http://www.digitalmagazine.com.au/2008/04/08/the-grid-at-cern-takes-grid-computing-to-new-levels/

What is grid computing? Grid computing, also known as distributive computing, is a way for incredibly large amounts of data to be spread over multiple network computers for analysis. By using either dedicated computers or the unused idle time and resources on volunteer computers, a project that might take hundreds of years can be completed in a few months. CERN’s new dedicated internet called The Grid, will allow the data collected from the Large Haldron Collider(LHC) to be distributed over networked academic and research facility computers in countries around the world.

The LHC, particle accelerator, will be used to conduct a variety of physics experiments. The experiments range from recreating the environment that existed just after the Big Bang to searching for Higgs boson (what gives mass to particles). These experiments will produce about 15 million gigabytes of data per year. No single computer or single site research center can analyze all of the projected data. So CERN, where the LHC resides, has constructed a dedicated fiber optic network that is 10,000 times faster than any current broadband internet speeds. 50,000 servers (projected to reach 200,000 servers in two years) are spread across the globe. This private internet will allow teams of physicists world wide to analyze and use the data.

Distributive computing has been around since the 1990s. The process gained notoriety when distributive.net in 1997, and SETI@home in 1999 began using volunteer computers as part of grid computing. Since then companies, universities and research facilities have used grid computing for a variety of projects. The volunteer projects range from searching for cures for malaria, cancer and AIDS; determining the range and effects of climate change; searching for extraterrestrial life in space; and solving intricate chess moves.

Check out this interesting article which is talking about grid computing, it was a great read.