The world-wide-web changed the way we share and distribute information but has done very little to make computing power or data storage more assessable. So the aim of the Grid is to build on existing Internet protocols and to develop `middleware` which will allow simple and transparent use of resources wherever they may be world wide. We will also need to change our applications so that they can take advantage of the Grid infrastructure and run efficiently in this complex environment.

There are many challenges to developing a Grid that will deliver the kind of robust, high-performance system required. Computing in the LHC era, for instance, will require computing clusters with tens of thousands of nodes, and each experiment will accumulate data at a rate of about one million gigabytes per year. To cope with this scale of computing and data, experiments will have to put globally distributed resources at the physicists’ fingertips. Particle physicists are therefore heavily involved in providing requirements for the Grid, in developing higher levels of the middleware, and in providing a real-world use case for the early deployment of software. They are working with researchers in computer science and many other fields, often pursuing novel solutions to these awesome challenges.

The Oxford E-Research Centre web site has a short but more detailed introduction.

For the big picture and overview please see The Anatomy of the Grid: Enabling Scalable Virtual Organizations