Towards a service-based collaborative framework for data-intensive Grid applications

As data-intensive applications increase continuously in various domains, scientists nowadays need to save, retrieve and analyze rapidly increasing large datasets. The long latency of data transfer over Internet results in a serious challenge on ensuring high-performance access to large quantities of data. Furthermore, the vision of Grid researches is to provide a collaborative space where scientists can share their data, resources and experiences with others. These issues motivate us to develop a service based collaborative framework for data-intensive Grid applications. In this paper, we shall describe our initial work towards this goal. We plan to provide an integrated and virtual infrastructure, which manages distributed and diverse data resources and services. By means of Web Services technologies, we can realize the interoperability among various types of resources and integrate computing services with storage services to serve data-intensive applications. By making use of portals and data-intensive applications built on the top of the framework, users will be able to manage and analyze large quantities of data with high performance and share useful data with others in a way of forming virtual groups to complete cooperative work.