A PROACTIVE AND CENTRALIZED AGENT-BASED SOLUTION FOR SERVICE DISCOVERY IN GRID

Abstract

The introduction of Grid Computing technologies has opened new ways in the computing field by allowing the collaborative usage of computational resources across the world. This situation has enabled to expand the computational barriers of organizations, and specially research centers, by performing time consuming executions on distributed resources from abroad. In fact, the Grid vision aims at delivering computational performance efficiently to a large community of users with such huge computing requirements. The overall aim of the Service Discovery is to efficiently schedule applications that need to utilize the available services or resources in the metacomputing environment. An agent is a representative of a local grid resource and is considered to be both a service provider and a service requestor. In this work "A Proactive and Centralized Agent-Based Solution for Service Discovery in Grid", a novel centralized framework is proposed in which mobile agents are used to discover the available resources in Grid and keep the information as fresh as possible. The proposed Centralized framework is shown to have important advantages over most of the existing Centralized and Decentralized Service Discovery mechanisms. Also, a new, itinerary scheme for mobile agents "Forking Itinerary" is proposed which has several advantages over the popular schemes, viz. Sequential and Parallel. Simply, maintaining grid-wide information about availability of services is useless, if it cannot be retrieved efficiently. A novel tree-based data structure is suggested for this purpose. The work has been carried out on grids created with Cougaar as the platform