A mobile peer to peer architecture for streaming media system was firstly proposed. According to the architecture, a serial scheduling and parallel scheduling algorithm were proposed for multi-video-sources. If quality of service monitored by the receiving-peer is degraded, the serial scheduling algorithm triggered the video source change event and synchronized the multi-video-sources by the time model of the streaming sequence. If the multi-video source concurrently sent the data to the receiving-peer, the parallel scheduling algorithm assigned the transmission task by frame-level bit assignment strategy. The algorithm firstly formulated the assignment problem to a nonlinearity programming problem and then transformed it into an integer programming problem based on the piece linear rate-distortion model. At last, the assignment problem was transformed into a general linear programming problem by relaxing the constraint condition. The algorithm was implemented by a greedy strategy and a piece approximate method was designed to map the solution between a linear programming problem and an integer linear programming problem. The experiment results show the algorithm is accuracy and valid and the distortion brought by the approximate method is neglected. The algorithm is very suited to a mobile P2P video streaming system.
展开▼