Encouraging Peer Cooperation In Mesh-pull P2p Streaming Systems Using Distributed Incentive Mechanisms

Recently peer-to-peer (P2P) systems have appeared to be the driving force to deliver streaming media on the Internet by offering attractive benefits, including robustness, reduced cost and scalability. Nevertheless, these systems have also been facing various security threats. In this thesis, peer cheating problem in P2P live streaming systems is studied and addressed. Legislative peers may starve for bandwidth, if selfish peers do not contribute bandwidth to other peers or if selfish peers request more than their fair share of bandwidth, even malicious peers intentionally drain all the available bandwidth in the system.To alleviate peer cheating behaviors, a distributed incentive mechanism is designed desirably with low complexity for mesh-pull live streaming systems. This incentive mechanism is implemented using a trustworthy distributed service-differentiation incentive scheme with low complexity. In the incentive scheme, peers'download rates are fairly differentiated according to the amount of useful upload bandwidth which they contribute for the system, thus peers are encouraged to contribute bandwidth to other peers in order to ensure the streaming quality-of-service.Compared with other schemes, the proposed incentive scheme has two properties: 1) the contribution of each peer is evaluated by considering the features of mesh-pull live streaming, not just by the amount of the transferred data alone. Both delay of the chunks from the serving peer to the requesting peer and the upload bandwidth are considered when evaluating peer contribution. 2) The potential cheating behaviors of peers are effectively discouraged. Peers may add or subtract the contribution value of a given peer based on the cheating identification during the peer contribution evaluation-process and such a peer hinders the desired performance of the entire system. In this thesis the above two issues are addressed in order to enhance the performance of the proposed incentive mechanism.The proposed incentive mechanism is evaluated in a simulation study by enhancing the simulator P2PTVSim with the proposed incentive mechanism. The simulation results demonstrate that the proposed distributed incentive scheme effectively encourages peer cooperation and the streaming quality can be maintained even though the cheating behaviors are severe (e.g. the percentage of the cheat peers up to 30%).