Although IP Multicast is an effective network primitive for best-effort, large-scale, multi-point communication, many multicast applications such as shared whiteboards, multi-player games and software distribution require reliable data delivery. Building services like reliable sequenced delivery on top of IP Multicast has proven to be a hard problem. The enormous extent of network and end-system heterogeneity in multipoint communication exacerbates the design of scalable end-to-end reliable multicast protocols. In this paper, we propose a radical departure from the traditional end-to-end model for reliable multicast and instead propose a hybrid approach that leverages the successes of unicast reliability protocols such as TCP while retaining the efficiency of IP multicast for multi-point data delivery. Our approach splits a large heterogeneous reliable multicast session into a number of multicast data groups of co-located homogeneous participants. A collection of application-aware agents--Reliable Multicast proxies (RMXs)--organizes these data groups into a spanning tree using an overlay network of TCP connections. Sources transmit data to their local group, and the RNLX in that group forwards the data towards the rest of the data groups. RMXs use detailed knowledge of application semantics to adapt to the effects of heterogeneity in the environment. To demonstrate the efficacy of our architecture, we have built a prototype implementation that can be customized for different kinds of applications.

The IP Multicast service model extends the traditional best effort Internet datagram delivery service for efficient multi-point packet delivery. However, in spite of a decade of research on multicast protocols and applications, a globally deployed multicast service is nowhere in sight, hindered by multitudes of problems such as manageability, lack of a robust inter-domain multicast routing protocol, scalability, and heterogeneity. In this work, we propose a new model for Internet multicast where we view multi-point delivery not as a network primitive but rather as an application-level infrastructure service. Our architecture relies on a collection of strategically placed network agents that collaboratively provides the multicast service for a session. Clients locate a nearby agent and tap into the session via that agent. Agents organize themselves into an overlay network of unicast connections and build data distribution trees on top of this overlay structure. This model effectively pa...