Edge of the Ledge
Federal agencies, corporations and nonprofit centers push storage capacity needs into the quadrillions
By John Makulowich
Driven by the popularity of the Internet, the increasing corporate penetration of intranets, the acceptance of data warehousing and the advent of the age of the knowledge worker, storage is a hot commodity.
And ongoing research co-funded by the government is likely to make it even hotter in the next few years as the envelope of innovation stretches into new media.
Ratchet the metric prefix up another notch. Here come petabytes - a quadrillion bytes or 1,000 terabytes. That's the stage we've reached in amassing digitized information.
One segment of the storage market that's already experienced strong growth since the late 1980s is the compact disk and optical disk drive industry. CD and digital versatile or video disk drives accounted for 97.65 percent of all drive shipments last year, according to a new strategic research study by Frost & Sullivan, an international marketing consulting firm in Mountain View, Calif.
But change is afoot in the CD and DVD drive segment toward newer technologies, according to the study. "These new technologies include compact disk rewritable, DVD read-only memory and DVD random access memory. By 2003, DVD [read-only memory] revenues will be the most significant market subsegment within the CD and DVD drive segment," the study found.
Rufus Connell, a Frost & Sullivan research associate and co-author of the study, feels that the capability of the newer equipment not only to record, but also to rewrite data, will be a market driver.
"It will allow CD and DVD drives to be used for new applications such as data storage, an area traditionally reserved for magnetic tape drives. By the year 2003, we expect CD rewritable and DVD random access memory to have totally replaced CD recordable drives in the market," says Connell.
| CD - compact disk|
CD-R - Compact-disk recordable
CDRW - compact disk rewritable
DDS - digital data storage
DVD - digital versatile or videodisc
DVDRAM - DVD random access memory
DVDROM - DVD read-only memory
FAQ - frequently asked questions
HTML - hypertext markup language
HTTP - hypertext transfer protocol
LAN - local area network
MO - magneto-optical
ODD - optical disk drive
Pb - petabytes (a quadrillion bytes)
RAID - redundant array of inexpensive disks
SQL - structured query language
SSD - solid state disk
Tb - terabytes (one trillion bytes)
URL - universal resource locator
He also points out the explosive growth of the CD jukebox segment, which experienced a growth rate last year of more than 158 percent. Connell quickly adds that not all three segments of the jukebox market are growing at equal rates, that is, CD, magneto-optical and large form factor (12- and 14-inch packages). In fact, he projects that the growth rate for magneto-optical jukeboxes will slowly decline over the next several years as this market begins to mature.
While the major trend is the transition from older drive technologies to newer technologies such as DVD, other trends noted in the Frost & Sullivan study include price declines, performance increases (faster spin rates, shorter access times and higher capacities), the introduction of incremental packet writing for producing in-house databases and using CD media for archival storage.
An affirmative nod for the study came from Eastman Kodak Co., Rochester, N.Y., one of the key players in the CD-recordable marketplace and reputedly the world's largest CD-recordable manufacturer.
In late January, the company announced the official opening of Kodak Cork Ltd., a new disk production facility in Ireland's Youghal, County Cork. It also announced plans to expand its disk production facilities in Guadalajara, Mexico.
When the current expansion is completed, manufacturing capacity will reach 18 million disks per month by the end of 1998, according to a Kodak statement. The facility in Ireland, occupying 45,000 square feet and employing about 250 people, is one of the first in Europe to manufacture recordable disks.
Having and Getting: Two Different Things
Storage may remain abundant and cheap, but having it and getting at it are two radically different issues. In fact, it's a business problem that Richard Tworek seeks to exploit. The executive vice president of Infodata Systems Inc., Fairfax, Va., is pushing the firm's newest product, Virtual File Cabinet or VFC, as the way to "bridge islands of information" within an organization and across the corporate local area network.
"Basically, VFC is a cost-effective and specialized Web-based system for accessing, organizing and sharing information," explains Tworek. "We currently target two kinds of customers. The first are those with no document-sharing system but who need to introduce files into a sharing environment. The second type have a document repository, but want to tie it together."
VFC allows companies to retrieve dissimilar documents on the server produced by other staff. In the company's words, VFC allows you to share information globally, but customize locally.
"The approaches taken in the past to solve this problem tried to replicate databases. We wanted to offer a solution where the customer no longer thinks about the technology and gets to the solution through a single access point.
Of course, software companies did not have the technology we have today nor the penetration of Web browsers and intranets," admits Tworek.
This type of software solution is likely to affect systems integrators in the short term and farther down the road. One reason, Tworek says, is that companies don't need systems integrators to set up the product. In one case, he cites, it took only an hour to completely install the system on an NT 3.51 server. However, he does see a need for integrators, value-added resellers and channels to offer clients the VFC product as one component to a general information technology solution.
The Other Extreme
Most users and companies may be sated in their present storage needs, but the research and development community, including federal agencies, corporations and nonprofit centers, sees the need to stretch the edge of the ledge for the storage solutions of tomorrow.
One approach showing results is the Advanced Technology Program for Data Storage Solutions administered by the National Institute of Standards and Technology in Gaithersburg, Md.
ATP was started in 1990 to help U.S. industry and business pursue high-risk technologies that have significant commercial potential by cost-sharing research projects.
In August 1995, NIST made six awards totaling $38 million to co-fund research on digital data storage technologies. The entire project is estimated at $75 million. The work includes three joint research ventures, involves a total of 20 participants and covers the areas of magnetic media, recording heads, tribology, tracking, channel electronics and software for the next generation of high-performance, mass data-storage devices.
From the standpoint of Tom Leedy, the ATP project manager who heads up the digital data storage, the program offers a number of benefits, not all of them technical.
"One of the lessons we've learned is that the program has allowed companies to work together in joint ventures that probably would not have come together otherwise," admits Leedy. "Further, it allows companies to work with a range of different organizations, such as federal agencies and universities. Probably one of its greatest values is the collaboration it's facilitated."
On the side of technology, he points to two different efforts on optical tape with their own approaches and markets. He believes them to be among the more interesting projects since they fill a niche in data storage that is not currently addressed. He estimates it will take at least three to four years to commercialize a product or pre-product.
With data from the National Storage Industry Consortium, San Diego, Leedy shows how the trends in the hard disk market worldwide have created an interesting problem. While the petabytes shipped per year from 1986 to 1996 have increased dramatically, and the cost per megabyte has decreased in just as spectacular a fashion, the revenues for the industry have remained flat, if not declined.
Thus, part of the rationale for government programs such as ATP is to help keep a critical industry competitive. With flat revenues, it's difficult for companies to support increased research and development budgets to compete effectively worldwide.
Overall, there are two major themes that drive the digital data storage program. First, audio, video and graphical information are being converted from analog to digital formats to meet massive information demands. Second, digital document storage and retrieval is big business, but the traditional ways of doing data storage research and development will not be adequate to meet market needs.
Leedy points to Pacific Bell as an example of future needs. The company records more than 200 million daily telephone calls in a database that exceeds 5 terabytes.
One of the digital data storage projects is the effort headed by Ted Schwarz, president of Peregrine Recording Technology Inc., St. Paul, Minn.
Named High-Performance, Variable Data- Rate, Multimedia Magnetic Tape Recorder, the project seeks to develop a small, reliable and affordable tape recording and cartridge system that can handle the high data capacity and transmission rates needed for specific digital formats.
Among the specific market targets for this technology are digital satellite system television, teleconferencing, terrestrial standard and high-definition television, fiber channel computer applications, and medical and publication imaging.
In the proposed system, data would be recordable at rates greater than 30 megabytes per second; the ultimate goal would be 100 megabytes per second.
Companies taking part in the project include Imation, St. Paul, Minn.; Storage Technology, Louisville, Colo.; Seagate Technology Inc., Santa Maria, Calif.; and Advanced Research Corp., Minneapolis.
| Technically, this [DDS] program is designed to be just beyond the leading edge, not wildly beyond it. Thus, we kept the need for invention to a minimum"|
|- Ted Schwartz|
Peregrine Recording Technology Inc.
Schwarz feels the project is on track with its original goal. "We built slack into our original goal, so even though we had a change in players, we still plan to demonstrate sample heads and some other components by the end of this year and the beginning of next year. The two biggest technical challenges are the heads and the cartridge. After we demonstrate feasibility, we estimate we can reach the beta stage in about 18 months."
The lessons he has learned so far cover a broad spectrum, from business to technology. One of the first was the need to address intellectual property and to work with lawyers from many companies.
"[Intellectual property] issues and legal departments are a major challenge. If you think you can avoid it, you'll be sadly mistaken. Technically, this program is designed to be just beyond the leading edge, not wildly beyond it. Thus, we kept the need for invention to a minimum," says Schwarz.
One major goal is to develop a linear tape drive at one-tenth the cost that can match and perhaps exceed the performance and capacity of high-end, helical-scan systems, a competing technology led by off-shore competitors. He's also seen the need, when dealing with technology in real products, for continuing communication among partners, especially in areas where the players are competitive.
"One of the keys to doing large multicompany projects is making sure everyone communicates. The first thing you have to do is appeal to each partner's self-interest. Each organization needs to see a real reward; to work at building trust. In our specific project, you have to have players with strong competencies. You can't use programs like this to build base competencies. It's just too complex," says Schwarz.
Another DDS program, named A Revolutionary, High-Density, High-Speed, Low-Cost Optical Information Storage Technology, is sponsored by Optex Communications Corp., Rockville, Md.
It addresses the need for data storage and retrieval systems that operate at far higher rates than those available today. The goal here is to develop a new approach to optical disk technology that, combined with high-speed communications links, could, for example, allow the capture and storage, or retrieval, of a two-hour movie in three minutes.
Donald Carlin, senior vice president and chief technology officer at Optex, says the project's goal is to develop a very high data transfer rate storage solution for information.
"There is a need for data capacity. We understand that when we rely on paper files. We understand it now when we look at personal computers. Everyone understands capacity. What we don't understand is the means of getting to and through capacity quickly. If we can't do that, the capacity is useless. One of the sexiest examples is video on demand or video nearly on demand," says Carlin.
This interests large firms such as the telephone companies, which would deliver the product into homes. To Carlin, many of the advantages are obvious: movies that are never out of stock and customers who never have to return a video. It fits well with the advantages of optical disks.
"The first is economics. You can press a CD inexpensively. With DVD, the 4.7 gigabyte capacity is enough for a two-hour, 15-minute movie. You can even add five or seven languages. It's the cheapest ways to publish lots of information. The second advantage is that optical disks are easily removed from drives or players. That's a lot different from hard disks, where a dust particle can cause a crash," notes Carlin.
Using Electron-Trapping Optical Memory, a disk and disk drive technology shown to be feasible under previous research done by Optex and supported by the ATP, the new DDS project could find application in a range of consumer products, including next-generation videodisc recorders, computer storage and virtual reality games. It could also be used commercially in post-production video editing and jukeboxes, that is, automated libraries for storing massive amounts of data.
All the Data on Storage
With the quickening pace of communication on the Internet and the commercial and research advances in storage, it's no easy task to keep up with industry developments. Enter the Usenet newsgroup, comp.arch.storage, and its invaluable two-part FAQ, or frequently asked questions, maintained by Rodney Van Meter.
The computer scientist at the University of Southern California/Information Sciences Institute, Marina del Rey, Calif., covers a broad array of storage topics in the FAQs, including disk, tape, magneto optical, RAID (redundant array of inexpensive disks), SSD (solid state disk), standards, file systems, hierarchical storage management, backup software and robotics.
The FAQ includes postings to the group about everything from prices to programs, URLs to key repositories on selected topics, mentions of bibliographic citations and journal articles, and editorial comments about the timeliness and quality of information. Thus, anyone wanting to explore the depths of a subject, such as DVD, will find enough resources to prepare a senior thesis.
The bearded Van Meter, whose home page sports the statement, "Ask me about my time in Japan, my trips to Nepal and Thailand, and my opinions about Tibet," holds clear and definite opinions about trends in storage.
"Where is the technology going in the next two to three years? I think DVD is going to be a little slower to hit the data market than the consumer market," says Van Meter. "And it will take a while for DVD to be readily available to the consumer market. Generally, the issues revolve around the technology itself, intellectual property and standards."
With cheap and ubiquitous drives, he feels that DVD will ultimately supplant CD-ROM as a distribution medium. However, he does not think that DVD will be used on a day-to-day basis for primary storage.
"DVD is a little bit behind the technology relative to magnetic disks. You have to remember that so far every prediction of a slowdown in magnetic disk technology has been false. It's very difficult to see five years out. Will we have 40 gigabyte disk drives on the desktop? Certainly, the stand-alone machine will continue to have a role. In five years, the idea of a computer not connected to the network will seem quaint," notes Van Meter.
A software person himself, Van Meter feels that the hardware technologists are way out in front of software technologists. For him, the hard problem right now in data systems is managing the data. Another trend he's noticed is robotically managed tape storage, where he feels demand will grow.
"Firms without some form of robotic storage will have trouble controlling their overall information system, especially if they depend on a particular set of disk drives. Still, lots of people with robotic storage [must] continue to update the way they manage their overall systems. With the increased level of connectivity, this will play a larger role in day-to-day operations. There's a real business opportunity for companies to provide Internet-attached, remote backup facilities for small to medium sized organizations," says Van Meter.
He also feels there's an opportunity in the storage area to manage and service the small business network. Part of the reason is that systems are becoming extremely sophisticated and most small businesses lack the knowledge to use them.