High Performance? Hiub Density? We Need Both.

The next uptick in the optical storage market will be driven by...what, exactly?

I asked two prominent advocates, each representing a trade group with a different core technology to promote. But both men see clearly that the next uptick in the optical market will come from multi-drive jukebox/library installations, rising sales of which car be achieved only through cooperation among developers, vendors, integrators, and resellers.

Dan Dalton is director of Optical Products at Fujitsu Computer Products of America He chairs the High-Performance Committee (HPC) of the Optical Storage Technology Association (OSTA), which focuses on moving 5.25-inch Magneto-Optic (MO) solutions forward.

Dave Doering is a networking consultant at TechVoice, who helped to form and now chairs the High Density Storage Association (HDSA), an OSTA spinoff that emphasizes 120mm (4.72-inch) phase-change media such as CD and DYD, particularly in libraries and jukeboxes.

In separate interviews each chairman described his organization's goals not only in terms of core technologies, about which they differ, but in terms of user benefits about which they are largely in agreement. Their visions of the marketplace come together in an understanding that no matter what recording technology or media form-factor is involved, the future belongs to the automated library or jukebox rather than to the single drive.

Scalability Determines Price

"The jukebox market sold less than 10,000 units last year," said Doering (citing IDG analyst Wolfgang Schlichting). You can't make an industry out of that! We need to sell at least 3,000 units a month to make an industry. And we need to sell 10,000 units a month to make a revolution.

"At such a low a number," he went on, "is it any wonder that we're not seeing innovation in NAS and SAN devices using jukeboxes? And yet, if we don't work on improving jukebox attachment now, or if we cut back on the R&D it takes to do that, nobody will sell anything. Then we'll all have graceful little solutions that nobody has ever heard of."

Dalton too feels that jukeboxes are the key to the market, but he believes that MO serves users-particularly users running heavy-duty applications-better than phase-change recordables such as CD and DVD can. "In a library application, performance is critical," he said. You have to look at factors like how long the disk remains in a drive, which may be more critical in some apps than in others. MO has the edge in over-writability: one million to 10 million overwrite cycles, vs. maybe 10,000-100,000 for phase-change. Remember: it's not so much that people overwrite disks that often, but that file systems do. The OS may be doing rewrites unseen by the user.

"Historically," Dalton added, "MO's key differentiator was that it had a longer archival life than tape. After twelve years in the industry, I can say that MO has the longest archival life of any removable media. MO systems are higher-priced than DVD-recordable systems, it's true. But cost is tied to manufacturing volume. Scalability determines price. CD-recordables and DVD-recordables are simply being made in larger quantities than MO. But MO has faster seek times and faster data transfer rates. It's really for the 'performance' segment of the market."

Archiving, of course, is where demand for capacity is most likely to grow. So Dalton declared: "the HPC's focus is to show not only the progression of the technology but also that customers who have invested in MO aren't obsoleting their product. I don't think you can characterize it only as a replacement market; our intent is to attract new customers, through individual marketing activities. The HPC has identified several potential new market segments, including e-commerce and Internet apps, where there's a critical need for maintaining the records of online transactions."

Doering concurred that the spread of Web browsing has helped to draw attention to record-keeping as a driving force behind growing storage demands. "We now have more users who want more information," he said, "and with the browser having become essentially a standard, we have a familiar user interface that nobody has to re-invent for every application and installation."

But on the hardware level, Doering considers the competition for the record-keeping market to come not so much from MO as from magnetic HDDs and RAIDS, which get a lot of attention because they're extremely fast at retrieval and throughput. "Suppose you want to read back-issues of a publication like Computer Technology Review. To put all those files on HDDs or RAIDs just because they're fast seems like a waste. A jukebox gives you access to those back issues, and users will wait through the latency because they're getting what they couldn't get cost-effectively otherwise."

Footprints Matter

Dalton concurred that RAIDs are inappropriate as archives, compared with MO. "The highest dual-sided capacity point ever [9.1GB] was accomplished through magnetic superresolution and land-and-groove recording," he said, "and by shifting from a 680nm laser to the slightly shorter 660nm wavelength, which enables new drives to read every earlier media generation back to 650MB, and yet also to write to 2.6GB and 5.2GB media."

Ensuring full downward compatibility with legacy systems while offering unprecedented capacity and swift data transfer rates certainly entitles the MO group to use "High Performance" in its committee's name.

The HDSA put "High Density" in its name, according to Doering, because "We wanted the user benefit in the name of the organization. 'Automated Library Association' is descriptive; but if we'd said that, it would signify only the players in the organization, and not the concept or the benefit. And libraries do offer high-density storage in a small footprint. There's a library that can hold up to 700 disks in a yard-high cylinder with a picker in the middle. The drive can be anything in a half-height form factor, so with DVD-RAM you can put about 3TB of data in that box. That's what we call 'high density,' and at relatively low cost.

"I hope the HDSA can unite the vendors and the large end-users. We've had good response from the U.S. government," Doering said, citing what is by anyone's definition a large end-user. "Government IT administrators tell us they want to make the "buying equation" easier. Right now, the purchase of an automated jukebox involves a hardware vendor, a software vendor, an application-tools provider, maybe even an OS manufacturer. That's too many steps in the buying equation. We want HDSA to be a place to work this out."

In the technology arena, however, HDSA members are focused on 120mm systems because, as Doering put it, "MO doesn't need cheerleaders any more: the MO advocates will continue to have their installed base for years. But whatever kind of drive is inside, the automated library should be a part of every [computer] system as its archive. And the way to do that is to have an organization that can create promotional materials that are vendor-agnostic, and to come up with ways to help give IT departments the ammo they need to promote the idea of libraries with senior management. You know the dairy industry's 'Got Milk?' campaign? Well, I'd like HDSA do a campaign to build awareness, called 'Got box?'