The European Particle Physics Laboratory, or Cern, has gone live with the first phase of a massive computing grid to process petabytes of data from its particle beam accelerator.

But multi-government-funded Cern, which does not have the resources to complete the project, is now looking for commercial IT vendors to help tackle the technical hurdles of processing the 12PB of data generated annually by its Large Hadron Collider (LHC).

In turn, Cern will encourage commercial IT vendors to use the LHC Computing Grid (LCG) project as a proving ground for their own grid computing technologies.

"The LHC is the world's biggest scientific instrument and the numbers are huge," François Grey, Cern Openlab development officer, told vnunet.com.

"But we are far away from where we need to be and we have to prove the technology. Industry sees grid as a set of disruptive technologies and it wants to see how it works."

The LHC is due to come on stream in 2007, to collide protons at high energy to probe the nature of matter. It will need compute-power equivalent to 70,000 PCs to analyse its output.

The LCG will become a worldwide computational grid service making use of the resources of scientific computing centres around the world. The first phase has just gone live with eight remote sites.

One overriding requirement for new systems is conformity to open standards as a form of future-proofing.

IBM has donated its TotalStorage file management system to handle the distributed data, including 28TB of high-end storage installed this week. TotalStorage uses a single virtual file naming to cover all data, irrespective of location or operating system type.

But both Cern and, in this case, IBM, need to know it will work within a grid scaling to what will be a larger storage environment than anything that exists today.

In the next two years the grid should expand to around 80 sites, incorporating a European grid of 70 locations, scaling to a further 200 sites including the US and Asia by the time the LHC is live.