非确定环境下组播路径QoS分解及资源预约

Partition of QoS in Multicast Tree and Method of Resource Reservation Under Uncertain Conditions

非确定环境中网络链路资源信息是动态变化的,为减少这种变化对组播路由的影响,提高资源预约的成功率,提出了一种将每一组播组成员的端到端QoS约束分解为局部QoS约束,以使组播路径资源预约成功率最大的QoS分解算法——AMOPQ,经分析新算法的时间复杂度为O(maxMe/σ×|T|2).为进一步提高资源预约的成功率,还提出了一种在资源预约阶段使用的新的资源预约方法NMRR.仿真实验结果表明,使用AMOPQ实现组播路径QoS分解后再使用NMRR进行资源预约可获得比单纯使用AMOPQ更好的效果.

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.