원문정보
초록
영어
Since the resource of nodes can access and leave the Grid system freely, it is possible some malicious nodes exist in the Grid environment. In addition, the numbers and qualities of nodes’ resource can be changed dramatically and optionally, so these malicious nodes must be affecting the utilization of Grid resource. Hence, the trust management mechanisms are developed and widely used in the resource scheduling for Grid environments. However, most of the trust management models are mainly focus on the nodes historic behaviors and few of them consider the future trust of the node based on its previous activities. This paper proposes a new method to detect the nodes behaviors of resource providers in Grid environments based on D-S theory. Through simulating experiments, the new mechanism could record the behaviors exactly and prevent the malicious ones accessing the Grid system effectively. It would give a strong support for the success of resource scheduling in Grid environments.
목차
1. Introduction
2. Related work
3. Uncertain inference and its representation
3.1. Uncertain inference theory
3.2. Trust function and likelihood function
3.3. Update strategies
4. Trust detection mechanisms
5. Expectation trust benefit driven algorithm (ETBDA)
6. Simulation experiments and results
6.1. The rationality of influence factor
6.2. Comparison of the ability of detection
6.3. Comparison of two algorithms
7. Conclusion
References