一种分布式多源应用层多播方法

为满足以分布交互仿真为代表的大规模多数据源多播应用的通信需求，重点研究了大规模多源应用层多播问题，提出了大规模多源应用层多播问题模型。参考基于P2P覆盖网络的应用层多播方法Scribe和Bayeux，设计了一种适用于大规模多数据源多播应用的分布式多源应用层多播方法DMALMM。DMALMM采用随机邻居结点选择机制，使用反向寻径方法构建数据分发树。仿真试验表明，DMALMM方法的结果优于采用Scribe方法的结果，是一种有效的分布式多源应用层多播方法。

A Distributed Multi-source Application Level Multicast Method

To satisfy the communication requirements of large scale many-to-many muhicast applications such as Distributed Interactive Simulation, The problem of large scale multi-source application level multicast is discussed and a large scale multi-source application level multicast model proposed. Using application level methods Scribe and Bayeux based on P2P overlay network for reference, a distributed multi-source application level multicast method （DMALMM） is devised. In DMALMM, random neighbor selection mechanism and reverse-path forwarding method are adopted to construct data delivery trees. Experimental results show that solutions obtained via DMALMM are better than solutions obtained via Scribe, thus showing DMALMM is an efficient distributed multi-source application level multicast method.