Mainstream industry are not early adopters, as is obvious from the only recent boom of commodity Linux, OpenSource and database solutions. Observing Miron Livny’s work in grid computing at the University of Wisconsin since 2000, I wondered why companies didn’t absorb and utilize this concept, when SETI@Home, Cancer Grid and academic researchers have put it to great use. A system like Livny’s Condor could crunch through the oil industry’s petabyte seismic data sets just as it iterated our academic research group’s forward and inverse electromagnetic models. Some companies have internal grids and clusters, but what if all the oil majors could get together and host a super-grid and make it available to internal clients and then schools and research labs?
Globus which started as a government-funded project is now in the news because it has become a Consortium with backing from industry bigwigs such as IBM, HP, Sun and Intel. A great idea, but I foresee some problems: security and resource ownership. How secure is proprietary information and how much is company X going to pay the Consortium for time and resources used? Grids work well because so far they have been internal or non-profit. I can see a huge amount of bureaucracy created and time and money wasted by the various business/finance departments over this. Someone out there had already thought of this and wrote about it back in 2001.
Industry paranoiacs like oil companies may fall into this late in the game for fear of information loss/theft, but you never know. I maintain that it would be in their best interest to develop their own grid for data crunching and customize the necessary software and networks to their own business needs.
Grid Computing resources:
Cheat sheet: Grid Computing
grid.org
