A plethora of applications begs for reasonably priced, high-capacity, fast, removable media. Examples include storing downloaded files from on-line sources, such as the Internet; transferring large data and code sets to computers at field sites; and storing and playing multimedia presentations. These applications strain the capacity of internal hard disks and could benefit from a better data-transfer/storage medium. On the other hand, many people haven’t used a 1.4-Mbyte floppy since they last needed to boot a new version of their operating system for the first time. It’s time to bury the floppy, and capable replacements now exist. One of those replacement candidates could best serve the industry by becoming a universal standard.

For most people, a floppy disk is, at best, fourth choice for storing or transferring anything more than a single data file. Better options for transfers include a network connection, such as most business users enjoy even with notebooks, a direct parallel- or serial-port connection that anyone with a cable can use, or e-mail. Better storage choices, other than a local hard disk, include a network server; a tape drive; or a second, third, or fourth hard drive.

The panacea— a universal 100-Mbyte to 1-Gbyte removable medium with drive performance approaching hard-disk performance— might be just around the corner. For brevity, let’s call it the A-Drive. An A-Drive could store and transfer data and provide primary on-line or near-line operation alongside the primary hard disk.

An A-Drive standard is important not only to PC designers and users. In today’s heterogeneous environment, the next A-Drive standard will pervade PCs and workstations, embedded systems based on PCs, and even embedded systems based on platforms such as the VMEbus. An example evidences the need for this standardization: Dennis Kramer, the team leader for next-generation products at Rockwell’s Automotive Division in Troy, MI, needs a 100-Mbyte, less-than-$100 removable device to use in auto navigation systems. The only way such a device can reach that price point is through standardization in PCs.

New consumer devices, such as video games, cameras, set-top boxes, and even the Internet appliance that some prognosticators envision, will all benefit from a high-capacity, universal medium. The PC is simply the platform that can act as the least common denominator to establish the standard.

By all rights, the industry should have already moved on to a new removable storage standard. Devices such as the 20-Mbyte floptical disk and the 128-Mbyte, first-generation 3½-in. magneto-optical (MO) drives were pretenders, but no device to date has made a serious attempt to unseat the floppy drive.

The reason that no new standard has emerged is the lack of a leader among the system vendors. Since IBM lost the leadership position in PCs in the microchannel days of the ‘80s, no system vendor has stepped forward as a leader. In its heyday, IBM could demand greater profit margins on its systems. These margins allowed the company to introduce a new technology into mainstream systems before the technology was truly cost-effective. The IBM introduction established a standard and ramped volumes, thereby lowering the price, and everyone benefited.

Today, all system vendors worship market share. Market share, especially with discount superstores selling so many systems, means profit margins of $50 or less per system. The margins allow for little experimentation with new storage devices, despite evidence from the retail add-on market that users will pay for a capacious A-Drive.

The leaders in the PC arena today are a software company based in Redmond, WA, and a chip company based in Santa Clara, CA. Unfortunately, neither cares about new removable storage devices unless the new device offers a direct way to boost the sales of software or µPs, respectively.

Well, the market has decided not to wait for a system vendor to lead the charge. Consumers have been scooping up devices, such as Iomega’s Zip, as soon as they hit the retail shelves, and Syquest isn’t far behind with the EZ135.

Whether you design or use desktop computers, embedded systems, or workstations, expect to confront the problem of choosing a high-capacity A-Drive soon. You must evaluate the available products on the basis of drive price, media price, capacity, performance, and reliability. Moreover, you have to judge the intangibles that could anoint a new A-drive standard, regardless of your engineering evaluation.

Eliminate pretenders

Before you consider each of the evaluation criteria against some contenders, eliminate some devices that won’t soon serve as floppy replacements. Tape drives can serve only in niches, but there are some advancements in that area (see box, "Software makes tape suitable for near-line storage"). Flash memory is far from cost-competitive with rotating storage, but, again, there are advancements for the compact niche (see box, "Flash targets miniature removable storage").

Recordable or rewritable CD-ROM drives can’t serve as an A-Drive. Despite falling prices, CD-recordable (CD-R) and soon-to-be-available CD-erasable (CD-E) drives lack the random-access and read/write characteristics an A-Drive requires. Writing to CD-R or CD-E drives requires that you first premaster, or arrange the data in a contiguous footprint, on your hard disk and then transfer the data to the CD-ROM. CD-R or CD-E may replace your standard CD-ROM, but, even with a bootable CD-ROM spec on the way, CDs are not the A-Drive answer.

The next-generation 4.7-Gbyte CD-ROM, now called the Super Density Disk (SD Disc), can’t fill the A-Drive role either. Originally called digital video disc (DVD), the SD Disc is now on track to standardization with competing camps agreeing on format issues. Drives will debut at Fall Comdex ‘96 but won’t be cost-effective. Expect widespread adoption in the fall of ‘97 as a replacement for the CD-ROM. Recordable and erasable drives will follow in ‘98 but will suffer the same drawbacks as today’s CD-ROM in the A-Drive role.

Another removable technology that enjoys success in niche applications, docking stations that allow you to remove an entire hard-disk mechanism, also won’t become the next A-Drive. These products range from PCMCIA hard disks to subsystems that accept a custom-packaged 3½-in. disk drive. Buying an entire hard disk as a plug-in cartridge is simply not cost-effective. Moreover, hard-disk drives are relatively fragile: You can’t throw them into a briefcase or drop them onto a table.

Luckily, we have a solid lineup of products that can fill the A-Drive role, and more should emerge over the next few months. The products fall into three categories (see Table 1). The first category includes drives that use flexible magnetic media but hard-disk-like recording heads. The second category includes rigid magnetic-media drives— essentially true hard drives with the disk platters in a removable cartridge. The third category is optical, including the hybrid MO scheme and a true phase-change optical product in Panasonic’s PD/CD-ROM.

Begin your comparison of A-Drive products from the angle of drive and media price, because the floppy disk is, quite simply, dirt cheap. OEM system vendors pay as little as $20 for a floppy drive, and you can walk into any superstore and buy a diskette for around $0.25 in a 50-pack box. Therefore, anything that replaces the floppy has at least to promise low price in the future.

Table 1 — Contenders for the floppy-replacement title

Technology	Capacity (Mbytes)	Product	Manufacturers	Form factor and interface	Seek time (msec)	Transfer Rate (Mbytes/sec)	Media Price	Drive price	Availability	Comments
Flexible magnetic media	100	Zip	Iomega and Epson	External SCSI or parallel 5¼-in. SCSI 5¼-in. SCSI with floppy 3½-in. IDE	29	1.4	$20	$199	Now	Customers have made Zip a best seller in the retail channel. Iomega now turns to OEMs with Micron Electronics the first to sign on. Epson is credible second source, but media is tough to manufacture and in short supply.
	120 floppy	High-capacity (Panasonic)	MKE Interface N/A	3½-in.	N/A	N/A	N/A	N/A	Q2'96	Uses optical-tracking technology and reads/writes standard floppies. Three partners — MKE, Compaq, and 3M — all have more pressing matters on their plates.
	128 drive	High-Density	Mitsumi floppy-disk	3½-in. IDE	N/A	2.45	$20	$130 (100)	Q3'96	Largest manufacturer of floppies has technology that can read/write standard floppies. Company has OEM focus and won't build more than a prototype before receiving a huge order.
Rigid magnetic media	100	SQ1100	Syquest	1.8-in. PCMCIA/ATA	16	1.3	$49	$299	Q1'96	Specialty product for miniature applications.
	135	EZ135 Drive	Syquest	External SCSI 3½-in. IDE	13.5	4	$20 (3)	$299 $199	Now	Giving Zip a run in the retail channel. Offers better performance at the same price, but can price drop as fast as Zip?
	170	HARDiskette	Avatar 3½-in. SCSI with floppy	2½-in. IDE	12	2.1	$49 $499	$399	Q1'96	Late arrival with OEM focus. Product scales from notebook to workstations, but Avatar isn't pushing hard to reduce price.
	270	SQ3270	Syquest	External SCSI 3½-in. IDE	13.5	4	$55	$400	Now	Two disk version of EZ135. Provides an upgrade path and can read/write EZ135 cartridges.
	540	MCD 540	Nomai	External SCSI 3½-in. IDE or SCSI	10	6	$50 to $60	$599	Q1'96	Former supplier of Syquest cartridges enters drive market with product that can read/write EZ135 cartridges.
	1000	Jas 3½-in. SCSI	Iomega	External SCSI	12	6.7 $599	$99	$499	Q1'96	Iomega has shipped samples of newest product and has the interest of verticle markets.
1300	SQ31300	Syquest	3½-in. SCSI	11	6		$600	Q1'96	Not formally announced. Will set new capacity/price target if Syquest delivers but, like Iomega's Jaz, is currently too costly as floppy replacement.
Optical media	140	MD DATA	Sony 3½in. SCSI	External PCMCIA or SCSI	500	0.15	$29	$749	Now	Sony has shipped more than 1 million audio players using this technology but hasn't pushed hard in data-storage market. Plans to use this drive in other applications such as cameras.
	140	MD DATA	Sharp	2½-in. floppy interface	500	0.15 (10)	$6.50 (10,000)	$200	Q1'96	Sharp, with help from National Semiconductor, has simplified interface, lowered price, and targeted OEM market. Will fit in notebook, and Sharp plans to extend capacity/performance ramp-up.
	230	Magneto Optical	Epson, Fujitsu, Olympus, Sony, and Toshiba	External SCSI 3½-in. SCSI and IDE	30	5	$25	$400 to $800	Now	Only removable technology with support from many vendors and established road map for capacity/performance. Has potential for low price if "chicken-and-egg" problem falls and volume ramps.
	650	PD/CD-ROM NEC	Panasonic and 5¼-in. SCSI	External SCSI	165	1.1	$60 $650	%500	Now	Suprise entry as Panasonic drives price of dual-mode drive down. Combines 4x CD-ROM with phase-change optical, and Compaq has started to ship in high-end systems.

Drive and media price

Zip is the price leader today. Moreover, the simplistic design of the product should allow for price cuts, as Epson becomes a second source and volume from both companies continues to grow. Robert Abraham, vice president of storage-analyst company Freeman & Associates in Santa Barbara, CA, believes that Iomega could sell the Zip for less than $100 on an OEM basis in the near future.

To date, Iomega has experienced more problems delivering the media than delivering the drive. The actual media and cartridge are not problems, but the medium must be servo-written once Iomega receives it. A shortage of Zip cartridges has existed on store shelves, but the price of the media can drop over time. The price will never approach floppy-disk prices, however.

Judging the rest of the lineup on price is more difficult. The other flexible-media entries, from Mitsumi and the Compaq (Houston)/MKE (Panasonic’s parent)/3M (St Paul, MN) partnership should be low-cost, but neither has yet delivered on product promises. Syquest is selling the EZ135 to compete with Zip, but, realistically, Syquest’s product must be more expensive to produce. Avatar, meanwhile, has just entered the market and isn’t currently cost-competitive. The Avatar prices in the table are from one of the company’s subsystem OEMs, APS Technologies (Kansas, MO), which sells PC and Mac add-on products.

The optical group holds the greatest long-term promise for low cost. An MO drive may not match the cost of Zip or one of the other flexible-media choices, but MO prices could quickly drop to much less than $200 in floppy-like volumes. Moreover, 3½-in. MO media could easily sell for less than $5.

The MD (MiniDisc) DATA camp has similar low-cost potential, despite the fact that Sony has not aggressively marketed the technology for the data-storage market. Sony has sold more than 1 million 2½-in. MD mechanisms in the worldwide audio market and plans to use the data-storage version in consumer products, such as cameras and personal digital assistants.

Sharp and National Semiconductor (Santa Clara, CA) have made inroads into lowering drive prices by developing an extension to the floppy interface. The interface links the Sharp MD DATA drive on the floppy cable. Blank MD cartridges already sell for as little as $6.50 in Asia, and, although the cartridges target audio recorder/players, they work equally well in data drives.

The contenders for the A-Drive slot feature capacities from 100 Mbytes to more than 1 Gbyte. Today, any of these choices would be a vast improvement over the floppy, although, as always in the computer business, the bigger the better.

Time is short for low-capacity products. If it takes two or three years to establish an A-Drive standard, then the capacity had better be much greater than 100 Mbytes, because, in three years, standard hard drives will store 4 Gbytes.

The key to capacity comparisons, therefore, may be future extensions. Mitsumi, for example, claims that it can double or triple the capacity of its announced high-density floppy. Iomega and Syquest both have 1-Gbyte removable magnetic products that could be price-effective in the A-Drive role in three years, and those products will certainly enjoy short-term success in vertical markets. Sharp plans to take MD DATA into the 650- to 700-Mbyte range in 1997, but Sony may not share the same road map.

MO drives are the only product class with a defined future road map (see Figure 1), and multiple major vendors have committed to the future. The ISO has adopted a 650-Mbyte MO standard, and the 650-Mbyte products should appear by midyear. Moreover, capacity jumps in MO drives typically happen with little or no increase in drive or media price. The Optical Storage Technology Association (OSTA) has clearly defined the MO road map through the end of the decade.

Hard-disk-like performance

Performance becomes an issue for removable media as capacity escalates. Low seek times and high transfer rates allow the removable devices to effectively transfer large amounts of data. Moreover, many applications for these new devices, such as playing multimedia presentations, require that the removable device work alongside the hard drive.

Table 1 shows that most of the A-Drive contenders offer relatively good performance, with the rigid-media devices leading the way. The optical products trail the pack. The 3½-in. MO drives are workable and will scale up in performance with the coming capacity ramp. Moreover, new direct-over-write media will boost write performance, which has always been a problem for MO devices.

The MD DATA drives could use a performance boost in seek time and transfer rate. The drives are currently faster than a floppy by a factor of five but are far from hard-drive performance. MD DATA uses an MO recording scheme and has adopted direct-over-write capabilities. The answer to boosting speed could be a faster-spinning drive with faster linear head movements.

Floppylike reliability

Any floppy-replacement contender had better be reliable. Users have been successfully spilling coffee on, sitting on, and throwing diskettes for years and still recovering data. The removable products that are already shipping have shown fairly good reliability in the field. The graphics-prepress industry has for years been stuffing Syquest’s 5¼-in. cartridges into overnight-delivery envelopes and shipping them worldwide. Moreover, all of the devices sport impressive MTBF specs.

After you review the anecdotal evidence on reliability, however, you can make few judgments on how reliable any of these devices would be if manufacturers shipped the devices in every PC. In general, the optical devices should have fewer reliability problems, because they don’t depend on heads that fly microinches above delicate media.

Among the magnetic vendors, Avatar, Iomega, and Syquest have invested thousands of hours on testing their products in chambers contaminated with everything from powder to sandpaper particles to road dust. With truckloads of Zip and EZ135 drives entering the market daily, a reliability profile should emerge from the market over the next few months.

Intangible factors

After reviewing engineering criteria, consider less tangible factors that could ultimately determine the A-Drive winner. Computer-industry history tells us that the best technology has little to do with market winners, so intangible factors will likely act as the trump card.

One factor could be backward compatibility. The Mitsumi and Compaq/MKE/3M floppy efforts will likely both produce a drive that can read and write 1.4-Mbyte diskettes. Both scenarios have downsides, as well. Mitsumi won’t produce one drive until they get an order for thousands of units per month.

Compaq, MKE, and 3M, meanwhile, all have distractions. MKE has found unexpected success with its dual-mode PD/CD-ROM, which Panasonic sells. NEC has also signed on to build the dual-mode drives, and Compaq has begun to ship the drives in high-end systems. At the same time, 3M is spinning off its Data Storage Division as an independent corporation, and no one knows what effect that might have on the company’s focus.

Portable form factors

Portable form factors could also play a role, because few of the A-Drive contenders can fit into a typical notebook. Avatar, for example, believes that its HARDiskette’s 2½-in. size will allow the product a back-door path through notebooks and into the A-Drive throne. Gailen Vick, vice president of marketing at Avatar, points out that high-end notebooks provide the profit margins necessary to support a new technology, such as the HARDiskette. Avatar also addresses the backward-compatibility issue by packaging both a 3½-in. floppy and a HARDiskette in a 3½-in. floppy form factor.

Other contenders that can fit into portable applications include the MD DATA drive, MO drives, and Syquest’s SQ1100. The SQ1100 is a specialty product with a cartridge that plugs into a Type III PCMCIA host drive. The SQ1100 might give the new miniature flash modules a run for their money in some applications.

MD DATA drives fit squarely into portable systems, including products smaller than notebook computers. Sony’s audio product that uses the MD DATA cartridge is the size of a deck of cards. Sharp’s MD DATA drive is already in a 17-mm-high package, and the company promises a 12-mm-high drive late this year.

Fujitsu, meanwhile, has shrunk its 3½-in. MO drive into a 17-mm-high package, making the drive suitable for use in a notebook. Apple Computer (Cupertino, CA) has begun shipping the Fujitsu drive in some PowerPC-based notebooks.

Iomega, on the other hand, doesn’t think the A-Drive must fit inside the notebook PC. The external Zip drive weighs only 1 lb, and you can easily carry it in your notebook-computer case. The external Syquest EZ135 weighs 2 lbs and is about the same size as the Zip.

Market makers

The large companies building drives would have you believe that you need "market makers" to establish a new technology, and this philosophy proved true in the development of CD-ROM. For the A-Drive slot, the market makers have yet to fire their best shot.

We still don’t know when the Compaq/MKE/3M troika will produce a product, but Compaq could be the closest thing we have to a leader among the system vendors. Sony and Sharp could be sitting on the best mechanical platform of all with an initial compelling capacity and synergy with consumer markets. The companies, however, are not yet cooperating on performance improvements and a capacity road map. The MD DATA drive also uses Sony’s Universal File Format rather than the OSTA format that 3½- and 5¼-in. MOs, CD-ROMs, and SD Discs use.

MO has the most impressive list of manufacturers that are sitting on the longest-running "chicken-and-the-egg" impasse in the data-storage industry. The vendors succeed at selling the drives in Asia, and Sony and Toshiba currently market their drives only in Asia. Meanwhile, the potential customer still faces the problem of low volumes leading to high prices leading to low volumes. . .

Market winners

Until some eggs hatch, you may be better off looking at market winners. Iomega is clearly winning in the retail market. Moreover, system vendor Micron Electronics announced at Comdex that it would begin selling the internal IDE Zip as an option starting this month. Micron (Boise, ID) Product Manager Jeff Moeser believes the option will immediately prove popular and lead to a standard slot for the Zip in Micron systems.

Syquest is also doing well in the retail channel but has yet to land a major OEM customer. Panasonic and Fujitsu have PD/CD-ROM and MO drives that Compaq and Apple, respectively, are shipping in their products. Adding NEC as a second source should bolster the A-Drive chances of the PD/CD-ROM.

The most impressive contender on the OEM side at Comdex may have been upstart Avatar. Acer America (San Jose, CA) is offering the HARDiskette as an option on some notebooks. Intergraph Computer Systems (Huntsville, AL) plans to offer the drive in some of its workstations. APS Technologies has announced add-on kits, and Olivetti (Ivrea, Italy) has announced plans to use the HARDiskette in some of its computers.

Avatar has also announced that several OEMs in vertical markets will use the HARDiskette, proving just how widespread the A-Drive market can be. The vertical applications range from medical imaging and record storage to multimedia presentation products and ruggedized embedded computers.

You can reach Technical Editor Maury Wright at (619) 748-6785; fax (619) 679-1861.