A Study On P2P Media Streaming Network Topolopy

With the rapid development of Internet broadband technology, IPV6and3G technologies,media streaming has become an important task for Internet. The low consumption and highscalability of the peer-to-peer media streaming system is a good solution to solve the highpressure of the traditional media streaming server systems and networks, and make it possibleto provide users with high-quality online video services. In the random peer-to-peer mediastreaming system, it will exit high pressures on the backbone network, long startup delay andplay delay and a big frame loss rate because of the random and uncertain neighbors’ selection.For the high pressure of backbone network, a long startup delay and a big frame loss rateof the random peer-to-peer network, we propose a neighbors’ selection algorithm based onRTT and bitmap. The algorithm uses the Tracker to store all the RTT and bitmap informationof nodes in the network. When a peer requests a new neighbors’ list, the Tracker uses the RTT,bitmap information and screening function to generate a best neighbors’ list for the peer.Through the simulation experiments on OMNET++platform, it shows that the neighbors’selection algorithm based on RTT and bitmap has a smaller startup delay and playback delayand a low frames loss rate and the average number of hops compared with the randomneighbors selection algorithm.And we use the RTT information between nodes to optimize the CoolStreamingalgorithm and propose an optimized CoolStreaming algorithm based on the RTT information.The algorithm takes the RTT information in consider when choosing a new parent. If thenumber of partners that satisfy the CoolStreaming restrictions is more than one, the peer willget the RTT from these partners and choose the node as a new parent node which has thelowest RTT between the requesting nodes. Through the simulation experiments onOMNET++platform, it shows that the optimized CoolStreaming algorithm based on RTTdoesn’t improve the startup delay well compared with the CoolStreaming algorithm and therandom algorithm, but in the frame loss rate, it achieves very significant improvement, whichgreatly reduces the frame loss rate and the average number of hops.Finally, we propose an incentive P2P streaming mechanism based on the users’ grade.We classify the nodes by their grade, and build a hierarchical P2P topology according to theirgrade when nodes join. The node with high grade will be located in the layer near the sourceserver and it is possible to obtain the high quality service. Then it will energize the othernodes in the system. Through the simulation experiments on OMNET++platform, it showsthat the nodes with high grade in the system have low startup delay and play delay, andincentive P2P streaming mechanism based on the users’ grade has low startup delay and play delay compared with the random neighbors’ selection algorithm.