Research On A Streaming Media Supported Application Layer Multicast Strategy

With the development of streaming media and peer-to-peer technology, the technology of P2P streaming media is presented. Compared with traditional streaming service employing a client-server unicast service model, P2P streaming media technology has huge potential on resource usage and system scalability. But it swallows large amounts of network bandwidth and brings out tremendous network flow. Futhermore, there are many problems in current P2P streaming media systems, such as big delay and redundancy.It can solve these problems to some extent by construction of the topology-aware overlay network and efficient application layer multicast (ALM). It aims to match the overlay network and the physical network by construction of the overlay network topology awareness. The current construction methods of overlay network topology awareness achieve the goals to a certain extent. But they are complex and of poor performance. The key of ALM is how to establish highly efficient data transmission paths and distribute the media data to all nodes in the group along these paths.Through thorough analysis of the physical network topology and current schemes of building overlay network topology awareness, the paper presents a new scheme which combines ISP (Internet Service Provider) and the delay between nodes. It partitions all nodes into different group by ISP, and then every peer periodly sends to some peers to find lower delay neighbors. The scheme is not complex but can match the overlay network and the physical network well.After researching on the gossip-based ALMs, combining the previously presented construction cheme of overlay network topology awareness, the paper puts forward an improved ALM TP2PM. TP2PM adopts some quick recovery arithmetics against P2P network high dynamic.Finally, experiments are done to validate TP2PM. The experimental results show that TP2PM can effectively reduce redundant network traffic and delay. Meanwhile the system is greatly improved in the robustness and of high performance.