DOM: a scalable multicast protocol for next-generation internet

In this article we propose a scalable and efficient destination-oriented multicast protocol for next-generation Internet. With DOM, each packet carries explicit destinations information, instead of an implicit group address, to facilitate the multicast data delivery; and each router leverages the unicast IP routing table to determine necessary multicast copies and next-hop interfaces. The fundamental issue is that we need to limit the bandwidth overhead for such explicit addressing, since it is impractical to attach all the destination addresses to each packet. We resort to the Bloom filter technique to encode the destination information carried by each packet for bandwidth efficiency, with elaborated design to accommodate the features of the practical Internet including longest-prefix matching, route aggregation, and asymmetric interdomain routing. Moreover, DOM enables a fast group joining approach to minimize the joining delay perceived by receivers. The scalability and efficiency of DOM are demonstrated with simulation results.

DOM

In the DOM service model, each receiver domain sends a join message to the source node of the groups in which it is interested; based on the join messages, each source node can know the members for each group it provisions. Each multicast packet will carry the destination addresses of all the members so that each router can retrieve the addresses and leverage the unicast routing table to compute necessary copies and output interfaces. DOM provides a uniformed intra-/interdomain multicasting service, but we focus on the interdomain case in this article for convenience of demonstration.