Thinking boxes
“ In 1999 the internet generated 2 exabytes (1.074 billion gigabytes) of information; by the end of 2007 it was handling 1 exabyte an hour.”
It’s not just the internet that faces this growth, the challenge of dealing with the demands of storing and serving the growing rate of information is being dealt with in different ways by different organisations. People are having to think big. A couple of articles highlight just how big…
The computer grid is the main way of dealing with this kind of problem – low cost machines clustered together to give the processing power and storage required. Cern (my first workplace, so holds a special pace in my heart) is building a grid of 100,000 CPUs to deal with the load and data generated by the Large Hadron Collider – up to 15 petabytes (15 million gigbytes) a year.
IBM is contemplating a slightly different approach. A distinctly different approach. It is taking the grid idea, removing a number of machines until there is one left:
“The Register has unearthed a research paper that shows IBM working on a computing system capable “of hosting the entire internet as an application.”
One machine but over 67 million cores. I want one of those… but I don’t think I have the money for the electricity bill. Or the real estate. The article continues:
“[The] platform is an order of magnitude more efficient to purchase and operate than the commodity clusters in use today…a more reliable computer – by more than two orders of magnitude – than commodity boxes which fail all the time. IBM has issued a novel approach to the utility computing, mega data center problem.”
To be honest, I’d never convince my wife anyway…
- Computer Weekly: Cern builds largest computer grid to study origins of the universe
- The Register: IBM explores 67.1m-core computer for running entire internet
- Wikipedia: Large Hadron Collider
