The Research And Improvement Of Peer-to-peer Live Media Streaming System

The application of media streaming service is gradually changing people's way of life, and brings people more experience about information and entertainment. However, as the continuous expansion of network scale and the increase of the number of users, the traditional mode of C/S largely hinders the wide application of media streaming service because of the limitation of server's performance and bandwidth. The technology of distribution of media streaming using P2P technology not only effectively reduces the load of server, makes more users be able to receive media streaming service, but also improves the service capabilities of system and the utilization of network bandwidth by using the idle resources of all nodes in the network.This paper firstly studies the relevant theories of Peer-to-Peer live media streaming, analyzes several typical models of Peer-to-Peer live media streaming, compares the advantages and disadvantages of these models, and summarizes the advantages and problems of these models in network bandwidth utilization, latency, and reliability.Secondly, combining the "push" mode of tree model for its small delay in propagation and multicast of MDC and the advantage of partner management and message exchange mechanism in mesh model, this paper designs an adaptive and equal Peer-to-Peer live media streaming system called AElive. AElive uses the server to maintain the information of all nodes to guide the rapid establishment of logical relationship between nodes. AElive proposes the concept of "family" and "generation" in order to assign the appropriate parent nodes for each node to make the network structure relatively stable and resources fully utilized and smaller delay in the transmission of media streaming. AElive also designs a customizable adaptive data scheduling proposal that consistents with the actual bandwidth of the users so that more users can receive media streaming service and more users with low-bandwidth can participate in the distribution of media data. Effective node management, partner management and data scheduling policy makes AElive achieve better effect in bandwidth utilization, latency and reliability.Finally, this paper carries out AElive in the platform of NS2, then makes a simulation of the system under the given conditions. The experimental results show that the design of AElive is possible, and has advantages in bandwidth utilization, latency, reliability, efficiency and adaptability.