The Study Of Techniques In Grid Market Model Based On Organic Computing

In the past several decades, computer industry mainly focused their attention on how to improve performances of hardware and how to enrich functions of software, which has made the computing system more and more complex to maintain and manage efficiently. Consequently, developers and users of complex computing system are confronting with great challenges.In order to solve this problem, utilizing intelligent techniques to implement self-x characters of complex system has become a hot subject at present. Organic Computing, as a new concept and a typical intelligent method, grants computing systems with some organic characters to decrease the complexity of computing system management. These characters include self-organization, self-configuration, self-repairing, self-adaptation and etc. Developers just only need to determine the goal of the system management. The inner-mechanism will give the system a certain degree of freedom and react to new situations intelligently.As a new network computing platform after the World Wide Web, grid computing system provides the infrastructure for users to share resources within a complex distributed environment comprehensively. However, with the characteristics of large-scale, distributed, heterogeneous, and dynamic of the grid computing resources, grid computing system has become an extremely complex system. This paper points out the shortage of grid market model after analyzing its usage in traditional grid resource management.Using the idea of Organic Computing in the design of grid system, this paper designs an OC model for the management of grid resources. Under the framework of this model, we present the structure and functions of each part. Further more, we establish grid market trust mechanism by analyzing trust degree of resources, and detect the malicious resources in grid to improve the robust and usability of grid computing system.Finally, this paper tests the trust mechanism on a simulating platform. The results demonstrate that this mechanism can effectively increase the success rate of task allocation inthe circumstance with malicious resources.