An Agent-based Grouping Strategy for Federated Grid Computing

Abstract

Characterizing users based on their requirements and forming groups among providers accordingly to deliver them the stronger quality of service is a challenge for federated grid community Federated grid computing allows providers to behave cooperatively to ensure required utility by users Grouping grid providers under such an environment thus enhance the possibility of more jobs executed whereas a single provider or organization might not be able to do the same In this paper we propose an agent-based iterative Contract Net Protocol which supports in building federated grid via negotiating distributed providers The main focus of this paper is to minimize the number of iterations using a grouping mechanism Minimizing the number of iterations would produce less communication overhead which results in the minimum queue waiting time for users to publish their jobs Simulation results further ensure the feasibility of our approach in terms of profit and resource utilization compared to that of the traditional non-grouped market