August 11, 2003 4:00 AM PDT
Midlife crisis for the hard drive
The amount of data that can be crammed onto a single platter in a computer hard drive has increased at a torrid pace for the past few years, even doubling annually during some periods. But that growth is beginning to slow as engineers run into technological obstacles and many PC buyers feel they have more than enough space.
The transition from 40GB per 3.5-inch platter to 80GB--the standard for most current PCs--took less than two years. A similar leap, however, may take three years, and could impact pricing, innovation and competition throughout the industry.
"There's no question" density growth has slowed, said Matt Massengill, CEO of drive maker Western Digital. "For some period of time you'll see some slower growth than 100 percent."
Ashok Kumar, an analyst with U.S. Bancorp Piper Jaffray, predicted data density may not double again for another two to three years.
"It is not the result of poor decision making or bad execution," Kumar wrote in a report last week. "It comes from fundamental limitations in the technology." In the near term, he said, the industry may develop platters that increase capacity by 25 percent to 50 percent.
The fallout could be broad and disruptive. Drive makers, for example, may face pressure to cut prices, and market dynamics may favor technology followers rather than leaders. It's also possible industry players will focus less on density gains and more on smaller drives for fast-selling items such as music players and laptops.
Developed in the 1950s, hard disk drives contain heads that read and write magnetized data onto platters. The devices are vital to PCs, servers, dedicated storage devices and consumer electronics, such as the iPod music player.
Drives with storage capacities of more than 200GB are currently on the market, but these include multiple platters. Increasing the amount of data that can be stored on a single platter allows manufacturers to reduce the number of platters while achieving the same overall capacity, which cuts costs.
Estimates on the growth in data density vary slightly, but the overall trend is clear. Currie Munce, vice president of research at disk drive maker Hitachi Global Storage Technologies, said platter density increased about 30 percent per year from the time the first hard disk drive shipped in 1956 to 1991. It then climbed to roughly 60 percent per year from 1991 to 1997, and it doubled annually from about 1997 to about 2001.
The pendulum may now swing back, with density growth slowing to the 40 percent to 60 percent range in the next few years, Munce said. "I think we're returning to the more historic growth rates that we've had in the industry."
One reason for the slowdown is reduced demand for pumped-up platters in desktop PCs, said Jim Porter, president of market research firm Disk/Trend.
"The part of the market that is clamoring to get more than 80GB per platter is relatively small," Porter said. He estimated that PCs make up about 75 percent of the market.
Hitachi's Munce agreed that "demand for increasingly larger hard drives has slowed down, in particular in the corporate desktop markets, where capacities are growing only about 20 percent to 40 percent per year."
Technology obstacles loom
There also are technological barriers to higher data density. Munce said one factor is the difficulty of creating ever-smaller heads. The next generation of disk drives will have heads with elements that are smaller than 0.1 micron, he said, which is as small as the features found on cutting-edge microprocessors. A human hair is about 50 microns wide.
"That's a challenge for manufacturers," he said.
Aside from head-manufacturing difficulty, another technological obstacle to higher densities is what's known in the industry as the "superparamagnetic limit." The term refers to the limit of storing a bit of digital information in an ever-smaller area on the hard disk.
As the space on the disk that is devoted to a bit--the magnetic domain--shrinks, problems are likely to arise, including the loss of magnetic charge, Porter said. "If your magnetic domain is too small, it won't be stable."
Engineers have surpassed earlier estimates on how small magnetic domains could get, but the limit seems to be approaching, Porter added.
Tech leaders could suffer
The slowing in capacity growth will affect the competitive landscape. Typically, drive makers prefer to increase density in their products rather than lower prices. With density gains declining, pressure will increase to reduce prices. For example, a company might sell an 80GB drive for $150 this year, but by next year consumers will expect 120GB for the same price, or to pay less for that 80GB drive.
On the other hand, stepping off the capacity treadmill may help companies that concentrate on low-cost production, such as Western Digital and Samsung, Kumar said, rather than those that gain revenue from pushing the edge.
Kumar mentioned Maxtor as an example of a technology leader that could be hurt because the advantages of having cutting-edge research and development are becoming less important.
"Longer product lifetimes are a disadvantage to technology leaders," he said. "They depend on having new products."
Stephen DiFranco, vice president of corporate marketing for Maxtor, disagreed with Kumar's description of his company.
"I would say that Maxtor is better at accessing and integrating technology" than doing pure research, he said. "We're not a white-lab-coat company." To back up his point, DiFranco said Maxtor relies on suppliers to provide disk drive heads, and buys about half of its disk drive media from external providers.
Then again, cutting-edge technology may still prove to be vital. For instance, companies that are able to achieve breakthroughs in reducing the size of disk drive heads--the devices that actually read and write data on a disk-?could grab a technological lead over their rivals. Western Digital essentially jumped into the race for better heads when it bought most of the assets of head maker Read-Rite earlier this year.
But Kumar viewed that move with skepticism, arguing that drive makers do not have a good track record in taking over head companies.
"It is something like driving an economy car all your life and suddenly taking the wheel of an 18-wheel tractor-trailer as it heads down from a mountain pass," Kumar wrote. "They need to learn a whole new way to run a business quickly, on the fly, and with no margin for error."
Massengill defended the Read-Rite deal. "It's certainly a complicated business, one that requires near-perfect execution," he said. "But that's the business we're in."
Given tepid demand for higher-capacity platters in desktop PCs, some companies are turning their attention toward small, lightweight, quiet and power-efficient drives that fit into laptop computers and consumer electronics devices.
Hitachi will soon produce 1.8-inch diameter hard drives. Drives of this size are currently found in the Apple iPod and have proven extremely popular. Today, Toshiba is the main producer of 1.8-inch drives.
Hitachi also plans to sell a 4GB 1-inch "microdrive" for use in digital cameras and other devices. Start-up company Cornice also is getting into the microdrive game, having developed a 1.5GB, 1-inch hard drive for consumer-electronics devices such as digital video cameras.
New tech to keep drives alive
Meanwhile, alternative data recording methods are in the works. One oft-mentioned approach is "perpendicular recording," which involves recording data in vertical, three-dimensional columns rather than in not in two-dimensions on a disk. Shifting to perpendicular recording methods will require new disk media, new heads and new electronics to control the heads, Massengill said.
Perpendicular recording promises a lot more capacity, similar to the way cities accommodate population growth by building high-rise apartments and office buildings.
"If land is really expensive, you build a skyscraper," Massengill said. "That's kind of what you can do with perpendicular recording." He expects it will take three to four years before companies start shipping large numbers of perpendicular recording devices.
A spokesman for drive maker Seagate said the company has successfully demonstrated both perpendicular recording and another method called "heat assisted magnetic recording." Seagate also has been working on ways to extend the current technology, called longitudinal recording, the spokesman said.
There's also the possibility that some other data-storage technology could emerge. Porter says one candidate is magnetic RAM, or MRAM, which is a semiconductor chip that can hold data even when electrical power is switched off. But such alternatives are probably five to 10 years away from potentially competing with hard drives on a price-per-gigabyte basis, Porter said.
So even with slower data density gains, the time-tested hard drive will likely keep whirring inside of computers and other devices for the foreseeable future.
"It's really one of the great inventions of the computer age," Porter said.