New frontiers for grid applications

Abstract

Grid research, rooted in distributed and high performance computing, started in late 90's when scientists around the world acknowledged the need to establish an infrastructure to support their collaborative research on compute and data intensive experiments. Soon afterwards, national and international research and development authorities realised the importance of the grid and gave it a primary position on their R&D agenda. The importance of the Grid was translated into large funding, from various national and international sources, channelled to various grid projects around the world aiming at building the socalled global infrastructure for eScience. Selected key projects, such as EGEE and Globus, play a key role in developing this infrastructure. This paper reviews the grid as a suite of concept, standards, programming paradigm, and tools intended to support compute and data intensive tasks in various application domains. The future of grid applications is then discussed in light of the major achievements and the future plans of key grid projects. The paper concludes by arguing for a metagrid infrastructure offering a superset of functionalities on the top of existing grid services and placing the user in a privileged position that grants him/her universal accessibility to the grid.