The rapid development of the Internet has changed almost every aspect of human life. In the services provides to people on the Internet, live video services with its timely, entertaining become so highly favored. At the beginning of video broadcast service development, people generally use centralized client/server architecture. Afterwards in order to address such issues as single points of bottleneck, IP multicast and P2 P technologies have been proposed. P2 P technology has been used in the wide range, because it’s high scalability, robustness and low cost advantages. By building overlapping network P2 P technology make the nodes share the resources and data. P2 P technology has not only succeeded in the video services but also in the distributed computing and storage.Firstly, we study and describe the live broadcasting system and its related technologies and then points out problems in the current research. With the development of mobile networks, network environment becomes increasingly complex. Clients of the P2 P overlay network also show the heterogeneity, mainly heterogeneous of clients are processing power, storage capacity, network access mode and network-bandwidth. How to service these heterogeneous clients at no additional cost? In P2 P networks, users who watch live services will provide data services to other users. The situation that users can join or leave the network anytime makes the dynamic of network very highly. How to reduce the impact of dynamics to other nodes on the network is also our study in this paper.Secondly, in order to solve the problem of dynamic and heterogeneous of network nodes, we construct a live video model that using Scalable Video Coding to transport video data, data-driven way to construct hybrid overlay network, multi-flow way to get video data. And this live video model are constructed on the basis of the existing research results and combined with the latest research trends.Subsequently, this article provides an overview of live broadcast model. Key problems and solutions are given, and give the overall framework of client as well as the introduction of the core modules. Moreover, we give the detailed design and implementation details of Scalable Video Coding live P2 P client.The Parent nodes selection strategy based on location information. The method of data acquisition, first get a base layer data broadcast in order to reduce the time that user has to wait. Then, dynamically request enhancement layer data according to its own network, and enhance the viewing experience.Finally, this article describes the test platforms of live P2 P client, as well as the functional and performance test. The result and data proved that it can properly perform the required functions and ensuring their efficiency. |