Adaptation Engine: an Agent-Based Framework for ad-hoc Service Life-Cycle Management for Meta-Computing

The cluster and grid-computing paradigm offers support for large scale, widely distributed, highperformance computational systems. Several of such architectures and frameworks have been developed aimed at primarily large parallel computations in support of scientific, engineering calculations and problem solving. With the emergence of the “service-oriented” business pattern for low-cost e-business, and increasing users’ need for high-volumes of multi-media contents and applications, we argue that the rapidly maturing grid technology will offer commercial opportunities for the development of a range of systems, infrastructures and services to support high-utilisation and availability of global computing and data resources for widely distributed enterprises. However, before this becomes a reality much research work is required to address a number of well documented grid issues including; ubiquity, high-assurance, flexibility, ease of use and alignment with emerging industrial standards. In this paper we argue for the potential of new distributed systems, aimed at ebusiness and commercial enterprises, leveraging the strands of research from cluster computing, grid computing, agent systems and data description languages. Based on an on-going research work, which focuses on high-assurance composable systems engineering, this paper will outline a service-oriented approach and associated agent description languages, which are used to facilitate the construction and management of ad-hoc federated software services.