The computing grid built to carry data from the Large Hadron Collider to scientists around the world is also being used to speed the development of life-saving drugs and uncover the causes of climate change, people involved in the project said Friday.
Scientists at the European Organisation for Nuclear Research, known by its French acronym, CERN, joined their colleagues in North America and Asia via live video-link on Friday, saying that the grid is ready for action when particle physics experiments resume at CERN next May.
But they emphasised that the grid is already being used for other projects. One-third of the experiments being conducted today are in non-physics areas, including life sciences and biomedical engineering. They said the grid is being used for simulations that could speed the development of new drugs for malaria, Avian flu and HIV/AIDS.
"The simulations that can be done on these resources is a kind of science that wasn't possible before. Now we see the type of collaboration that can be generated on the fly," said Ian Bird, project leader for the LHC computing grid.
About one-third of the grid infrastructure in Europe is not directly owned or funded by the LHC, so that capacity should be available for non-physics experiments, he said.
Wolfgang von Rüden, head of CERN's IT department, said scientific datacentres were once "essentially silos controlled by local administrators. The grid has encouraged opening up the digital infrastructure for science on an unprecedented scale."
Underlying the comments was an acknowledgement that some people simply aren't very interested in particle physics.
"Most of the world at large doesn't really care if we live in a super-symmetric universe or not, or whether we live in one that has 11 dimensions of space-time instead of four," said Glen Crawford, head of the US Department of Energy's high energy physics division. For them the LHC may be remembered for its "sheer audacity," he said.
Von Rüden was asked what impact the grid will have on people's daily lives. He compared its significance to that of the World Wide web, which began life at CERN almost 20 years ago.
"Will I be able to use it at home? Well, indirectly. If you go to your doctor with a CT [scan] and the CT can be analysed quicker by grid technology, then the patient will profit. Similar to the World Wide web, I don't think even Tim Berners-Lee could imagine what it could become one day."
The grid uses a tiered structure that fans out from CERN to progressively smaller computing centres around the world. Data from the collider, expected to be produced at about 15 petabytes (or 15 million gigabytes) per year, will be recorded locally on tape storage at "Tier 0," which is the CERN Computer Centre. From there it will be filtered and distributed to 11 Tier 1 computing centres, including seven in Europe, three in North America and one in Asia, in Taipei. These sites distribute the data to about 160 Tier 2 sites that have enough computing and storage capacity for specific analysis tasks.
The data is carried from CERN to the Tier 1 sites over new fiber-optic cables using a modified version of FTP (file transfer protocol), and from there down it uses standard Internet protocols.
In a recent test, Brookhaven National Laboratory, a Tier 1 site close to New York City, received data from CERN at 800M bytes per second, but expects that to reach 1.5G bytes per second in 2010 and 2.5G bytes per second in 2012, said Michael Ernst, Brookhaven's director of computing facilities.
Brookhaven has already amassed 8 petabytes of scientific data from its own experiments, he said. Data from the LHC will be stored on tape, with 4 petabytes cached on hard disks so it's ready for faster access. The centre has about 2,500 x86 servers, each with up to eight processor cores and running Scientific Linux.
Current security systems at the various facilities are sufficient for the grid, Bird said. "We care about data integrity, but data being stolen is not a problem for us - people are welcome to take it." Knowing what to do with it is another matter, he said.
Bird said the grid can keep expanding.
"How big it can be is anybody's guess, there's no limit here. It's a distributed system, it doesn't depend on a central point so it can grow and grow."
Find your next job with computerworld UK jobs