Keeping Data Bits From Burning Away

Research grants target problems associated with superparamagnetism.

Miniaturized colonies of data and laser-assisted magnetization guns may sound like "Honey, I Shrunk the Server Farm," but thanks to government-funded research and some healthy competition, these tools could be in your next disk drive.

Two research grants were awarded last month for novel solutions to the age-old data storage dilemma of superparamagnetism, which is the problem of bits getting so small and close together on a disk that ambient heat makes them magnetically unstable. Over time, this will cause data to evaporate.

The National Institute of Standards and Technologys Advanced Technology Projects group, in Gaithersburg, Md., awarded the grants.

The first approach tackles superparamagnetism by using new disk patterning technology and materials with "mountains and valleys," similar to the arrangement of a CD. In that approach, created by General Electric Co., IBM, Imation Corp. and the University of California at San Diego, the mountains and valleys surround groups of data, making them into small islands and protecting them from the climate of a disk drive.

"Our approach is to put a topography on the media. The media itself has physical patterns on it, of under 50 nanometers in size," said Brad Reitz, principal investigator at GE Research and Development, in Niskayuna, N.Y. "The approach is somewhat of a marriage of optical-disk manufacturing to magnetic storage."

About 20 researchers from the four organizations support the two-year program, with the goal of making a mass-producible, terabyte-capacity disk drive, Reitz said. GE brings the materials expertise; Imation, of Oakdale, Minn., will handle the media itself, and IBM, of Armonk, N.Y., contributes the drive components and production skills, he said. NIST contributed about $2.9 million, and the three companies each gave about $1 million, he said.

Other experts said they believe that even with better media materials, such as the ones being developed by GE and Imation, actually writing to such drives could be an equally daunting task.

The second research grant targets this problem. The solution, called HAMR (Heat-Assisted Magnetic Recording), is led by the National Storage Industry Consortium and Seagate Technology LLC.

HAMR project scientists said they have media materials that are more heat-resistant than normal materials, but the problem is theyre so stable, normal disk drive magnets cant exert a strong-enough force to make the bits writable. So a very short, hot laser beam is used to temporarily make the material susceptible to drive magnets while the data writing occurs, said Mark Kryder, senior vice president of research at Seagate, in Pittsburgh. Then the material cools, and the data is safe again, Kryder said.

For HAMR, there are still problems left to deal with before such a drive can be mass-produced within five years, Kryder said.

"Theres two big challenges. One is to make ... an integrated head that allows you to place the magnetic field and the heated spot in the same location at the same time," Kryder said. "[The other] is that the heated spot needs to be as small as the written bit." A third problem is how to deal with the lubricant built into read/write heads that makes the heads move more freely.

Kryder declined to say how many scientists are working on HAMR, but research is also being done by Advanced Research Corp., of Minneapolis; MEMS Optical Inc., of Huntsville, Ala.; Carnegie Mellon University, of Pittsburgh; and the University of Arizona, in Tucson. NIST gave about $10.7 million; the vendors contributed about $11 million.

Some users think their data may be too valuable to trust to new technology. "When youve got 10 years of data collected, you are going to be conservative about migrating it to something thats just out of the research lab," said systems engineer Ben Allen, of the University of Minnesota Laboratory for Computational Science and Engineering, also in Minneapolis.

Industry analysts generally agreed, adding that the research will probably take longer than the researchers goal of two to five years to succeed.

"We have to find a way to make those bits stay in the polar position theyve been put in," said Dave Reinsel, an analyst with International Data Corp. Like tape technology, disk drive technology probably will be mainstream for longer than researchers like to admit, said Reinsel, in Hutchinson, Minn.

"As long as they can continue to do what they do and push technology and stay ahead of the game, itll keep the others from investing heavily because what they do is simply niche," Reinsel said. IBMs Pixie Dust announcement this summer was a more compelling solution, he said, because its layered chemical approach to the problem requires less new end-user investment.

Rob Enderle, an analyst at Giga Information Group Inc., agreed. The new technologies "appear to be much too aggressive. ... If we could do it in decades, it would be a miracle," said Enderle, in Santa Clara, Calif. "This is interesting, but the answer is just too different."