| The issue of resource sharing in distributed systems is raised since the appearance of computer networks. It is one of the oldest research topics in computer science. However, the diversity of distributed systems and their inherent characteristics have changed significantly in the last 30 years. First, the scale and openness of the resource sharing platform has increased considerably. A typical example is the Internet, the major paradigm of distributed systems. Moreover, many new paradigms of sharing resources has emerged, such as peer-to-peer (P2P) computing and Grid computing. These features make the investigation of effective mechanisms for resource sharing in open computing communities much more challenging than before. How to coordinate participators into a healthy and robust community to effectively sharing resources is the core of the research on the resource sharing in distributed environments.;In this dissertation, a general trust based economic model is proposed to attack this problem. This approach includes two components. One is a personalized trust model, and the other component is a multiple currency based economic model. We make our approach general enough so that it can be applied to different systems. The design space of trust model is first evaluated comprehensively from both the angles of trust model design related factors and trust model independent related factors. Implied by the evaluation results, we augment the trust model by introducing flexible adaptation. To validate our approach, we have chosen three two application scenarios as the case studies which include ISP Peering and Grid Based Resource Scheduling. The experiment results show that our model can improve system's performance, robustness, and efficiency comparing the classical approaches. |
