A heuristic overlay multicast routing algorithm for minimum delay optimization

Multicast services can be provided either as a basic network service by IP multicast or as an application layer service called overlay multicast. How to build an efficient data delivery structure is the most critical problem for overlay multicast, which depends on a good routing algorithm. Many routing algorithms have been presented recent years. Some algorithms modeled overlay multicast's routing problem as a degree constrained spanning tree problem based on undirected completed graph, and some others were built on some special peer to peer routing protocols. However, detecting and obtaining a completed graph's all network state information needs consume expensive cost and restricts its scalability and sometime be impossible. On the contrary these routing algorithms built on p2p scheme are of good scalability but maybe fail in practical applications. This paper redefines the overlay multicast routing problem as to find a degree constrained sub-tree with minimum delay optimization for a general undirected connected graph not a completed graph. It is a NP-hard problem. A heuristic genetic algorithm for this problem is given and the simulation results verify its effectiveness.