Sharing The Load

          Chances are, your computer isn't working as hard as it could be. Much of the time most computers sit around, twiddling their virtual thumbs, waiting for the next excruciatingly slow key stroke or mouse movement. Millions of operations a second of wasted processor time multiplied by millions of computers around the world. This is changing though with the notion of distributed computing which sees wasted processor time as a vast untapped resource.

          Still in its early stages, distributed computing uses the processor time that you don't use on your computer to work on some larger project. The larger project is broken down into smaller pieces, sent out to computer users all over the world who have signed up to take part in the project, then processed, and the results sent back and a new piece downloaded. The project gets the equivalent of a giant supercomputer working for them for free and the volunteers who make up the supercomputer get the satisfaction of working on something they believe in, or increasingly, the possibility of money.

          One of the first organizations to tackle the notion of distributed computing was SETI, the Search For Extra Terrestrial Intelligence. This group searches the skies for signs of intelligent life in the form of radio signals. Funding for this kind of research tends to be scarce in the commerce-driven world we live in, but undaunted, scientists at SETI have become masters in conserving resources and making do. Getting time on overbooked radio telescopes is expensive so they designed Project SERENDIP to run on radio telescopes engaged in other tasks to collect raw data for SETI. Processor time for the kind of number crunching needed to examine this data is very expensive indeed so a group at the University of California at Berkeley developed SETI@home to run as a screensaver and in the background on home computers. It was (and is) successful beyond their wildest expectations. It turns out that there are a lot of people willing to help out who never had the ability to do it before.

          SETI@home downloads a small piece of data from the host server, processes it then signals when it is done and downloads another piece. Each piece of data is a small download, around 360 K, and requires some 175 billion equations to process, taking typically around 20 - 40 hours of CPU time that was not being used for anything else to run. Consult the SETI@home website for system requirements and more information.

          There are currently about 1.5 million people and their computers helping out with this effort world wide including some 93,000 Canadians. Chebucto users are the second highest contributors to this effort in Nova Scotia, closely following the first place N.S. provincial government computers. Current stats on this are here. Even people in countries one would not typically associate with computers and the internet, such as Iraq, Mongolia and East Timor are contributing time to this effort.

          There is no money involved with the SETI@home project though should signals be discovered, the user would be given co-discoverer credit and a place in the history books.

          Money is involved with another distributed computing project, the cracking of code keys.

          In what has become almost a cottage industry, companies involved with making cryptographic products give rewards to people or groups who crack or break the key on a particular encrypted message. The companies do this mostly to demonstrate the inadequacies of current world standard cryptography using 56 bit keys. Current strong cryptography involves the use of 128 bit keys, but there are export restrictions on such software so companies in this field use these contests in an effort to lobby for change.

          Distributed.Net is currently engaged in several of these projects, the latest being to crack a 56 bit encoded message put out by CS Communications & Systems for a €10,000 prize (about $15,000 CDN) which is supposed to run until March 17, 2000. The software is similar to SETI@home, running in the background on your computer and getting out of the way when other programs on your computer want resources. The process uses "brute force" methods, literally trying every possible key. Distributed.Net will put a share of the prize money towards non-profit groups voted upon by the users and give €1000 (about $1500 CDN) to the individual computer user who successfully cracked the key and another €1000 to their team (or if there was no team, this goes to the user as well). They have signed up some 25,000 users, of whom 610 are Canadian.

          They are not alone in using distributed computing to crack this particular contest. DCypher.Net has some 2559 people signed up working on this same problem, however they plan to give the entire €10,000 to the successful user. This group plans on offering other distributed computing projects as time goes by and look to eventually have a program that will sit on the user computer able to plug in different project modules at will.

          Perhaps the most inventive of the distributed computing projects on the web is being offered by NanoComputer.Org, which seeks to design and build a molecule-sized supercomputer and is looking for some processor time. Looking to complete the project by 2011, the group of some 200 scientists, including Dr. Richard Smalley of Rice University, who recently won the Nobel Prize for co-discovering Buckminsterfullerene, a new arrangement of carbon atoms with new properties, is working with the most cutting edge technology in existence such as robots made up of fractal cubes which can change shape at will. Scientists and engineers are welcome to sign up as is anyone with some spare processor cycles.

          There are limits to what can be worked on with distributed computing. A project has to be capable of being broken down into smaller sized pieces that can be processed individually then re-assembled. However with the advent of more and more powerful home computers using less and less of their resources to run the individual user's programs, it seems an idea whose time has come.

You may direct comments or suggestions about this article to:

Andrew D. Wright,  au141@chebucto.ns.ca

Back To The Beacon Index