More than a century after its invention, the venerable punch card is fueling a quiet contest between the worlds largest computer makers. IBM and Hewlett Packard are racing to develop a nanoscale version based on atomic force microscopy, with the potential for vast data storage at low cost. Both companies say the technology, called probe storage, has moved beyond laboratory curiosity. Today, IBM researchers announced that they have packed bits together at a density of one trillion per square inch, more than 20 times denser than todays disk drives.
It allows us to store a very large amount of data in a small form factor, says Peter Vettiger, the projects director at IBM Research in Zurich. Although a single atomic probe is too slow to serve as a useful memory device by itself, Vettigers group has built an array of 1,024 probescodename millipedethat read, write and, importantly, erase data on a plastic medium. And in less than a year, Vettiger says, his team is on track to complete a fully-functional prototype, connect it to a lab computer via a standard interface, and use it as a working memory devicea big step toward putting a commercial version on the market as early as 2005.
|A prototype chip, with the wiring to address millipede’s 1,024 probes. (Photo courtesy of IBM).|
Hewlett Packard has taken a different approach. Instead of punching holes like millipede, the HP system writes bits by changing the mediums crystalline structure, much like an optical disk. Chuck Morehouse, head of HP Labs information access laboratory, says his group has achieved results comparable to IBMs, although they are still unpublisheda disparity Vettiger is quick to point out. They dont publish very much, says Vettiger, whose name has appeared on seven papers published since 2000. But Morehouse says his team has built probe arrays that do nearly everything that a production device needs to do, and calls Hewlett Packard very serious about developing revolutionary devices based on probe storage.
One reason for the Big Blue researchers prolificacy, Morehouse suggests, is that they faced an uphill battle demonstrating that thermomechanical storage can be successfully erased and rewritten. But in a paper published today in a new journal, IEEE Transactions on Nanotechnology, Vettigers group reports that millipede has achieved hundreds of thousands of rewrites in a single location. By comparison, a rewriteable CD or DVD can be rewritten only about 10,000 times.
Both companies are tight-lipped about their plans to develop probe storage-based products. There are no plans that we could communicate, says IBM Zurich spokesman Martin Hug, who emphasizes that millipede is still in a research state.
At Hewlett Packard, commercialization manager Mike Altree says the company definitely has plans to productize our probe-based storage. Altree, who works for HPs personal storage business line, is leading the group charged with developing scenarios for entering the market, and is in the process of winnowing the list down to one or two. He declines to discuss a timeline, saying that the initiative is getting well enough defined that it would be misleading to be very vague, and inappropriate to be very specific.
Nor will Altree say what those scenarios are. But talk to the researchers, and a picture begins to emerge. Morehouse and Vettiger sound surprisingly alike in their excitement about the integration of portable devices such as handhelds and mobile phones, and their growing need for memory. Probe storage is an ideal system for storage in mobile systems, Vettiger says. If I could store your complete CD library on this little storage device, you could play your entire collection at any time. He envisions a product that looks and functions like a flash card, but stores 1,000 times the data at the same cost. Like flash, he says, it could be integrated into chip, or packaged as a portable memory card.
IBM and HP arent the only companies to take notice; Samsung, SeaGate and Hitachiwhich last week purchased IBMs hard drive businessare also investigating probe storage. If IBM completes its prototype on schedule, it will have built on what is already a visible lead. But, Morehouse says, until you can go down to Best Buy and buy a fistful of cards, the game isnt over.