Abstract
To be effectively adopted in different application domains, nextโgeneration Grids need to address different issues such as: an increasing complexity and distribution of applications; different goals, skills and habits of Grid users; availability of different programming and deployment models; heterogeneous capabilities and performances of access networks and devices. Moreover, scientific and commercial applications, as well as Grid middleware, will increasingly produce an overwhelming quantity of application and usage data. Although the ongoing convergence between Grids, Web Services, and the Semantic Web constitutes a milestone towards a serviceโoriented Grid architecture, which has the potential to face important issues such as application programming and business modelling, many other issues need research and development efforts. The great availability of data and information at the different layers of Grids, the maturity of data exploration techniques able to extract and synthesize knowledge, such as data mining, text summarization, semantic modelling, and knowledge management, and the demand for intelligent services in different phases of application life cycle are the driving forces towards novel knowledgeโbased Grid services. Guided by those considerations, the paper first introduces main requirements of nextโgeneration Grids and then describes some representative knowledgeโbased Grid services for both applications support and system management. Simple cases study showing how such services could be employed are discussed. Copyright ยฉ 2005 John Wiley & Sons, Ltd.