The world’s largest IT grid is due to start work in earnest tomorrow, as the $15bn (£7.6bn) Large Hadron Collider (LHC) in the Cern nuclear research facility near Geneva begins hurling protons and capturing, storing and sharing data on conditions at the birth of the universe.

To provide the computing power needed to sort and store this information, Cern’s IT department has created the LHC Computing Grid (LCG), which comprises 200,000 processors in 11 academic institutions around the world connected by optical fibre links.

Approximately 30,000 processors are located at the Cern centre itself, along with at least five million gigabytes of disk storage and 16 million gigabytes of tape storage capacity.

The Cern IT department is focused on handling the startup of the LHC by “improving the reliability and scalability of the LCG critical grid services”. It said the main challenge is to “focus on LCG-relevant work and provide better services with fewer resources”.

Once fully operational, the LCG will handle between 12 and 14 petabytes of data each year. In an average day, the LHC is expected to produce more than 40,000GB of usable information.

“We get data from the accelerator, which is stored to tape here at Cern and simultaneously a second copy is distributed to the 11 tier-one academic institutions around the world at 1.6Gbit/s. This is done via specially installed 10Gbit/s links,” said LCG project leader Dr Ian Bird.

“The tier-one institutions process it, join the dots and provide a representation of the raw data to tier-two academic institutions over the academic internet. This is the information most scientists accessing the data will use.”

Wolfgang von Rüden, head of the Physics Data Processing Group in Cern’s IT division, said highly reliable and stable IT services are needed throughout the grid.

Part of the final preparations was to co-ordinate large-scale “dress rehearsals” known as the Common Computing Readiness Challenge. In his report to staff earlier in the year, von Rüden said: “Problems are inevitable, we must focus on finding solutions as rapidly and smoothly as possible.”

The key to IT success at Cern is to support LHC “data taking”, which involves:

- distributed data management and analysis support;
- experiment integrations and “gridification” support;
- dashboards, monitoring, logging and reporting.

The IT department also provided all the support and hardware involved in the construction of the LHC. This involved handling 180,000 calls a year, 5,200 audio conferences and managing 18,000 mailboxes and 7,500 web sites.

Another huge job for the department is maintaining security ­ a challenge when the network is so dispersed. But there were no major outages due to computer security incidents in 2007 and the team reduced the number of compromised computers from 162 to 95.

Once the LHC is operational, the IT team will concentrate mainly on storing, processing and exporting the data, making sure information is stored as efficiently as possible and providing support to the various different sites.