Cern scientists build the web of the future

There are two notable things about Cern, the European Organisation for Nuclear Research.

As the world's leading particle physics laboratory, it is staffed by people with very large brains.

It is also where Tim Berners-Lee, a former scientist at Cern's Geneva headquarters, invented the worldwide web as a way to share information among the global research community that contributes to the organisation's work.

So when the people with very large brains tell you they are undertaking an experiment so complex that they need to build the next generation of the web to support it, the IT industry should sit up and take notice.

Cern is a funded jointly by 20 European countries, with 3000 staff supporting 6500 researchers in 35 nations.

The web was created in 1989/90 when Berners-Lee devised a way to share information between the computers used by Cern's scientific community. The first browser was developed and the web as we know it was born.

But Cern has moved on since then and its next experiment, scheduled for 2007, will drive new developments in technology and manufacturing that will benefit all of industry.

The laboratory is building the Large Hadron Collider (LHC) - a 27km circumference tube 100m underground that will accelerate beams of particles to near-light speed, then crash them into each other to study what happens.

The aim of the project is to simulate the events taking place one millionth of a millionth of a second after the universe was created - information that could revolutionise our understanding of how the natural world works.

The LHC will be the world's largest scientific instrument, says Derek Mathieson, deputy group leader of internet development services at Cern.

Building such a device will challenge the manufacturing industry to develop new processes to produce the precision components needed.

'The steel collar for the accelerator is made from a special steel in Japan, and cut in Europe to one-fiftieth of a millimetre precision - and we need six million of them to build the LHC,' said Mathieson.

The challenge for the technology industry will be just as great.

Scientists will have to analyse vast amounts of information - two petabytes of data will be generated every second. Not all of this is needed - but 10 petabytes will be retained every year during the 10-year project, which would require the power of 100,000 of today's fastest PCs to process.

To share and manage this data Cern is building the LHC Computing Grid - and contributing to the development of grid technologies that will be commercialised by IT suppliers to build the future internet that every consumer and business will one day use.

Grids are distributed networks of computers that share resources such as storage and processing power, making use of spare capacity to create a single, virtual system. In the grid-oriented future of the web, all applications will run in the network, rather than on corporate servers that are connected to the internet. Grids could act as the basis for computing utilities, with processing power provided through a socket in the wall and charged for in the same way as electricity.

Leanne Guy, a section leader in Cern's LCG team, says the work will have commercial benefits, because it is part of the European DataGrid project, a European Union initiative to create a grid for collaborative research. DataGrid software is already installed on hundreds of systems throughout the world.

To make sure the results can be widely used, developments are done using both commercial and open-source software - Cern uses Oracle's 9i database and application server, as well as open-source equivalents MySQL and Tomcat.

'We are required to provide an open-source alternative for all our developments in grid data management for research institutes that are smaller and can't afford the investment in commercial software and skills,' said Guy.

The first release of LCG is scheduled for June, and Guy hopes to have a 'real production system' in place by the end of next year.

It will be several years before the next generation web becomes an everyday business tool - but the work of the people with very large brains will inevitably bring another huge advance for the IT industry.

Cern : The mind-boggling facts

Cern is building the Large Hadron Collider (LHC) particle accelerator to study subatomic matter and the forces that hold it together.

* The LHC is a 27km ring being built 100m beneath Switzerland and France. It will be live in 2007

* The ring will be cooled to -271C, just 2C above absolute zero

* Particles will circle around the LHC 1000 times per second

* The collision between particle beams will simulate what happened in the first millionth of a millionth of a second after the Big Bang

* Particles will collide 800 million times a second

* Two petabytes of data will be generated every second - a petabyte is 1,000,000,000,000,000 bytes