Data management in grids
Jean-Marc Pierson IRIT, Universite´ Paul Sabatier, 118 route de Narbonne, 31062 Toulouse Cedex 9, France
As it was sketched in previous chapters, Grids have been so far very successful to handle at large scale a lot of computing intensive applications in domains like high energy physics, weather forecast, genetics, . . . . Most of these applications deal with data, either as inputs, outputs or both of their processing. Some of them are dealing with gigabytes, terabytes or even petabytes. In 2006, the digital data production of the humanity reached 161 exabytes. Grids do not deal yet with these numbers, but are each day getting closer. For instance, the Large Hadron Collider at CERN 1 produces 15 petabytes of data each year starting in 2008.