Trust enabled CARE resource broker

Abstract

The objective of the grid system is to allow heterogeneous geographically distributed computational resources to be shared, coordinated and utilized for solving large scale problems. Grid users are able to submit and execute tasks on remotely available grid resources. Resource management is a vital part of a grid computing system. The notion of trust is used in resource management of grid computing system. In this work, we apply behavioral trust to the problem of resource selection. Through CARE Resource Broker (CRB), we demonstrate that our behavioral trust scheduling scheme significantly maximize the throughput while maintain the high success rate of task execution. We also analyze results of our behavioral trust based scheduling and rank based scheduling.