Deployment issues for application level multicast

ALM (Application Level Multicast) has been suggested to remove the biggest barrier to block deployment of IP multicast, scalability problem on core networks. ALM can provide more extended multicast service by moving functionalities for group communication from network routers to higher application layer components on hosts, but it has also difficulties in providing scalable structure for deployment on Internet because its mechanism is largely based on hosts for actual services instead of network nodes. Therefore, this paper suggests a scalable and common architecture in deployment aspect, and proposes a scheme for multicast supports particularly based on this mechanism. The proposed mechanism is based on two-level architecture: (1) existing host group model can be used between DR (Designated Router) and hosts on the DR's subnet, (2) application level multicast mechanism can be applied between DRs. Simulation shows performance results of end-to-end delay, maximum degrees, overhead for a new path computation, and bandwidth utilization within a subnet with our service model.