ABSTRACT

A Optimal Grid should have Three characterstics : It should not allow resource providers to increase prices continuously, No provider should reject to provide resources to certain type of jobs due to risk associated in executing them, Resource Providers and Consumers are allowed to take autonomous decisions. There will be no force on providers regarding the price for which they should offer their resources, but at the same time the price cannot be increased to any extent by providers because if no consumer is selecting him to offer his resources he should definitely get the price down. The two objectives identified are to schedule all type of jobs for execution and to minimize fairness deviation among resources. We present a scheduling scheme, which utilizes a peer-to-peer decentralized scheduling framework. Here the Greedy nature of provider is taken into account. we need not adjust competition degree explicitly , the provider automatically decreases his competition degree depending upon his capacity and number of jobs he is offered and automatically increases competition degree when jobs offered is more than his capacity. Here we give different levels to jobs to be executed depending upon the application type to which they belong and resource quota is given for different type of jobs while allocating resources by a provider so that starvation of particular type jobs can be avoided.