Incentive Compatible Mechanism Design And Implementation For P2P Systems

Peer-to-Peer (P2P) systems are receiving considerable interest in recent years. An important goal in P2P networks is that all peers provide resources, including computing power, bandwidth, storage space, and content. Most of the existing P2P networks are typically designed around the assumption that all peers are most willingly to contribute their resources and assist others. This assumption ignores the peers'ability to modify the behavior of an algorithm for self-interested reasons. However, as experience with P2P networks shows, rationality and selfishness are real issues in P2P networks. Selfish behaviors of peers may lead to serious problems of P2P network, such as free-riding and tragedy of commons.In order to solve these problems, there are increasing considerations on reputation system design in the study of P2P networks. The current work in the field can be roughly divided into two groups. One is reputation-based trust models that rely on the well known probabilistic estimation techniques use only a limited fraction of the available feedback. The other is game theory based incentive mechanisms that rely on aggregating the entire available feedback in the social network in hope achieving as much robustness against possible misbehavior as possible.In our work, we combined both techniques into a semi-structured mechanism which divide all the peers into 2 parts: super peers and normal peers. Super peers compose a structured P2P network using a micro-payment platform that we design to exchange services. While, normal peers compose an unstructured P2P network using a recommonand-based trust model.The micro-payment platform that we mentioned above has two layers: pricing layer and accounting layer.In pricing layer, we provide a general pricing mechanism that can maximize P2P networks'social welfare in a way of Vickrey-Clarke-Groves family, while assuming every peer in P2P networks is rational and selfish. This pricing mechanism has some desirable properties using an O(n) algorithm: (1) incentive compatibility, every peer truly report its connection type; (2) individually rationality; and (3) fully decentralized, we design a multiple-provider multiple consumer model, concerning about the service provider and service consumer individually.In accounting layer, we storage each peer's account information in several third party peers. These peers are chosed by a remote, distributed, and redundant way.Simulation results show the efficiency and robustness of our mechanism. This solution can deploy peers'resource efficiently with rational overhead. Additional, individual rationality is not necessary condition of our mechanism. This solution can deal with malicious behaviors well.