Trust Audit On Grid Resources

As the rapid development of computer and network technology, the gridcomputing which is mainly towards the wide-area resource sharing and collaborationis greatly enhanced in computer science. The realization of large-scale resourcessharing and collaboration in grid computing is built on the basis of grid security. GridSecurity Infrastructure ensures the identity trust of the grid environment, but trust isnot enough emphasis on behavior. To some extent the emergence of trust model makeup the security problems which is caused by the dynamic and uncertainty of gridenvironment, because of the subjectivity and randomness of trust evaluation, there arestill unreasonable conclusions on it. The subjective evaluation which is distorted andmalicious can be reduced by the audit of trust. In this paper, the concept of trust auditon grid resources and a model of trust audit on grid resources are introduced, they areboth based on the traditional trust model in grid environment. The impact of distortedand malicious evaluation is reduced and the trust mechanism is further improved. Themain work includes in this paper:(1) The major contents in this paper are all based on the traditional trust model,and the concepts of trust and trust evaluation are further understand, and thetraditional trust model is analyzed and compared.(2) The application of the audit system in the field of economy and computer isanalyzed, the concept of trust audit on grid resources is introduced, and an audit dataacquisition method is proposed and normalized.(3) A trust audit model on grid resources is introduced, mainly including twosub-modules which are exception judgment and exception analysis to the trustevaluation values. In the exception judgment module, an exception judgment methodis proposed which is based on trust cloud, it can determine the exception evaluationvalues. In the exception analysis module, an exception analysis method is proposedwhich is based on the QoS parameters of grid resource. The exception values arereviewed on three aspects, which are resources, real-time network and users. Then thereason for the exceptions is found. At last, the audit feedback to trust evaluating modeis provided and an audit algorithm is described based on the model. The model reduces adverse effects on users'unfair exception evaluation, the trustevaluation value on resources more accurately reflect the quality of services so thatusers can make the right choice. The experiments show that the audit model canimprove the utility rate of resources, ensure the success rate of scheduling, reduce theaverage completion time of task, and improve the stability of system.