Research On A Grid Trust Model Based On Anti-Strategical Fraud Behavior

Grid is a new technology, Which is rapidly developing with the development of the Internet technology. It makes the complete sharing of resource come true and offers transparent superior services to the users. Along with emergence of grid technology, the user management isn't central, closed and controllable. Grid system is facing different kinds of threats include threats from the inner of system and threats from the outer of system. Traditional computer security mainly researches on preventing intrusion from the outer of system.As openness of the grid environment, whether to lager-scale parallel computing, or to information sharing and interoperability, there will be a wide range of security issues. And the Grid-domain management, heterogeneous, dynamic characteristics, making trust management different network environment. Research on Grid security has become the hotpot.The thesis starts from the basic background knowledge and analyses the characteristics and structures of Grid environment, which bring challenges to security and trust mechanism of the Grid. In this paper, the research status of the grid security is described; the typical grid security technology and related grid technology are elaborated. According to grid security characteristics, this paper proposes a new trust model to handle the trust relations between grid entities. After a close observation of all kinds of malicious nodes, we divided them into three categories: simple malicious behavior, concussive fraud behavior and corporate fraud behavior. This model proposes "Service Level" and "Recommend Trust" concepts, which can prevent signal or several grid entities' trick effective. The model divides the grid into different domains and advances the methods to compute and update the trust value between or in every domain. Through the judging of the service result given by the nodes, this model can update nodes' trust credits and there recommended trust credits. During the submitting process, nodes would chose the node that has the highest trust credits to do the job. In this way, we can greatly improve the task's success rate and reduce the overall damages made to the grid by malicious nodes.In the end, this paper devised a model experiment to testify this new model's reactions to various kinds of malicious behavior. The results of experiments and analysis show that this model is able to evaluate the trust relations between entities in grid more precisely, thus solving the safety problems existing in grid environment more effectively.