For the last several decades, most storage advances have been an evolution of the magnetic techniques that made the hard disk drive a mainstay of today's computing architectures. The latest attempt to change that is driven by the many backers of TeraStor Corp., who feel that despite a number of setbacks, their concept for adding optical recording techniques to magnetic storage will take capacities well beyond the capability of magnetic storage alone. While there are a number of earlier techniques for boosting magnetic capacities with optical technology, TeraStor's

founders feel their concept has far more potential. TeraStor burst on the scene early in 1997, boasting that it had "breakthrough technology which will enable a new class of rewritable mass storage products delivering a sustainable ten-fold capacity advantage at a lower cost-per-gigabyte than all existing storage technologies."

The industry stood up and listened because TeraStor had a number of well-known founders and a solid financial support. Although prospects of near-field recording were downplayed by some in the drive industry, this form of optically assisted Winchester disk drive was examined by other leading companies, including Seagate Technology.

TeraStor unveiled a disk drive with removable media in June 1998. But it didn't ship. The bad news about TeraStor's delays was compounded by a fadeout of research dollars being spent by Seagate and others on optically assisted technology.

But that hasn't dimmed the gleam in the eyes of those who believe in near-field recording. Investors definitely feel that the company is on to something.

In 1999 alone, TeraStor completed two rounds of financing, with $20 million and $23 million. Many companies are also hoping to cash in on the technology. TeraStor has announced a slew of strategic relationships, many with prominent companies-Hitachi, Quantum, Olympus and Maxell. Many of them have invested R&D dollars to create components for the TeraStor drives.

Now, TeraStor executives expect to unveil a working product in the middle of 2000, saying that a few novel technologies were more difficult to perfect than expected.

"This has taken us a lot longer than we expected. Things like evanescence and flying the heads over plastic were all things that were new," said Gordon Knight, chief technology officer. "When we started, we felt our near-field techniques provided a 20:1 advantage over far-field optics and other techniques. Nothing has changed. We will have a 20-Gbyte product: two sides with 10 Gbytes that uses two heads. I don't see any reasons we can't go well over 200 Gbytes per disk."

But first they have to get the 20-Gbyte version out. That effort suffered a major setback last year when Seagate's autonomous thin-film head operation pulled out of TeraStor's supplier base for reasons unrelated to its decision to scale back its optically assisted Winchester efforts. If ever there was a depressing time in the oft-delayed program, that was it.

"Our lowest point was when our head partner dropped out," Knight said. "That set us back 18 months."

Heads are one of the most critical aspects in the design, even without the trauma from Seagate's departure. One of these head-related issues is the same difficulty that magnetic drive designers are facing. Given the low-flying heights and thin track widths used today, the temperature changes can have a great impact in positioning the head over the desired track. An obvious difference is that when TeraStor's laser heats up the disk, the temperature change is "tremendous," Walker said. The company has focused on laser heat this year.

"The head-to-disk interface is where we solved a lot of problems in 1999," Knight said. "What happens is that when we heat up the optical layer, the head is flying a few microinches above the surface, so there are major heat issues we had to solve. We have about 35 patent applications pending, and many of them are in the head-to-disk interface."

With all that effort, every aspect of the business was in an upheaval. Marketers who thought they would soon have a product were once again resigned to talking about potential after the head vendor upheaval.

"Last year, we thought we were a lot further along than we were," said Richard Walker, marketing vice president. "One reason is that we lost our head development partner. We spent a lot of time finding a new partner, Yamaha. For us, 1998 was quite painful. In 1999, we made a lot of progress."

He explained that the time to get Yamaha's production line ready slowed an already lengthy development cycle for heads. During that transition, the head developers decided that they could upgrade the design, something that might not have been done if they did not switch partners.

"The impact of technical issues were very evident in the last year, and they caused us to go through a fundamental redesign of the head," Walker said. "The head development process is very time consuming. Once you complete the design, you wait three months until you see the first head come off the wafer production line."

After all the delays, the company is eager to get out and start competing in the open marketplace. Marketers will focus on three markets for removable storage.

"There's the traditional backup architecture, where tape used ranges from Travan to the low end of the digital linear tape technology. Where we'll come in is where the DDS drives fit," Walker said. "Our removable products also fit in creative applications, like the Hollywood people who are dealing with very large files that they need to pass from one to another. The third market is vertical applications like medical imaging that are done today with magneto-optics because of the long shelf life. Our media will be good for 30 years just like M-O."

In these markets, pricing will be critical. If the drives are more costly than tape drives, it will be extremely hard to get potential customers to switch from tried-and-true tape. Walker said that TeraStor's approach of spreading development costs between several partners will help keep the end-unit pricing low.

"That's one of the fundamental strategies of our approach. The people who partnered with us have put millions of dollars into this to get us as far along as we are."

Though TeraStor spokesmen once seemed ready to take on the well-entrenched magnetic disk drive industry, that rhetoric has cooled. As with others who have predicted that new techniques would put the old war horses out to pasture, TeraStor's managers now concede that the hefty expenditures of the firms involved in an established technology often lead to advances that take it well beyond perceived limitations.

"There's so much money for R&D in magnetics that we're seeing new areal density levels all the time. IBM's latest demonstration was 30 Gbits/square inch" Knight said. "Everyone is talking about the super paramagnetic effect, and we still believe that will occur around 40 to 50 to 60 Gbits/square inch. The temperature at which bits self-erase gets closer to room temperature all the time. With all the money being put in, there's sure to be a solution to that particular problem."

Knight doesn't have any concerns about the viability of near-field recording, saying that it also has plenty of backing.

"The beauty of near-field recording is that it can go well beyond both optics or magnetics," Knight said. "We've got an extremely powerful set of partners who have invested heavily in our technology. Some of our concepts are even being explored by the magnetic guys. We're seeing a lot of magnetic companies starting to look at thermally-aided media."

The Century of the Engineer: Misunderstood Milestones