A Brief History of SETI@Home

The year was 1999, and the people were going online. AOL, Compuserve, mp3.com, and AltaVista loaded bit by bit after dial-up chirps, on screens across the world. Watching the internet extend its reach, a small group of scientists thought a more extensive digital leap was in order, one that encompassed the galaxy itself. And so it was that before the new millennium dawned, researchers at the University of California released a citizen-science program called SETI@Home.

The idea went like this: When internet-farers abandoned their computers long enough that a screen saver popped up, that saver wouldn’t be WordArt bouncing around, 3-D neon-metallic pipes installing themselves inch by inch, or a self-satisfied flying Windows logo. No. Their screens would be saved by displays of data analysis, showing which and how much data from elsewhere their CPUs were churning through during down-time. The data would come from observations of distant stars, conducted by astronomers searching for evidence of an extraterrestrial intelligence. Each participating computer would dig through SETI data for suspicious signals, possibly containing a “Hello, World” or two from aliens. Anyone with 28 kbps could be the person to discover another civilization.

When the researchers launched SETI@Home, in May of ’99, they thought maybe 1,000 people might sign up. That number—and the bleaker view from outsiders, who said perhaps no one would join the crew—informed a poor decision: to set up a single desktop to farm out the data and take back the analysis.

But the problem was, people really liked the idea of letting their computers find aliens while they did nothing except not touch the mouse. And for SETI@Home’s launch, a million people signed up.  Of course, the lone data-serving desktop staggered. SETI@Home fell down as soon as it started walking. Luckily, now-defunct Sun Microsystems donated computers to help the program get back on its feet. In the years since, more than 4 million people have tried SETI@Home. Together, they make up a collective computing power that exceeds 2008’s premier supercomputer.

But they have yet to find any aliens.

SETI is a middle-aged science, with 57 years under its sagging belt. It began in 1960, when an astronomer named Frank Drake used an 85-foot radio telescope in Green Bank, West Virginia, to scan two Sun-like stars for signs of intelligent life—radio emissions the systems couldn’t produce on their own, like the thin-frequency broadcasts of our radio stations, or blips that repeated in a purposeful-looking way. Since then, scientists and engineers have used radio and optical telescopes to search much more of the sky—for those “narrowband” broadcasts, for fast pings, for long drones, for patterns distinguishing themselves from the chaotic background static and natural signals from stars and supernovae.

But the hardest part about SETI is that scientists don’t know where ET may live, or how ET’s civilization might choose to communicate. And so they have to look for a rainbow of possible missives from other solar systems, all of which move and spin at their own special-snowflake speeds through the universe. There’s only one way to do that, says Dan Werthimer, the chief SETI scientist at Berkeley and a co-founder of SETI@Home: “We need a lot of computing power.”

In the 1970s, when Werthimer’s Berkeley colleagues launched a SETI project called SERENDIP, they sucked power from all the computers in their building, then the neighboring building. In a way, it was a SETI@Home prototype. In the decades that followed, they turned to supercomputers. And then, they came for your CPUs.

The idea for SETI@Home originated at a cocktail party in Seattle, when computer scientist David Gedye asked a friend what it might take to excite the public about science. Could computers somehow do something similar to what the Apollo program had done? Gedye dreamed up the idea of “volunteer computing,” in which people gave up their hard drives for the greater good when those drives were idle, much like people give up their idle cars, for periods of time, to Turo (if Turo didn’t make money and also served the greater good). What might people volunteer to help with? His mind wandered to The X-Files, UFOs, hit headlines fronting the National Enquirer. People were so interested in all that. “It’s a slightly misguided interest, but still,” says David Anderson, Gedye’s graduate-school advisor at Berkeley. Interest is interest is interest, misguided or guided perfectly.

But Gedye wasn’t a SETI guy—he was a computer guy—so he didn’t know if or how a citizen-computing project would work. He got in touch with astronomer Woody Sullivan, who worked at the University of Washington in Seattle. Sullivan turned him over to Werthimer. And Gedye looped in Anderson. They had a quorum, of sorts.

Anderson, who worked in industry at the time, dedicated evenings to writing software that could take data from the Arecibo radio telescope, mother-bird it into digestible bits, send it to your desktop, command it to hunt for aliens, and then send the results back to the Berkeley home base. No small task.

They raised some money—notably, $50,000 from the Planetary Society and $10,000 from a Paul Allen-backed company. But most of the work-hours, like the computer-hours they were soliciting, were volunteer labor. Out of necessity, they did hire a few people with operating-system expertise, to deal with the wonky screensaver behavior of both Windows and Macintosh. “It’s difficult trying to develop a program that’s intended to run on every computer in the world,” says Anderson.

And yet, by May 17, 1999, they were up, and soon after, they were running. And those million people in this world were looking for not-people on other worlds.

One morning, early in the new millennium, the team came into the office and surveyed the record of what those million had done so far. In the previous 24 hours, the volunteers had done what would have taken a single desktop one thousand years to do. “Suppose you’re a scientist, and you have some idea, and it’s going to take 1,000 years,” says Anderson. “You’re going to discard it. But we did it.”

After being noses-down to their keyboards since the start, it was their first feeling of triumph. “It was really a battle for survival,” says Anderson. “We didn’t really have time to look up and realize what an amazing thing we were doing.”

Then, when they looked up again, at the SETI@Home forums, they saw something else: “It was probably less than a year after we started that we started getting notices about the weddings of people who met through SETI@Home,” says Eric Korpela, a SETI@Home project scientist and astronomer at Berkeley.

The SETI astronomers began to collect more and different types of data, from the likes of the Arecibo radio telescope. Operating systems evolved. There were new signal types to search for, like pulses so rapid they would have seemed like notes held at pianissimo to previous processors. With all that change, they needed to update the software frequently. But they couldn’t put out a new version every few months and expect people to download it.

Anderson wanted to create a self-updating infrastructure that would solve that problem—and be flexible enough that other, non-SETI projects could bring their work onboard and benefit from distributed computing. And so BOINC—Berkeley Open Infrastructure for Network Computing—was born.

Today, you can use BOINC to serve up your computer’s free time to develop malaria drugs, cancer drugs, HIV drugs. You can fold proteins or help predict the climate. You can search for gravitational waves or run simulations of the heart’s electrical activity, or any of 30 projects. And you can now run BOINC on GPUs—graphical processing units, brought to you by gamers—and on Android smartphones Nearly half a million people use the infrastructure now, making the système totale a 19 petaflop supercomputer, the third-largest megacalculator on the planet.

Home computers have gotten about 100 times faster since 1999, thank God, and on the data distribution side, Berkeley has gotten about 10 times faster. They’re adding BOINC as a bandwidth-increasing option to the Texas Advanced Computing Center and nanoHUB, and also letting people sign up for volunteer computing, tell the system what they think are the most important scientific goals, and then have their computers be automatically matched to projects as those projects need time. It’s like OkCupid dating, for scientific research. BOINC, and SETI@Home can do more work than ever.

The thing is, though, they’ve already done a lot of work—so much work they can’t keep up with themselves. Sitting in a database are 7 billion possible alien signals that citizen scientists and their idle computers have already uncovered.

Most of these are probably human-made interference: short-circuiting electric fences, airport radar, XM satellite radio, or a microwave opened a second too soon. Others are likely random noise that added up to a masquerade of significance. As Anderson says, “Random noise has the property that whatever you’re looking for, it eventually occurs. If you generate random letters. You eventually get the complete works of Shakespeare.” Or the emissions are just miscategorized natural signals.

Anderson has been working on a program called Nebula that will trawl that billions-and-billions-strong database, reject the interference, and upvote the best candidates that might—just might—be actual alien signals. Four thousand computers at the Max Planck Institute for Gravitational Physics in Germany help him narrow down the digital location of that holiest of grails. Once something alien in appearance pops up—say from around the star Vega—the software automatically searches the rest of the data. It finds all the other times, in the 18 years of SETI@Home history, that Arecibo or the recently added telescopes from a $100 milion initiative called Breakthrough Listen have looked at Vega. Was the signal there then too? “We’re kind of hoping that the aliens are sending a constant beacon,” says Korpela, “and that every time a telescope passes over a point in the sky, we see it.”

If no old data exists—or if the old data is particularly promising—the researchers request new telescope time and ask SETI colleagues to verify the signal with their own telescopes, to see if they can intercept that beacon, that siren, that unequivocal statement of what SETI scientists and SETI@Home participants hope is true: That we are not alone.

So far, that’s a no-go. “We’ve never had a candidate so exciting that we call the director and say, ‘Throw everybody off the telescope,’” says Werthimer. “We’ve never had anything that resembles ET.”

And partly for that reason, the SETI@Homers are now working on detecting “wideband” signals—ones that come at a spread spectrum of frequencies, like the beam-downs from DIRECTV. Humans (and by extension, extraterrestrials) can embed more information more efficiently in these spread-spectrum emissions. If the goal is to disseminate information, rather than just graffiti “We’re here!” on the fabric of spacetime, wideband is the way to go. And SETI scientists’ thinking goes like this: We’ve been looking mostly for purposeful, obvious transmissions, ones wrapped neatly for us. But we haven’t found any—which might mean they just aren’t there. Extraterrestrial communications might be aimed at members of their own civilizations, in which case they’re more likely to go the DIRECTV route, and we’re likely to find only the “leakage” of those communication lines.

“If there really are these advanced civilizations, it’d be trivial to contact us,” says Werthimer. “They’d be landing on the White House—well, maybe not this White House. But they’d be shining a laser in Frank Drake’s eyes. I don’t see why they would make it so difficult that we would have to do all this hard stuff.”

And so humans, and our sleeping computers, may have to eavesdrop on messages not addressed to us—the ones the aliens send to their own (for lack of a better word) people, and then insert ourselves into the chatter. “I don’t mean to interrupt,” we might someday say, “but I couldn’t help overhearing...” And because of SETI@Home and  BOINC, it might be your laptop that gets that awkward conversation started.