Scalable group-tree matching algorithm in aggregated multicast

Abstract

IP multicast suffers from scalability problems when the number of simultaneously active multicast groups is very large. In this paper, we propose an algorithm called SGTM (Scalable Group-Tree Match algorithm) which reduces the number of trees by allowing several groups to be aggregated to the same delivery tree. SGTM improves the state scalability because it can reduce the required multicast state and the management overhead for the distribution trees. The paper introduces metrics to measure multicast state and tree management overhead for multicast schemes. SGTM performs faster aggregations than previous aggregation algorithms by evaluating fewer trees for each group, while keeping the same performance. We then show the scalability and the fastness of SGTM by several simulations and we compare its performance to the previous algorithm.