Performance Evaluation using Nimrod/G: A Grid Resource Broker

Abstract

Grid computing is the next generation IT infrastructure that promises to transform the way organizations and individuals compute, communicate and collaborate. It offers untapped processing cycles from networks of computers spanning vast geographical boundaries. Grid resource broker is responsible for resource discovery, resource selection, binding of software, data, and hardware resources, initiating computations, adapting to the changes in Grid resources, and presenting the Grid to the user as a single, unified resource. In general, a brokering system of Grid computing environments aims at delivering better performance. Desirable performance goals of Grid brokering include: maximizing system throughput, maximizing resource utilization, minimizing the execution time and fulfilling economical constraints. In this thesis work, we have simulated the environment of Nimrod/G Resource Broker: an Economic based Grid Scheduler, which manages all operations associated with remote execution including resource discovery, trading, scheduling based on economic principles and a user defined quality of service requirement. The Nimrod-G resource broker is implemented by the services provide by Globus. Performance evaluation has been done in order to demonstrate how the performance is affected if by varying the number of jobs submitted to Nimrod/g or by varying the number of computing resources available in Grid environment.