Grid to supersede internet

Alison Goddard reports on a powerful successor to the net being developed by British scientists.

British scientists are developing a successor to the internet that will give researchers access not just to information but to raw computing power.

Plugging your computer into the grid - as this next-generation internet has been dubbed - will allow you to find seamlessly the powerful answers to your queries. The fact that you have combined data from the United States and Japan and processed it on a supercomputer in Germany will be concealed behind the scenes.

The grid is the highest priority for particle physicists and astronomers in the current spending review. The Treasury is considering a bid for an extra Pounds 8.4 million over the next three years to establish a United Kingdom prototype of the grid needed to handle the data from future experiments at Cern, the European particle physics laboratory.

In the United States, Nasa is spending some $60 million a year on developing the grid in collaboration with the national laboratories and universities. The members of Cern, where the worldwide web was invented, are preparing a proposal for funding from the European Union's Framework 5 programme for a European grid.

Particle physics is driving the grid's development. Experiments to identify how a point-like particle such as an electron can have mass are due to start in 2005, on the Large Hadron Collider at Cern. The LHC experiments will produce so much data that a new way of handling the information will be needed.

"Particle physics and arguably the scientific world is facing the biggest computing challenge ever seen. Some 5,000 physicists from 500 countries will collaborate on the LHC, and the only way to meet their computational requirements is to use the grid. The grid makes it possible for physicists based in their home institutes to assert analysis priorities, maintain local control and leverage resources," said Paul Jeffreys, who is leading work on the grid at the Rutherford Appleton Laboratory.

Jim Sadlier, of the Particle Physics and Astronomy Research Council, added:

"The grid won't be developed by a single scientific community. It is too big for particle physicists and, if you are looking for general applications, you will need people from computer science and industry."

Dr Jeffreys added: "It is essential that we have industry working with us from the outset, so we can ensure that the software developed can be used as widely as possible."

Particle physics is not alone in facing a very substantial increase in the size of data samples - the work in this area will spin-off to help many other disciplines.

Julia Goodfellow, vice-master of Birkbeck College, London, said: "There is a tremendous opportunity for biologists to use the grid. Biology is a large, diverse community with many areas that could benefit, especially when you consider the exponential increase in data coming from genetic sequencing and functional genomics. We need to set up ways of handling and analysing that data."

Professor Goodfellow has identified at least seven areas that could benefit, including neuroinformatics, population biology and computer modelling. Groups of biologists will identify the IT problems that are specific to these subject areas and develop plans to solve them. Environmental scientists are also involved in drawing up plans for the grid.

Meanwhile the scientists and policy-makers will be fleshing out the details of who pays for what in a system where one could potentially get expensive computer time almost for free. "It is all very well saying you are going to grab data from around the world, but we are going to need rules about who has access and how it is paid for. As well as real technical challenges, there are policy rules that will underlie access and its cost. We have only just started the debate," said Mr Sadlier.