Paper Review: < HC99 >>
IP Multicast Channels: EXPRESS Support for Large-scale Single-source Applications

Reviewer: Oleg Elkhunovich

Problem

IP multicast model lacks a basis for charging, access control and is difficult to scale which can be a problem for a variety of applications.

Contribution

This paper presents an alternative model for IP multicast that addresses the problems mentioned above - EXPRESS (Explicitly Requested Single-Source multicast).

Main Ideas

  • Extending the IP multicast service model to support multicast channels. A multicast channel is a datagram delivery service identified by a tuple (S,E) where S is the sender's source address and E is a channel destination address.
  • Express is implemented using ECMP, a single common management protocol that both maintains the distribution tree and supports source-directed counting and voting.
  • Multi-source applications can be build on top of EXPRESS channels by using multiple channels, one per source, or by allowing several sources to share a channel using higher level relaying through the channel's source host.

    Critique

    Significance: 4
    The paper is significant as it presents a novel approach to IP multicast that solves a number of problems that are in the existing IP multicast model and allows for deployment of a variety of new applications that require the properties given by EXPRESS.
    Methodology:
    Methodology of this work is very good with authors stating the problems, presenting a solution and showing how it solves each of the aspects of the problem. They then also analyze different performance and other issues theoretically. However many calculations seem to be unconvincing and no real-world testing on large scale is done.
    Limitation:
    It's unclear how simulations were run. No real-world testing. Cost calculations seem to be wrong.

    Lessons:
    Present IP multicast model is not sufficient to support variety of commercial applications of content delivery.