Grid computing goes very large

09 Sep 2008

As all you physics fans know, today (September 10) is the day that the European Organization for Nuclear Research (CERN) switches on its Large Hadron Collider (LHC), the world's biggest particle accelerator that will either unlock the secret of the Big Bang or cause another one that destroys the universe, depending on who you believe.

Assuming we all survive (which we probably will), one of the side benefits of the LHC could be a major breakthrough in grid computing, according to Scientific-American.

While the LHC smashes atoms, by the time it reaches full capacity next year, it will be generating raw data at a rate of 5 GB per second, with works out to an annual data output of 15 petabytes worth of snapshots of particle collisions that scientists will access and analyze.

To handle the load, universities, laboratories and institutes participating in LHC have developed the LHC Computing Grid, which differs from standard grid computing via the Web (which faces issues such as bottlenecks and bandwidth costs) by taking a tiered approach.

Tier 0, located at CERN, stores and manages the raw data, and transmits portions over dedicated 10-Gbps fiber links to 11 Tier 1 sites across North America, Asia and Europe, which in turn parcel out data to 140 Tier 2 computer networks where scientists actually access and process the data.

The linchpin of all this is Globus, an open-source middleware platform that can gather requested information seamlessly from across the entire grid.

That's the theory, anyway. But if the LHC Computing Grid performs as promised, says Sci-Am, it would mean a major milestone in grid computing:

If project scientists can tame massive, worldwide fields of networked data and computing cycles in particle physics, their solutions could well apply across the Internet"”in much the same way that [Tim] Berners-Lee's specialized HTML invention morphed into the very backbone of modern technological society.

Translation: the Internet could finally leverage the power of grid computing in a way that can cost-effectively process heavy traffic in a manner completely transparent to the end-user. If nothing else, it could help you sift through all your digital home videos that you never look at anymore.

Related content

No Comments Yet! Be the first to share what you think!