Research On Incentive Mechanism For Nodes Cooperation In Peer-to-Peer Networks

The Peer-to-Peer (P2P) network, which adopts the Peer-to-Peer mode, can make full use of the edge resources in the Internet and greatly improve the efficiency of the network. The application of P2P network technology has increasingly reflected its advantages and been widely used in file sharing, streaming media, computing, instant communication and so on. However, the widespread existence of free-riding actions will reduce the performance of the P2P network, and increase the vulnerability of the system. In addition, there are also malicious node behaviors such as White Washing, Sybil Attack and Collusion. Therefore, in order to guarantee that the P2P network can work effectively, safely and reliablely, it is necessary to take the suitable measurement to restrain the serious free-riding.Currently, many domestic and foreign scholars propose different theories for incentive mechanism, in view of massive existence of free-riding and malicious node phenomenon. In summary, it can be approximately divided into the following kinds of incentive mechanisms:virtual payment incentive mechanism, direct reciprocal incentive mechanism, reputation-based incentive mechanism, and selfless node incentive mechanism. As for current problems such as the free-riding of P2P file systems, the most effective analysis tool is the game theory. How to construct and quantify the incentive mechanism, and promote the cooperation among P2P network nodes are problems that need to be constantly studied. The effective incentive mechanism in P2P networks can reduce the ubiquitous phenomenon of free-riding, realize the steady-state operation of P2P networks, and thereby improve the Quality of Service (QoS) of P2P network applications. However, there is no universal and effective incentive method; hence a specific mechanism must be given according to the specific overlay structure and application type.The research work of this dissertation is under the support of Natural Science Foundation of Hubei Province in China (No.2010CDA011), Natural Science Foundation of Hubei Province in China (No.2008CDB342) and t Natural Science Foundation of Education Department of Hubei Province in China (No. D20111409), deeply studies BitTorrent, Gnutella, Chord and streaming media of P2P network, and proposes corresponding incentive mechanisms. The main contribution of this dissertation is listed as follows.1) This dissertation proposes the incentive strategy named AGTFT (Adaptive Generous Tit-for-Tat). The AGTFT strategy is improved based on the TFT strategy. If the node betrays the interactive node in the last game, then in this game, the node will select cooperation with a certain probability according to the environment dynamically. In the unstable network environment, the AGTFT strategy is more robust and has a better fault tolerance than the TFT strategy. And at the same time, the AGTFT strategy is friendlier to cooperative nodes than that of the TFT strategy, and promotes cooperation between the cooperative nodes. Hence, the AGTFT strategy inhibits the free-riding phenomenon directly and indirectly, and effectively improves the QoS of the BitTorrent system.2) On the basis of the Gnutella protocol, this dissertation proposes Gnutella incentive mechanism based on the interest cluster and the prestige improvement, which is mainly composed of interest clusters, and each cluster consists of a super node and some leaf nodes which managed by the super node. The super node in the interest cluster adopts the reputation incentive mechanism and evaluates the reputation of leaf nodes according to the service qualificaiton of the cluster nodes. And reputations among the interest clusters can be evaluated according to the service qualification of the interest clusters, so as to effectively restrain the free-riding behavior. The proposed mechanism can effectively prevent the flooding phenomenon, and has an obvious effect to restrain free-riding behavior.3) On the basis of the Chord protocol, this dissertation proposes the IChord protocol based on priority queuing service with incentive mechanism. The IChord protocol has characteristics including:the target of the IChord protocol is to incent the Chord network working in a steady state; the quantification is according to the contribution value of nodes in the IChord protocol; the incentive mechanism of DiffServ is in accordance with priority queuing, and dynamic DiffServ in the protocol. In the Chord network, service based on priority queuing has the incentive function, which can then restrain the free-riding behavior effectively, reduce Churn phenomenon in P2P networks and promote P2P networks to work in a steady state.4) By dynamically selecting a cooperative game for modeling nodes based on the Markov prediction mechanism, this dissertation proposes an incentive mechanism for P2P streaming media system based on dynamic alliance to form a stable alliance. In the streaming media system of P2P network, the alliance management server manages the union and its member in P2P streaming media network. Through issuing certificate and feedback of the union to judge whether a node provides streaming media download service to the downstream node, it can judge whether the node is a free-riding node. The proposed mechanism can effectively restrain the selfish behaviors of network nodes and can maintain the stability of the network.