The implications of a distributed computing paradigm on multicast routing

Significant interest exists within the military in moving towards an integrated service environment in which various real-time services can co-exist on the same network. An integrated service network can provide a distributed computing environment. To realize this, a number of advances are required principally in intelligent use of the network environment including the reservation of resources, management of the traffic access to the network resources, and effective utilization of the network resources. Multicast routing is an important means of providing effective utilization of the network resources. In developing multicast routing protocols, operational paradigms based on specific application classes are utilized. In the past, two paradigms have been used, the lecture model and the conference model. Both make specific assumptions about the role of participants, how the groups are initiated, the system dynamics, and scalability of the protocol. This has lead to specific routing protocol solutions being biased toward a lecture or conference applications at the expense of distributed computing applications such as distributed interactive simulation. This paper defines the distributed computing paradigm for multicast routing and contrasts it with other multicast paradigms. This paradigm is then used to compare center-based trees and shortest-path trees on the basis of delay, cost, traffic concentration, scalability, protocol overhead, and state. Conclusions are then drawn from this analysis as to which multicast routing protocol is best suited for distributed computing applications such as distributed interactive simulation.