AN APPROACH FOR CLUSTER-BASED MULTICAST ROUTING IN LARGE-SCALE NETWORKS

This paper addresses the optimum routing problem of multipoint connect ion in large-scale networks. A number of algorithms for routing of mul tipoint connection have been studied so far, most of them, however, as sume the availability of complete network information. Herein, we stud y the problem under the condition that only partial information is ava ilable to routing nodes and that routing decision is carried out in a distributed cooperative manner. We consider the network being partitio ned into clusters and propose a cluster-based routing approach for mul tipoint connection. Some basic principles for network clustering are d iscussed first. Next, the original multipoint routing problem is defin ed and is divided into two types of subproblems. The global optimum mu lticast tree then can be obtained asymptotically by solving the subpro blems one after another iteratively. We propose an algorithm and evalu ate it with computer simulations. By measuring the running time of the algorithm and the optimality of resultant multicast tree, we show ana lysis on the convergent property with varying network cluster sizes, m ulticast group sizes and network sizes. The presented approach has two main characteristics, 1) it can yield asymptotical optimum solutions for the routing of multipoint connection, and 2) the routing decisions can be made in the environment where only partial information is avai lable to routing nodes.