高级检索

    全部到全部组播的核管理机制及其算法研究

    Core Management and Algorithms in All-to-All Multicast

    • 摘要: 组播是因特网上的重要应用和具有挑战性的研究问题,现在大多数已有的组播算法是基于有源树的,存在带宽利用率和传输效率问题.有核组播是在网络中选取一个核节点,并以它作为根来建立一棵连接组播组中所有成员节点的单一共享树,而不是为每个源节点建立一棵组播树,这是一种有效的方法.由于核节点选择的不同将导致组播树的不同,核节点实际上决定着有核组播的性能.因此,核选择及核迁移是有核组播里的两个重要问题.针对全部到全部的组播模型,先给出了一个以总代价最优为目标的核选择完全算法,该算法复杂度较低;然后针对实时应用设计了两个核选择近似算法,并分析了它们的近似比;最后提出了一种可扩放的核迁移算法,该算法中的树代价估算机制为核节点的迁移时机提供了一个可信的评价标准.

       

      Abstract: The multicast network is one of the important and challenging problems that must be effectively resolved before multicast applications can be fully deployed on the Internet. Most existing multicast schemes focus on source based multicast trees. They either cannot be directly deployed in core based multicast trees or may render sub-optimal performance when directly deployed. Core based multicast selects a core node in the network, and then uses it to establish a single sharing tree covering all multicast group members, other than building trees for each source node, and so that it can save bandwidth enormously and enhance the data transmission efficiency. In core based multicast, core nodes in different positions will bring on different multicast trees, so the core node actually determines the performance of core based multicast. Therefore, core selection and migration are two important issues in core based multicast. In allusion to all-to-all multicast model, a complete core selection algorithm targeting at total cost optimization is given and the algorithm has low complexity; then two approximate algorithms are designed for real-time application and the analysis of approximation ratio is given; finally a scalable core migration algorithm is designed and the tree cost estimation mechanism in this algorithm provides a credible criterion for the migration time.

       

    /

    返回文章
    返回