In recent years, streaming media applications based on peer-to-peer technology have become a very important application in network new-media field for its scalability, easy deployment, low-cost features and Etc. It's a very hot spot in the industrial and academic circles.The existence of node heterogeneity and dynamic nature in P2P-based streaming systems bring great challenges on how to make full use of node resource to reduce media-server load ; various channels in P2P streaming systems have different popularities,which cause the imbalance of channel service capability and waste of peer bandwidth resources, weaken the system scalability and increase the deployment cost.In this thesis, we focus on research in peer service capability which is related with how to make full use of peer bandwidth resources to reduce server load. The main research contributions of this thesis include:Firstly, we propose a peer service capability evaluation method based on historical information in data scheduling process, designed the data scheduling and service scheduling strategy based on peer service capability for P2P Live streaming. The data scheduling and service scheduling strategy can improve system Quality of Service (QoS), reduce startup delay and server load.Secondly, we design and implement the data scheduling algorithm based peer service capability for P2P VoD system, this algorithm can make full use of peer bandwidth resources and reduce servers load.Thirdly, we propose a peer service capability self-aware method based on peer's feedback, through the co-ordination with data scheduling, this method can get peer service capability accurately to support cross-channel collaboration. |