Skip to Content

The Big, Bad Bit Stuffers of IBM

The ferocious progress in disk storage densities has come thanks to an IBM lab that was slated for elimination–until it met the “gigabit challenge.”

Bob Fontana, research member at IBM’s Almaden Research Center in San Jose, Calif., is only half joking when he says Silicon Valley should have been called Iron Oxide Valley. Or even Rust Valley. Because for Fontana, it’s iron oxide-the original material used to coat the disk drives that store magnetic bits of information-that fueled the growth of Silicon Valley.

Of course, he may be a little biased. IBM invented the disk drive in San Jose in 1956, when this part of the world was better known for cherry orchards than industrial parks. Since then, Almaden researchers have repeatedly smashed the record for how much data can be stored on a disk. They were up to their old tricks again last December, when Fontana and his colleagues squeezed more than 11 billion bits (gigabits) onto a single square inch of magnetic material. That more than doubled the previous record of 5 billion bits per square inch, set in the same lab only a year earlier. How much is 11 billion bits? It’s roughly equivalent to 725,000 pages of double-spaced text, which would stack up higher than an 18-story building. By any measure, this was a great scientific achievement.

This leapfrogging has had a dramatic effect on what personal computers can do. It is these huge capacity hard drives that have made it practical for computer users to keep large amounts of extremely sophisticated software on their machines, for example. Vast hard disks have also fostered the transformation of computing from a textual activity to one filled with pictures and sounds. What’s more, the way the disk-drive project is managed highlights an effort by IBM to recouple basic research to product development in the service of innovation.

Almaden’s accomplishments are by now so well accepted in the world of computing that the announcement in December of yet another new record didn’t make big headlines. Even competitors shrugged. “Everyone in the audience was saying, sure, that’s what we’ve been waiting to hear,” says Gordon Knight, chief technical officer for TeraStor, a Silicon Valley startup that is championing a different kind of storage technology than IBM’s. But beneath this calm surface of expectations fulfilled lies a surprising story.

Filling a Hall of Fame

The world wasn’t always so nonchalant about IBM’s breakthroughs. After all, magnetic storage was supposed to be dead by now, replaced by optical storage devices or some other technology. Even IBM thought so: In 1970, an IBM research scientist published a paper “proving” that the technology would never go beyond 200 megabits per square inch.

But instead of believing the company’s own experts, the team at Almaden topped one predicted limit for magnetic storage after another. They discovered engineering work-arounds for what were once thought to be hard physical limits. By 1989, the Almaden lab was packing 1 gigabit per square inch. In the following years, Almaden has upped the ante, demonstrating densities of 3, 5 and now 11.6 gigabits per square inch. The market has begun to take for granted that magnetic storage capacities will double every 18 months, following roughly the same feverish pace set by the semiconductor industry. The big news will be when IBM slows down.

Talk to Currie Munce, director of storage systems and technology at Almaden, and he’ll complain that magnetic storage scientists are the unsung heroes of the information age. Like everyone else at Almaden, Munce likes to evangelize about storage: “We’re trying to move things mechanically over millimeter distance in milliseconds and to get them to settle within tenths of microns on track,” he says. “It’s great science.”

A visit to one particular room at Almaden shows how far IBM has run with the technology. Hanging on one wall is a single rusty platter from the original 1956 drive-proudly displayed as a rock star might show off a platinum record. Back in 1956, IBM’s disk drives were refrigerator-sized boxes that held a mere 5 megabytes, on 24 platters, each 2 feet across. Today the company ships a standard PC drive that holds more than 16 gigabytes, some 3,000 times the capacity of its original product. Put the 1956 drive and the 1998 drive side by side, though, and they look the same except for scale. Dave Thompson, director of Almaden’s Advanced Magnetic Storage Recording Laboratory, says the inventors of the original disk drive could walk into his lab today and know exactly what was going on.

Storage density depends on the size of the magnetic bit: that portion of disk real estate that is given a particular magnetic orientation-north or south-to represent a binary one or zero. At the most elementary level, the goal is simple: Shrink the bits and you expand the storage capacity. But smaller bits emanate smaller magnetic fields, which in turn requires positioning the reading head-the device that senses these fields and converts them to electrical signals-closer to the spinning disk surface.

Over and over again, shrinking the technology has forced disk-drive developers to confront physical limits that first appeared to be insurmountable. For example, the head rides on a cushion of air created by the spinning disk. Conventional wisdom held that bringing the head too close to the surface would squeeze the air molecules into a space so small the supporting cushion would disappear. “There was a lot of math to back those conclusions up,” says Barry Schechtman, executive director for the National Storage Industry Consortium (NSIC), an intercompany consortium of storage manufacturers that funds basic research at several universities. The happy reality, however, was that this theory was not true. “Nature turned out to be smarter than our equations, which needed to be modified,” says Schechtman.

Rethinking Research

Almaden’s dramatic progress in magnetic recording is all the more remarkable when you consider its institutional history. In the early 1980s, IBM’s research division had a reputation for performing brilliant work that had little relevance to the company’s business. And even when the labs did produce findings that had commercial implications, the handoff to product groups was often fumbled, allowing other companies to capitalize on IBM’s research breakthroughs before IBM did. By 1981 IBM had fallen so out of touch with the market that Big Blue had to cobble together its first PCs out of components-including disk drives-made by other companies.

The Almaden building itself is a throwback to the great research labs of the past, surrounded by hundreds of millions of dollars worth of empty real estate, where the only sound is the wind sweeping over the Santa Teresa foothills. Conceived in the late 1970s when IBM had money to burn, Almaden was to be a showcase institution for pure research on a par with AT&T’s Bell Labs, Xerox’s Palo Alto Research Center, and IBM’s own research facility in Yorktown Heights, N.Y. By the time the building was completed in 1985, though, the climate for corporate R&D had changed. New CEO John Akers-a man who liked to use words like “cost containment” and “streamlining” when describing IBM’s mission-toyed with the idea of dismantling the research division altogether and dispersing its employees into various product groups.

Out of the Ivory Tower

The 1-gigabit challenge was the catalyst for other deep changes that brought the research scientists at Almaden out of their ivory tower and into the real world of profits and losses. Almaden began to do regular joint development work with IBM’s Storage Systems Division, the product group down the hill. Soon, one-third of Almaden’s research budget was coming from the product division. Munce feels that the ratio of funding gives the research group clarity without compromising independence.

The change in funding has also created a new kind of thinking about innovation, Munce says. “Ten years ago the attitude here was: if I didn’t invent it, I don’t want to work on it, because I won’t get credit for it,” he explains. “Today we’re trying to say: If you invent it in the lab, or if you’re the first one to grab it out of someone else’s lab and make it relevant, we don’t care.”

Munce holds a joint managerial position that reports to both the research and the product divisions. It was a position created in the early 1990s to make the ties between the two groups even closer. He sits in on the meetings of the product divisions and then tries to figure what research is needed to serve their missions. Putting on his research hat, his job is to influence IBM’s product divisions to move in directions that take advantage of the work coming out of the labs. “My job really is to manage innovation,” he says. “We need to be separate so we can innovate, create and motivate people to do good research. But we need to be connected to get technology to market.”

A graphic example of how far the research team has evolved from the original vision for Almaden is the Advanced Magnetic Recording Laboratory itself. The lab, designed by research staff member Fontana, is jointly staffed by research and product groups. Two years after Almaden first opened its doors, Fontana convinced his managers to rip the guts out of one wing on the first floor. This renovation gave him a 5,000-square-foot lab to do prototyping work-the kind of work that used to be done by IBM’s product engineers instead of its research staff. This lab provides facilities for building components quickly, allowing the researchers upstairs-specialists in read heads, write heads, materials science and other areas critical to disk-drive technology-to test whether their innovations would work together.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.