A New Credit-Based Incentive Mechanism for P2P Scheduling with User Modeling

Peer-to-peer (P2P) computing offers new research challenges in the domain of distributed computing. Encouraging peers to collaborate in file-sharing has been widely investigated. Incentive policies in P2P distributed computing systems is a new research field that requires specific policies to deal with resources mutability and the modeling of peers behavior. Unlike file-sharing, which uses such immutable resources as files, the mutable ones, such as CPU and Memory are the principal resources involved in P2P computing. Furthermore, the dynamics of peer participation are an inherent property of P2P systems and critical for design and evaluation. This paper presents a global P2P credit-based scheduling model that reflects P2P user dynamics, penalizes free-riders efficiently and participation of peers is also encouraged. Simulation results show the good performance and functionality of our proposal.