: "Review 11-10"

    Title: The PIM Architecture for Wide-Area Multicast Routing
    Authors: Stephen Deering, Deborah L. Estin, Dina Farnacci, Van Jacobson, Ching-Gung Liu, and Liming Wei
    Reviewed by: Yuhan Cai

    Main results of the paper:

    , It presents a multicast architecture that guarantees an efficient routing in wide-area internets, while existing mechanisms are only for regions where a group is widely represented.

    , It uses soft-state mechanisms to provide adaptation to different network conditions and group dynamics.

    Strengths of the paper:

    , The protocol supports both shared and shortest path tree types.

    , It is independent of specific underlying unicast protocols.

    , It is robust, scalable and flexible and therefore it is suited for large heterogeneous networks.

    Key limitations:

    , Few experimental results are given. More empirical data that demonstrate the superiority of PIM in a sparsely distributed environment should be more persuasive.

    Relevance of the paper:

    , PIM serves as a good complement to existing routing protocols when network nodes are widely distributed over a large area.

    Future work:

    , Future research work is expected on scaling issues such as the memory bottleneck. As the size of the network increases, both the memory requirements and the message capacities might become unmanageable.

    , Finding a way for PIM to interact properly with other routing protocols might be of further interest.

    , How to interact with Receiver Initiated Reservation Setup is another issue.


  • Next message: Erika Rice: "Review 11-10"

    This archive was generated by hypermail 2.1.6 : Tue Nov 09 2004 - 23:42:36 PST