Multicast tree rearrangement to recover node failures in overlay multicast networks

Overlay multicast makes use of the Internet as a low level infrastructure to provide multicast service to end hosts. The strategy of overlay multicast slides over most of the basic deployment issues associated with IP multicast, such as end-to-end reliability, flow and congestion control, and assignment of an unique address for each multicasting group. Since each multicast member is responsible for forwarding multicast packets, overlay multicast protocols suffer from multicast node failures. To cope with node failures in the overlay multicast networks, the employment of multicast service nodes (MSNs) is considered which allows relatively high processing performance to cover the disconnected nodes. We are interested in minimizing the cost of both the MSNs and additional links when a node failure occurs. Overlay multicast tree rearrangement to connect multicast members is discussed and formulated as a binary integer programming problem. The tree rearrangement problem is solved by a heuristic based on the Lagrangean relaxation. The performance of the proposed algorithm is investigated by carrying out experiments in 50 and 100 node problems. The employment of MSNs is illustrated to be dependent on the end-to-end delay bound in overlay networks and the degree constraint of member nodes.