Partition of QoS in multicast tree and method of resource reservation under uncertain conditions

The information of link resources in network changes dynamically under uncertain conditions. To reduce its influence on multicast routing and improve the success ratio of resource reservation, an algorithm for multicast optimal partition QoS, i.e., AMOPQ, was proposed, by which the restriction on end-to-end QoS of every multicast member is partitioned into the restriction on local QoS so as to maximize the success ratio of resource reservation. The analysis of AMOPQ reveals that the time complexity of the algorithm is O (maxMe/σ × |T|2). Besides, a new method of resource reservation, i.e., NMRR, is proposed to increase further the success ratio. The simulation result showed that using NMRR to carry out resource reservation after using AMOPQ to partition QoS along multicast path outperforms using AMOPQ only.