Agent-based service management in large datacentres and grids

Abstract

Increased computational demands and data mining present the IT world with new challenges. Amongst others, the maturing grid technology aims to address them. To take full advantage of the grid capabilities and enhance its effectiveness in complex and dynamic computational environments, we must make service management more stable, less computationally expensive and more automatic. In this paper, we propose a synthetic approach to deal with service management in large Unix datacentres that involves the employment of intelligent agents and ontologies. These agents can automatically detect and correct faults at run-time and manage services.