Many projects and companies have sprung up in the past few years to take advantage of distributed computing and the expanding high-speed Internet.
The largest and most successful distributed computing project to date is SETI@Home, an effort by the Search for Extraterrestrial Intelligence (SETI) at Berkeley. The project uses the computers of its more than three million volunteers analyze the radio signals that are collected by the Arecibo Radio Telescope in Puerto Rico, the largest radio telescope in the world. Users first download and install the program they will run and then continually download sets of data, process them and send the results back to SETIís servers. The program methodically sorts the radio data, searching for patterns and characteristics which indicate that the signal may not have originated from Earth. SETI@Home then follows this up with additional testing on those signals which the user computers set aside. The project has been so successful in gaining computational time that the project is now making plans to expand, adding new recording systems at Arecibo and/or in the southern hemisphere to collect more data and cover more of the sky.
Many organizations are finding that distributed computing can be used in scientific research, especially in the life sciences. The recently completed Human Genome Project uses distributed computing for comparing data with its recently compiled database. One of the most successful of these projects is the FightAids@Home project, which uses independent computers to model the development of drug resistance and design AIDS medications.
Distributed computing is so useful to so many groups that it has become a lucrative business. United Devices and Entropia are two companies which supports multiple independent distributed computing projects. They offer software to businesses which allows them to use the otherwise wasted CPU cycles of the computers in their local area network. They also offer the service of vast grids of volunteered personal computers across the Internet so that projects donít have to have their own network of computers and donít have to recruit the users to utilize the technology. United Devicesí current projects include cancer and genetic research, and Entropia is also supporting life sciences projects as well as corporate and financial use of their platform.
Stanford is running two distributed computing projects at the moment: Folding@Home and Genome@Home. Folding@Home deals with simulating the dynamics of how a protein self-assembles, and Genome@Home tries to design new protein sequences. Between the two of them, the network includes about 35,000 PCs. Dr. Vijay Pande, an assistant professor of chemistry, and the program director sees the research as possibly being useful in future for work in nanomachines, drug design, and fighting diseases relating to proteins, such as Parkinsonís Disease or Alzheimerís Disease. Recently, both projects had rather major breakthroughs, which are being written up in papers to be published soon. However, their main initial goal was as a proof of concept - to show that itís a viable model - and that has already passed with flying colors.
Dr. Pande feels very strongly that distributed computing will be very big in the next five years. What once was considered simply impossible, and not attempted, can now be analyzed and handled in a relatively short period of time. The impossible has suddenly become the possible.
They have faced some minor challenges, but keep plowing ahead. The folding project needs the results from one set of analyses before it can start on the next set, so itís very difficult to effeciently distribute the information over all the thousands of people. Writing their own software was also very difficult, considering they are mostly chemistry majors and graduate students, but now they are about to release version 2.0 of their software. They are also trying to increase the number of users up by a factor of 10 or 100, particularly through intense corporate relations and sponsorship (Iíve heard Googleís name thrown around.) They currently process a 1/2 terabyte of data per year, but thereís still a lot of room for improvement. They are also planning new projects, such as one to find a more accurate way to do drug design.
|[Home] [What Is It?] [History] [Future] [Concerns] [Efficiency] [Curr. Projects] [Resources]|