Over the past two years [email protected] has taken off, harnessing 15 teraflops per second of processing power from 2.3 million personal computers to analyze radio signals from outer space in the search for intelligent life.
Now, several companies are looking to bring the large-scale computer power of unused PC processor time down to earth and into the world of bioinformatics.
Since much of genomic analysis is well suited for parallel processing, it seems a natural application for distributed computing.
But are large pharmaceuticals likely to trust their drug discovery data to millions of desktops worldwide?
Ed Hubbard, CEO of United Devices, one of the emerging firms looking to make [email protected] services a reality, believes they will. “The potential speed-up and the advantage to their business model can be so great it’s worth the risk,” he said.
[email protected] founder David Anderson is chief technology officer at United Devices, which recently raised $13 million in a first round of funding and is on the verge of penning a deal with one of three big genomics companies — Celera Genomics, Incyte Genomics, or Human Genome Sciences.
The distributed computing start-ups claim they can increase the speed of data processing by a factor of 10.
“What is a month or a year worth in coming to discovery faster? That value could be almost infinite,” said James Madsen, CEO and president of San Diego-based Entropia, another company looking for big pharma business.
And besides, he said, the security issues are minimal. “If you think about how distributed computing works, you’re taking a problem and sending it to perhaps a million different PCs. It’s like breaking the Sistine chapel into a million pieces and picking up one piece of stained glass and saying now I can recreate the Sistine chapel,” said Madsen. In other words, even if someone were to break into any one computer, a single piece of information has very little value.
Parabon, yet another company that has recently entered the market, signed a deal with the Genomics and Bioinformatics Group at the US National Cancer Institute’s Laboratory of Molecular Pharmacology.
And Entropia’s Madsen said, “I can tell you that the majority of the big pharmas have sought out Entropia to help them out with their research.” The company has secured $7 million in funding from Mission Ventures.
“We ourselves are exploring those applications right now,” said AstraZeneca global head of R&D informatics Ken Fasman. “We’ve also considered rolling our own in the past, although it looks like there are enough commercial offerings that we may well be able to do the job with somebody else’s software, which would be ideal.”
But sending AstraZeneca’s data to computer users across the globe is out of the question, he said. “I’m absolutely uncomfortable with that. It would be used internally. In a big company like ours there are a lot of desktops,” he said.
The distributed computing companies all offer the option of limiting the service to within the corporate firewall, a much more palatable alternative to pharmaceutical companies.
What will it take to convince big pharma to exploit the enormous potential compute power waiting to be aggregated from the fabric of the Internet?
Nothing less than a paradigm shift in how these companies look to meet their computing needs, according to Hubbard. “It’s the kind of fundamental change that really drives industries to new plateaus,” he said. “This technology represents the kind of classic disruptive technology that comes around ever few years.”
What your PC time is worth:
•Entropia offers its service free to special projects like AIDS research and plays on the altruistic instincts of the general public. Once hooked, it also uses the computer for commercial projects.
•United Devices hands users frequent-flier miles and e-commerce gift certificates.
•Parabon has said it will dole out cash for idle computer time, but has not said how much.
—Aaron J. Sender