A self-configuration model for execution management in Grid computing

This paper presents a model for self-configuration of execution management in Grid environment. This model makes it unnecessary to use brokers. No broker in large systems like Grid has many advantages such as elimination of point of failures for the system, increasing accessibility, and reducing the complexity of brokers' communication. Networks formed logically between Grid's nodes have been used to present this model. Using this model, nodes able to implement submitted jobs to the system are found without using a central node autonomously by questioning neighbor nodes. This model also can improve function during runtime when submitted jobs to a node are almost similar to the former jobs. In the end of the paper, presented model is simulated by Netlogo multi-agent environment and using given scenario. Conclusions show function improvement during runtime.