Bioinformatics Incubator Buys Into Juno's Virtual Supercomputer
Page 1 of 1
The Juno Virtual Supercomputer Network (JVSN), a distributed computing scheme thought up by engineers at Juno Online Inc. as a way to further monetize the company's free subscriber base, signed its first customer Tuesday.
LaunchCyte, which expects to help start more than two dozen bioinformatics companies over the next seven years, anticipates that many of its companies will use the processing power of Juno's subscriber base as a tool to speed medical research in areas like genomics and the creation of new drugs.
The JVS is an alternative to traditional supercomputers. It breaks a computational problem into small pieces which are then sent to subscribers when online. Subscribers' computers analyze the data when the subscribers' screen savers are active, and send the results back to Juno the next time they connect to the Internet.
The Seti@home project, managed by a group of researchers at the Space Sciences Laboratory of the University of California, Berkeley, has used just such a distributed computing scheme for years to analyze radio signals for evidence of extraterrestrial intelligence.
The world's fastest traditional supercomputer, IBM's ASCI White, is rated at 12 TeraFLOPS and costs about $110 million. Seti@home said it gets about 15 TeraFLOPS and has cost about $500,000 to date. Seti@home has about 3 million volunteers who let the project utilize their computers. But with 15.9 million registered subscribers and 4.1 million active subscribers as of March, has the potential to be much more powerful. According to Juno President and Chief Executive Officer Charles Ardai, if all Juno's subscribers were working on the same project at the same time, the JVS would hypothetically break the petahertz barrier with a projected effective processor speed in the order of a billion megahertz.
Juno launched the virtual supercomputer project with the bioinformatics field -- the convergence of life sciences and information science -- in mind. It even brought in Yuri Rozenman, formerly of Applied Biosystems and with 13 years of experience in the bioinformatics field, to head up the project as vice president in charge of the Virtual Supercomputer Network.
Juno's alliance with LaunchCyte will begin with an initial pilot project -- slated to begin later this quarter -- to demonstrate the JVS's potential to LaunchCyte's companies and lay the groundwork for revenue-generating projects the companies might conduct with Juno.
Juno said at least two LaunchCyte-funded development stage companies are expected to use the results of the pilot to evaluate purchasing time on the JVS. One of the companies is a genomics endeavor developing tools to help identify complex clusters of relationships among genes which could potentially lead to new treatments for diseases like heart disease and cancer. The other is a proteomics initiative which is looking at a proprietary protein screening technology with the goal of accelerating drug development.
"Although the bioinformatics market is still in its infancy, it is already highly competitive," said Thomas Petzinger, chief executive officer of LaunchCyte. "Anything that speeds up the development process has the potential to provide a significant competitive advantage. This is why we are excited about the prospect of working with Juno. Juno's virtual supercomputing resources could enable companies we are launching to perform analyses in weeks that might take months on traditional supercomputers."