Research On Pyramid Incentive Embedded Peer Management Architecture For P2P Network

During the years of P2P application, Non-cooperation behavior is one of the major problems that threaten the P2P application systems, and it has become one of the hottest topics among P2P research. In the P2P application systems that base on peer cooperation, peer user may think altruistic cooperation and contribution to the network will cost his resources (computing power, storage capacity and network bandwidth) and will automatically decrease the performance. Hence come the non-cooperation behaviors against the"Free Cooperation". Although this kind of behaviors could not be simply regarded as"malicious behavior", it does influence the total performance of the P2P systems and may eventually cause the collapse of the whole P2P system. Therefore, motivating cooperation among users and restricting non-cooperation behaviors is one of the key topics to protect P2P application systems.By bringing in some social experiences into P2P research, this paper proposes a complete peer management architecture, with a novel and effective incentive mechanism embedded, to guarantee the security, stability and efficiency of the P2P system.The importance of this work mainly rests with providing an incentive embedded peer management architecture for P2P application systems, witch could not only effectively increase the degree of user cooperation among the whole network, but also restrict Free-Riding as well as malicious behaviors such as Whitewashing, Collusion and Traitor. Other value of the work is as follows:The proposed architecture bases on the extended Gnutella protocols and could be easily applied to the practical systems.The architecture applies quantilized trade-based credit scheme, which easily solve the problems like Asymmetry of Interest and Collusion that are difficulties for reputation-based schemes. Moreover, the monetary-like trust scheme may smooth the deployment of the future commercial P2P applications.The architecture first groups the peers, according to the underlying topology information, into CC (Candidate Community) and then groups the peers that are within the same CC into the IBC (Interest-Based Community). Moreover, optimized search strategies are used to decrease traffic load of critical links of the...