y desktop computer has a hobby. When it's not busy being my word processor, or balancing my checkbook, or being my portal to the Internet, it casts its eyes skyward and searches for extraterrestrial life. It's listening for an ET - not an ET that is phoning home, but one that is calling to us.
My computer isn't doing this alone, of course. Searching for ET requires a lot of things I don't have, the first one being a multimillion-dollar radio telescope. Even if I had my own telescope, my computer lacks the power to analyze all of the radio signals coming in from space, filter out the background noise, and scan for the sounds that might be characteristic of intelligence. In fact, there is no computer on the earth today that is powerful enough to accomplish this task.
That's why my computer has joined the ranks of more than a million other machines in the international Search For Extraterrestrial Life At Home (SETI@Home) project, which is being coordinated by the University of California at Berkeley.
SETI@Home is one of several projects under way right now that is harnessing the power of multiple computers for a single purpose. Another is Distributed.NET, which is using spare CPU time on several thousand computers to crack encryption keys. Both these projects distribute a special-purpose screen saver that runs on Windows and Macintosh-based PCs. When you aren't using your desktop computer, the screen saver comes on and starts crunching mathematical problems. When you start using your computer again, the program politely saves its work and gets out of your way. (Versions of the SETI@Home program are also available for more than a dozen versions of UNIX.)
The Internet is the key to coordinating these distributed projects. The first time you start up the SETI@Home program, the software makes a connection over the Internet to a computer in California and downloads a ''work unit'' - that is, a set of measurements from a particular part of the sky. The work unit is about 350 kilobytes of data, so it downloads in less than three minutes over a dial-up telephone line. That's not a lot of data, but it takes about 40 hours of CPU time on a 550 Mhz Pentium III computer to crunch it. When my computer is done with the data processing it makes another call on the Internet to Berkeley, uploads its results, and downloads a new work unit. So far my desktop has completed 10 work units.
I run a second copy of SETI
@Home on my laptop computer. Now, between my trips, the laptop spends its days searching the stars. It's about a third as fast as my desktop, but it has still completed more than 10 work units all on its own.
Of course, it costs money for me to leave my computers on all the time. My laptop's power supply draws 82 watts, which translates to 5 to 10 cents per day in higher electricity costs. But that's money that I'm happy to pay to support this scientific project. Indeed, this nickel-a-day donation is key to the success of the Berkeley project. Even if Intel or Compaq had donated a million desktop computers to the SETI project (an unlikely proposition at best), it is hard to imagine that the Berkeley team could have raised the money to pay for the electricity, cooling, and laboratory space those machines would have needed.
To further mobilize the community, the SETI@Home researchers have put together a comprehensive Web site that allows participants like me to see how their contributions to the effort stack up against the project as a whole. The Web site reports the number of work units you have completed, the total time you have spent crunching data for the project, and the average amount of time spent per work unit. The Web site further appeals to the competitive streak that most humans have by allowing individuals to join together in teams. Schools, clubs, and companies all have fielded SETI teams. Not surprisingly, the top-ranked teams are at Sun Microsystems and Silicon Graphics, where engineers run the software on some of the world's fastest workstations. Not only is this great for science, but it's also a not-so-subtle form of advertising.
Some commentators have said SETI@Home is a model for how many science projects requiring massive amounts of computation will work in the future. Others have taken the idea further, saying we'll be able to make a few extra dollars by renting out our computers to corporations, electronically selling cycles to the highest bidders. I don't think this is likely.
The first barrier is security. Everyone who runs SETI@Home on their computers has put a tremendous amount of trust in the program's authors. We trust that the astronomical screen saver isn't doing anything nefarious, like searching for credit-card numbers or spinning calculations for nuclear weapons. Likewise, the SETI@Home coordinators trust that misguided pranksters won't modify their programs and make them report false sightings of ET. It might only take a few such cases to poison the community spirit that has led more than a million people all over the world to download and run the SETI software.
A second potential threat is competition. As more of these parallel computation projects spring up, there will be more and more competition between them. Clearly, if there are hundreds of projects like this on the Internet, they all won't be able to get a million participants.
But despite these fears of mine, SETI@Home is off to a wonderful start. And it would be auspicious indeed if the only way that we could hear ET's call is by working together as a human race, without national or language barriers separating us.
You can learn more about the SETI@Home project, and download the software, by visiting setiathome.ssl.berkeley.edu.
Simson L. Garfinkel has set up a Web-based chat system for readers of this column. The address is chat.simson.net.
Dow:
