CERN Accelerating science

The European DataGrid project (EDG) is now entering its third year of activities. Thanks to the achievements in 2002, it successfully passed its second year review at the beginning of February, when a team of four European Union experts came to CERN for two days to carefully evaluate the latest progress of the project. And progress was, indeed, quite considerable.

A series of successively more sophisticated and reliable versions of Grid middleware, the software that enables computers to work together in a transparent manner, has been deployed on a production style testbed covering major European sites. The three application groups designated to test the system - high energy physics, earth observations and bio-informatics - have been given access to the testbed in a continuous manner through-out the year. They have tested their applications and explored the system's various possibilities, and their feedback has been used to improve the software and define the priorities for the development work.

Within the physics application group, task forces for the ATLAS and CMS experiments were set up to determine whether the existing system could be used at production level. Initial results have been successful and helped to identify a number of improvements that can be made to the DataGrid software and the way the experiments applications are managed. This work allowed improvements to be identified and will serve as a basis for the LHC Computing Grid (LCG). The pilot project LCG-0 is based in fact on the current release of the EDG software, and the next release of the software (version 2.0, being built now) will be used for the LCG-1 initial deployment.

The collaboration with DANTE and EU DataTAG projects has been particularly successful. Numerous National research and Education Networks have been upgraded and inter-connected with the European GEANT network. In particular, the testbed sites have been linked by high-speed networks, enabling the project to face the users needs and perform advanced networking functionality investigations. On several links, high throughput has been demonstrated as well as network Quality Of Service benefits. The collaboration with DataTAG on the GLUE (Grid Laboratory Uniform Environment) activities has increased the level of inter-operability between European and US testbeds for the benefit of all users.

DataGrid has also made important contributions to international standardization activities (e.g. GGF), as demonstrated by the significant number of members involved in organizational activities (e.g. chairing working groups) and the number of papers and presentations (see EDG website for the list of publication).

To allow the test users to make use of the prototype grid, the DataGrid team has developed a "DataGrid tutorial", a two days event which includes practical hands on exercises where participants connect to production testbed. During 2002 the tutorials have been repeated in 9 sessions held in 7 different cities for approximately 200 people. This activity has turned out to be a major success and an important instrument to train a wider community of users to the use of the project middleware and testbed and also to learn Grid technology in a more general sense.

The project is now ready to get on with its third and last year, which foresees an extension of the testbed to more sites and a new major release of the middleware software, intended to provide new functionality and higher levels of reliability and scalability.