Today, the hard drive is found everywhere–from the PCs we use daily to MP3 players and memory keys so small you can toss them in your pocket and forget you’re carrying around a hard drive. But when the hard drive was first introduced on September 13, 1956, it required a humongous housing and 50 24-inch platters to store 1/2400 as much data as can be fit on today’s largest capacity 1-inch hard drives.
Back then, the small team at IBM’s San Jose-based lab was seeking a way to replace tape with a storage mechanism that allowed for more-efficient random access to data. The question was, how to bring random-access storage to business computing?
Enter the RAMAC, 1956
IBM’s answer to this quandary was the Random Access Method of Accounting and Control, dubbed the RAMAC for expediency. The device’s name is a direct reflection of the need for such capabilities in the enterprise. Led by project leader Rey Johnson, IBM’s San Jose lab brought the RAMAC 305 to market.
Recalls Al Shugart, who worked as a field engineer at IBM before joining the RAMAC project and went on to later found Seagate Technology: “They were starting from scratch in the lab. The RAMAC was not just a disk drive, it was a whole system. Nobody had made disk drives before.”
The approach IBM’s engineers came up with represented a clean approach to random data access, notes Shugart: “The concept of the whole disk drive was random access.” To achieve random access, the device would have to move its read/write heads around to different data tracks. “The easiest way to do that,” he says, “was a stack of disks.”
The integrated RAMAC was about two refrigerators in width and not quite as tall, and it literally weighed a ton. Its 50 24-inch platters were in a stack inside the unit, in an assembly that spun at 1200 revolutions per minute. The unit used two magnetic recording heads. The RAMAC could hold 5MB–about the storage that today is needed for one 5-minute MP3 encoded at 128 kilobits per second.
In order to read and write the data, the RAMAC heads moved across a series of circular tracks on each disk surface. Albert Hoagland, who helped build the first drive and is working to preserve the history of magnetic disk technology as executive director of the Magnetic Disk Heritage Center, elaborates: “A shaft ran the length of the disk stack, with a horizontal arm that moved in and out; that arm, which weighed three pounds, had to get to another track in less than a second.”
“The disks’ surfaces were covered with a paint that had magnetic properties–very similar to the paint used on the Golden Gate Bridge,” says Bill Healy, senior vice president at Hitachi (which bought IBM’s storage division in 2003). “They needed a disk with magnetic properties, so it would be magnetically susceptible to recording 1s and 0s; and they needed a read element, such as a disk head, to detect, read, and write that data,” he explains.
The initial prototype, remembers Shugart, “was a relay machine, it wasn’t even a vacuum tube machine. They ended up building 12 of them. From there on, we would design a system for production, including a disk drive. The production [version] was a vacuum tube machine, and I was in charge of designing the computer system for the vacuum tube machine.”
Although the RAMAC shares only passing characteristics with today’s hard drives, it is the drive that launched the industry. “The [technology] industry reinvented itself as the applications for the hard drive changed,” says Healy. “In the fifties and sixties, these devices were made for large corporations, government–the enterprise. The 24-inch diameter platter reduced in size in time. As the devices got smaller over that time, they were mainly aimed at the enterprise environment,” he continues. Disk capacity doubled every two years, a 40 percent compound growth rate.
Adds storage industry analyst Tom Coughlin of Coughlin Associates: “Many companies started to make hard drives for computers, because it was a relatively inexpensive, high-performance way to make mass storage.”
From the late fifties to the early seventies, hard drives were largely used in mainframe computer systems, the kinds found in large corporations and government. The rise of personal computers in the late seventies and early eighties opened the door of opportunity for hard drives–and in turn dramatically influenced where computer technology could go. “With the introduction of hard disk drives,” notes Coughlin, “you had large amounts of storage that were always attached to the computer, and that enabled personal computers to achieve the levels of success they had. A hard drive allowed you to create higher-performance computers with more features because you could have a richer operating system running off the hard drive.”
Related StoriesTrack milestones in hard drive history on “Timeline: 50 Years of Hard Drives.”Learn about the underlying technology in “How It Works: Hard Drives.”Read buying tips in “How to Buy a Hard Drive.”Fast Forward: 50 Years Later
The disk drive has come amazingly far since its introduction: “Today, on 2.5-inch platters we have 15,000 times the capacity of the original IBM RAMAC,” says Seagate Technology Chief Operating Officer Dave Wickersham.
Wickersham notes that the advancement is startling when compared to the pace of other industries: “In the auto industry, to keep that same pace, they’d have gone from fitting five people in the car in 1956, to fitting 160,000 people in that car; or, from getting 25 miles per gallon to 62,500 miles per gallon.”
Today we have drives that cover a range of sizes (the smallest is Toshiba’s 0.85-inch drive, initially introduced in 2GB and 4GB capacities) and specialties. Vendors offer drives optimized for uses in servers, desktops, notebooks, digital video recorders, music players, and more; and you’ll find hard drives in cars, planes, and a wealth of other commercial and military applications.
Prices have dropped dramatically. The RAMAC 305’s cost per megabyte was approximately $10,000–that’s about $70,000 in today’s value. Today, a typical desktop hard drive can deliver that same megabyte for 3/100 of a cent.
Over time, the core recording technology–longitudinal magnetic recording–has remained the same, but the way drives are designed and built has changed. Coughlin reflects: “Heads have gone from the original metal cores, to harder ferrite cores, then to thin-film inductive heads; magneto-resistive heads; and then giant magneto-resistive heads. Now we’re moving to tunneling magneto-resistive heads. They’re now using complex nanotechnologies in magnetic recording heads.”
The media has changed, too. “In the beginning, they used iron-oxide particles dispersed in a plastic binder; then they transitioned in the early eighties to the development of the initial thin-film disks,” continues Coughlin. “By the nineties, thin film disks were the standard, and since then they’ve become more complex, with multiple layers of thin films performing different functions. At the same time, the heads are flying increasingly closer to the disk’s surface,” he explains. The closer the heads fly to the surface, the more data can be stored in a given area–and the drives have become quicker and more accurate. “Increasing electronic integration over the years has led to impressive improvements in head positioning in the detection and decoding of very small signals, and in the correction of errors,” he says.
After 50 years of relying on longitudinal magnetic recording, the industry is shifting production to perpendicular magnetic recording. (For details on these technologies, read “How It Works: Hard Drives.”) The technology was initially explored decades ago, but is only now being used in drive production. Toshiba was the first out the gate in 2005, with its 1.8-inch 40GB mobile hard drive. Seagate was next to the party, with the release of the first 2.5-inch 160GB notebook hard drive and the 3.5-inch 750GB hard drive earlier this year.
Wickersham elaborates: “From an areal density perspective, perpendicular has changed the industry. For a while, areal density was growing at north of 100 percent per year. Then that came down to 10 to 20 percent a year–demonstrating that longitudinal was out of gas. Perpendicular got us back to this 40 percent per year areal density growth; with it, you can quadruple your capacity every four years.”
In our storage-hungry universe of digital downloads and digital photography, increased capacity is a good thing. “There’s up to 60 percent per year growth in storage demand, for the next five years,” continues Wickersham. “The demand for storage is greater than the ability to grow areal density,” he says.
Hard Drives: Future Watch
Hard drives have been indispensable to our computer use for about the last 20 years. Today, hard drives are increasingly indispensable in other ways. “The whole lifestyle has changed–content is king, and we’re carrying data wherever we go. The hard disk drive is the enabler of this,” says Seagate’s Wickersham. “We have 20 disk drives in our home–and there are four of us,” he continues.
Hard drives are in everything from cell phones and digital audio players to set-top box video recorders. That trend will grow, according to industry experts–and provide a fertile new opportunity for the proliferation of high-capacity, hard drive-based storage.
Hitachi’s Healy suggests that the beginning of what he thinks of as the “consumer era of hard drives” can be traced back to 1998, with the introduction of the 1-inch IBM Microdrive. At the time, it stored 340MB in a space just a bit thicker than a standard CompactFlash card.
“A tech-savvy home could easily generate 5 terabytes of cumulative data from 2002 to 2010,” estimates Coughlin. He elaborates: “About half of that would be personal content and half of that would be commercial content. I’m projecting that by the next decade, as consumers become creators of content–a camera on your cell phone is just the beginning–the demand for storage will mushroom, and the line between what’s commercial and what’s personal will be blurred. Personal content will significantly overwhelm commercial content for the people who are comfortable with the technology–especially the younger generation.”
Universally, industry experts expect the cost per gigabyte to continue to fall and capacity to continue its march onward and upward. Estimates Gartner Research Vice President John Monroe, by the end of 2006 you’ll see 80GB to 160GB 3.5-inch drives sell for less than $50. By 2010, Monroe predicts that you’ll pay that same price for 750GB to 1TB drives. The pace of areal density boosts, he notes, won’t be quite as rapid as they have been in the past decade, but they will continue.
Wickersham outlines what he expects for 3.5-inch drives: “In 2005, for a three-platter drive, 500GB was standard. By 2009, that will be a 2TB drive. And if we continue for 2013, using Heat Assisted Magnetic Recording technology, we’ll have 8TB drives.” Wickersham throws out similar numbers for 1-inch drives: From a standard of 8GB in 2005, he expects we’ll see 30GB in 2009, and 100GB in 2013.
We can also expect to see hybrid hard drives that integrate flash memory to take advantage of Microsoft’s upcoming Windows Vista operating system. In addition, server technologies such as faster rotational speeds and more robust design should trickle down into standard desktop drives–that’s something, says Wickersham, “that’s closer to [happening in] 2007 to 2013. I think it will be sooner than anyone realizes.”
In the near term, one potential technology tweak could be a shift to using what’s called long data block. In long data block, you’ll move from 512 bytes to 4 kilobytes–a change that requires operating system support. This could boost a drive’s capacity and efficiency during video streaming.
Looking ahead, other technologies that will help keep the areal density race on include patterned magnetic media and Heat Assisted Magnetic Recording, also known as HAMR.
Patterned magnetic media is a less random and more structured recording process in which the bits of info are akin to small islands of magnetic material.
In HAMR, the drive will have a heating element, perhaps even a laser, to heat tiny bits of information and change the state of the material, as the data is written; the changed state will allow data to be recorded. “It’s quite an integration challenge, integrating a laser into the disk drive,” says Wickersham.
As these future technologies show, magnetic disk recording has plenty of innovation ahead. “The technology can be re-engineered and reinvented and extended for at least another couple of decades,” Hitachi’s Healy enthuses. Engineering roadmaps extend another two decades, at this point–although the sharpest engineers can’t completely anticipate where storage is going. Back in 1956, after all, a storage scientist’s wildest dreams would not have foreseen the developments we take for granted today.