Author_Institution :
Fac. of Software & Inf. Sci., Iwate Prefectural Univ., Iwate, Japan
Abstract :
IP multicast has been suffering from a large scale deployment due to various issues such as upgrading of routers, policy making, negotiation and so on. To overcome problems of IP multicast, overlay multicast in which multicast functionalities are implemented at end hosts has been proposed. However, unlike IP multicast non-leaf nodes in the multicast delivery tree are normal end hosts which can join and leave the tree freely. This causes the multicast tree unstable and disrupted nodes to rejoin the tree. In this paper, we propose that each node when joining the multicast tree sets its leaving time, i.e. how long it will stay in the tree, and sends join request to a number of nodes which are already on the tree. Using the leaving time of nodes, new nodes are joined at the tree such that a child´s leaving time is earlier than its parent, i.e. the child will leave the tree earlier than its parent. This makes the tree more stable than joining at random nodes in the tree. Furthermore, we propose a proactive recovery mechanism so that even if a parent leaves the tree and children do not, children can rejoin at predetermined nodes immediately, so that the recovery time of the disrupted nodes is minimum. We have shown by simulation that there is less overhead when joining the multicast tree and the recovery time of the disrupted nodes is much less than the previous works.
Keywords :
IP networks; media streaming; multicast communication; IP multicast; media streaming; policy making; proactive fault resilient overlay multicast; router upgrading; Bandwidth; Buildings; Information science; Information systems; Large-scale systems; Multicast protocols; Standardization; Streaming media; Topology; Unicast;