The work reported in this paper starts from the observation that one cannot make a single reliable multicast delivery scheme that optimally meets the functionality, scalability, and efficiency requirements of all applications. It has been concluded elsewhere that the best way to meet diverse application requirements is to leave as much functionality and flexibility as possible to the applications. This approach has been successful in the design of wide-area, large-scale conferencing applications.

The scalable reliable multicast (SRM) framework is based on the group delivery model of the IP multicast protocol. SRM enhances the multicast group concept through data sharing among the members, and makes each member responsible for its own correct reception of the data.

The SRM protocol framework follows the core design principles of TCP/IP. It requires only the basic IP delivery model; no change or special support is required from the underlying IP network. SRM adjusts its parameters based on the observed performance within sessions. This allows applications using the SRM framework to adapt to a wide range of group sizes, topologies, and link bandwidths.

The paper is well written and worth reading for anyone interested in this specific area. The presentation proceeds logically and systematically from describing the design principles of the SRM framework to analysis and simulations demonstrating the performance of the framework. Related work and planned future developments by the authors are discussed.