高级检索
    黄东军 陈松乔 王建新. 结合集中式与分布式特征的多路径QoS组播路由协议[J]. 计算机研究与发展, 2005, 42(8): 1403-1408.
    引用本文: 黄东军 陈松乔 王建新. 结合集中式与分布式特征的多路径QoS组播路由协议[J]. 计算机研究与发展, 2005, 42(8): 1403-1408.
    Huang Dongjun, Chen Songqiao, and Wang Jianxin. A QoS-Aware Multicast Routing Protocol Based on Centralized and Distributed Algorithms[J]. Journal of Computer Research and Development, 2005, 42(8): 1403-1408.
    Citation: Huang Dongjun, Chen Songqiao, and Wang Jianxin. A QoS-Aware Multicast Routing Protocol Based on Centralized and Distributed Algorithms[J]. Journal of Computer Research and Development, 2005, 42(8): 1403-1408.

    结合集中式与分布式特征的多路径QoS组播路由协议

    A QoS-Aware Multicast Routing Protocol Based on Centralized and Distributed Algorithms

    • 摘要: 提出了一个结合集中式算法与分布式算法优点的多路径启发式QoS组播路由算法和协议,它以单播路由协议OSPF传播链路的代价信息为基础,运用最小代价Dijkstra算法计算端节点到当前在树节点的最小代价路径,然后启动一个分布式计算过程得到一个可选路径集,加入节点通过一个综合性启发式选择其中的最佳路径连接到组播树.算法能够有效地支持延时和带宽受限的代价优化组播树构造,具有无环选路、呼叫接收成功率高、呼叫建立时间短、伸缩性好等特点.

       

      Abstract: QoS-aware multicast service is an essential component in the future high speed Internet. QoS-aware multicast routing algorithm and protocol are the core of QoS-based multicast architecture. Proposed in this paper is a multi-path and QoS-aware multicast routing protocol integrated with both a centralized algorithm and a distributed one. A node initially takes, from a special node called the manager, some information about the existing tree the receiver wants to join. Then it uses Dijkstra's minimum cost algorithm with some costs as the parameters to calculate several minimum cost paths from the joining receiver to the source node and the on-tree nodes of the group, and regards them as a set of feasible paths for selection. When a joining receiver gets the set, it employs a heuristic to select the best path to join multicast tree. Analysis and experiments prove that this algorithm can be used to solve the delay and bandwidth-constrained minimum-cost multicast tree problem for it has many advantages such as loopfree, high average call accepted ratio, shorter average call setup time, and better scalability.

       

    /

    返回文章
    返回