Design And Simulated Implementation Of Resource Trading Mechanism In Social Cloud Environment

The gradual maturity in technology and popularization in daily life of both cloud computing and social network provide firm foundation for the appearance and development of social cloud. As a virtual social application, the objective of social cloud is integrating resources of social network users and utilizing social relationship to realize resource sharing. As a novel paradigm of resources sharing, how to make full use of the superiority brought by social network and propose essential mechanisms to realize the goal that users can share and exchange resources fairly, conveniently and safely is one of the critical challenges to be solved in the research of social cloud.Taking the thought of cooperative exchanging model in microeconomics as the basis of sharing, a kind of social cloud resource trading mechanism composed of an auction based public trading model and a negotiation and contract based private trading model is proposed in the paper. Utilizing the convenience for communication of social network, the private trading model is presented based on bargaining model as a subsidiary trading way to deal with peculiar situation. Public trading model as the primary trading method can be divided into four functional modules, i.e., auction protocol, bidding strategy, winner determination, auxiliary mechanisms and systems. First, merging continuous auction and call auction together, adding the property of combinatorial auction, as the critical part of trading mechanism, an improved double combinatorial auction protocol is the critical part for controlling the trading process. Second, GD-FL bidding strategy is applied for user agents making scientific decision in the dynamic and changeful market environment, which combines GD (Gjerstad Dickhaut) and FL (Fuzzy Logic) in addition to some essential changes for the adaption to protocol. Then, winner determination becomes the final step in auction for the feature of call auction, but it has proven to be an NPC (Non-deterministic Polynomial Complete) problem in combinatorial auction, due to the limitations of traditional algorithms, an extended krill herd optimization algorithm with the application of genetic operators is introduced to derive the solution. Last, auxiliary mechanisms and systems mainly including social relationship topology system, mutual feedback evaluation based QoE (Quality of Experience) system, RBF (Radical Basis Function) neural network based user recognition system, user contribution system and social currency incentives are designed and proposed to form a complete trading mechanism for users fairly, safely and conveniently sharing resources.The proposed trading mechanism has been simulated, and then performance comparison is fairly conducted between it and the benchmark on several indexes, in addition, improvement analysis of related algorithms and mechanisms is performed. Simulation and evaluation results have demonstrated the feasibility and effectiveness of the proposed trading mechanism.