Research On Reputation-based Trust Management Mechanism In Cloud Computing

With the development of the grid computing and the distributed computing,cloud computing has become a popular service model with many providers and consumers swarming into the cloud computing market.The trust relationship between transaction entities become the main obstacle in cloud computing development.One of methods to solve the problem is the reputation-based trust management.Currently,the main method of calculating reputation value is to put the deviation value between each feedback with the majority as the feedback's weight in order to eliminate subjective factors and then obtaining a relatively objective reputation value of a cloud service.However,the method has following deficiencies: firstly,the reputation value of cloud services cannot reflect requesters' subjective preferences,and secondly,the reputation value cannot change dynamically with the cloud services' QoS.Aimed at these shortcomings,this paper concentrates on the research on reputation-based trust management mechanism.The main work of this paper is as follows: firstly,in order to reflect the subjective preferences of requesters,we propose a subjective reputation value-based trust management framework,and also the reputation value change along the time.While calculating the reputation value,we adopt the decision tree algorithm to predict the similarity of preferences among the requester and consumers as the subjective factor,according to consumers' background attributes.Secondly,we adopt the time-decay function to calculate the impact on the reputation value of feedbacks as the time-decay factor.Thirdly,we combine the result of subjective factor with time-decay factor to calculate the reputation value of the cloud service.Finally,we conducted the simulation experiment which illustrated that the trust management framwork we proposed could provide the reputation value of the cloud service which can reflect the requestor's subjective preference and the reputation value cloud update with the cloud service's QoS value in time.