Research And Implementation Of Node Selection Mechanism In P2P Streaming System

Because of the large traffic data that produced while the streaming media is transporting in the network, the traditional streaming media system that based on the C/S mode bring a lot of pressure for the server and bring about the question of single point failure. P2P streaming media system make the service users involved in it as well as the service provider.P2P model will be applied to streaming media system to reduce the pressure on the sever side an ease the bottleneck problem of the server., That means, the client in streaming media system which uses P2P technology are forwarded to the other client streaming media data while receiving streaming media data. In P2P streaming media system, node selection is a key technology. Good node selection mechanism can provide a steady stream of data transfer rate to ensure that the streaming media file playback fluency and reduce the pressure of network traffic on the backbone of overlay network and provide more stable quality of service.On the basis of analyzing the problems of the past node selection algorithm, this paper presents a new node selection mechanism that base on the IP database and node capacity. At the server, we aim to solve the topology mismatching problem between underlying physical network and application layer of streaming media system and present node selection mechanism that based on the IP database. At the requesting peer, we use the available bandwidth, packet loss rate and network delay between requesting node and serving node to present node selection mechanism based on the capability of serving node.By analyzing the structure and code components of streaming media system of P2PCenter, due to the deficiencies of the node selection in this system, we analyze and discusse the feasibility to use the node selection mechanism of base on the IP database and node capacity on this system.Simulation tools of PeerSim is used for authentication for the node selection mechanism proposed in this paper. Aggregate flow rate and overlay network pressure are used as two standards to compare the node selection proposed in this paper with the simple random node selection strategy and node selection that based on the node's ability. The comparing figure shows the node selection proposed in this paper has an obvious advantage in the reduction of overlay pressure and the increase in aggregate flow rate.