Self-adaptive Schedule Mechanism for Peer-to-peer Multi-rate Live Streaming System

Traditional peer-to-peer live streaming systems usually have one specific streaming rate for one single channel. However, with the terminals diversifying, especially smart mobile phones bursting, the desire for multi-rates from one single channel topology of P2P live streaming becomes urgent, while considering that the network environments and topologies always change momentarily. High rate stream would make some terminals lost media data due to limited bandwidth and CPU etc., while low rate one would bring low quality data for terminals even if they have enough resources. To solve above problems, a new method called self-adaptive scheduling mechanism is proposed. A stream is divided into several sub-streams with different bit rates. These sub-streams are delivered by only one channel topology. Terminals with different capacities and bandwidths play different roles and try their best to deliver sub-streams while cooperating with other peers in the channel. Users can change the bit rates of channels they are watching. Moreover the bit rates also can change according to the bandwidths of the network adaptively. This mechanism is implemented for live streaming in real system, called MoSee. The experimental results prove that the efficiency of the mechanism is good.