A scalable multicast routing

Abstract

With the growing interests in distributed applications, efficient and scalable multicast is of great concern. Widely used multicast routing protocols, represented by the DVMRP, however become very inefficient when used over large scale internetworks. New multicast routing protocols (RPs) like PIM (protocol independent multicast) and CBT (core based trees) being developed by the IDMR working group of the IETF, trade efficiency of multicast distribution trees with scalability, but they have some issues open to further study. The work presented in this paper, designated rendezvous point multicast routing (DRP), is essentially complementary to PIM-SM (sparse mode) and CBT, yet introduces new concepts to resolve their residual problems. While presenting basic features of the shared-tree structure found in PIM or CBT, the proposed DRP scheme takes an intelligent approach to tree creation and management so that it can generate multicast delivery trees of lower average cost with less control overhead. Improvement on scalability is sought by using a hierarchical architecture, which requires the multicast addresses be administratively scoped in accordance with the hierarchy. Management of candidate RPs are restricted within individual hierarchical regions. Meanwhile, the DRP scheme is recursive, meaning that all routers run the same algorithm regardless their hierarchical levels.