Sunday, October 30, 2011

Distributed Computing


This is probably the more boring side of cloud computing, but it does mark something amazing in research. Cloud computing is all about an array of computers or super computers with tons of memory and storage space and a metric shit ton of processing power. Distributed Computing is all about making that super computer a reality.
Companies like Amazon, Microsoft and Apple have the means, capitol and technology specifically, to pull off an amazing cloud service for all to enjoy. Cloud computing is part of buzz marketing today but maybe someday it will be the future where no one person actually owns a computer. That's a high concept, impossible future, but it is a fun idea to toy around with.

Unlike the large corporations, researchers don't have those means, specifically capitol, to make super computers. To make a sub-standard super computer (which compares with general standards) costs well over a million dollars. To make a fast super computer, it would take much more than ten times that. You could probably guess where this is going; Distributed Computing is about amassing computers that aren't super computers and distributing processing load across all of those computers.

Folding@home is probably the most popular Distributed Computing software because it is on the PS3 (a great folding machine). To explain it in great detail would take some time, but I'll sum it up best I can. Protein chains make up important 'machines' in our bodies, like hemoglobin, the protein red blood cells use to carry oxygen. Any number of proteins can have the same chain, but drastically different mechanics and uses. The differentiation comes with how the chain is folded against itself. With immense study we can discover the folds ourselves, but a computer can do it much faster (within an hour instead of a week).

Folding@home allows anyone to participate in this. The speed is dependent on the computer, but when a million computers are doing it at the same time, the work gets done very quickly, perhaps even quicker than a super computer would. There are many @home projects going on, Folding is just one of the many. It's a fun and interesting way to help medical science.

There is another form of Distributed Computing that is much like Folding@home, but it doesn't involve computer processing; it involves human puzzle solving. Foldit is a program with a much similar goal to Folding@home, but it is a puzzle solving game where one folds a protein chain as hard as possible. When we understand the chain, it's easy to draw its shape and figure out how it would react. The difficulty is finding the best possible way to fold it. It may take over 15 years for a small group to solve this issue, but two weeks for 100,000 puzzle solvers to do the same thing.

No comments: