Recovering data from shrinking disk drives
Digital devices have come a long way in the last few decades, particularly as they continue to shrink in size. Many advances, including new recording media, disk drive heads, and disk architectures, have contributed to making today’s computer drives compact while being able to store and read amazing amounts of data. But as the size of disk drives became relentlessly smaller and were able to store tremendously more data, a major challenge developed.
How can the computer successfully recover and read the bits of data stored on miniature disk drives?
Over the last 30 years, Carnegie Mellon engineers have contributed in many different ways to the progress of disk drive storage and, in particular, significantly contributed to solving the challenge of accurately reading bits of data crammed into miniscule places.
Back in the 1950s, a typical computer disk drive was approximately five feet in diameter and could store no more than 1 million bits, or a megabyte (MB), of data. For reference, an MP3 audio file a few minutes in length, or a 10 million-pixel image from a digital camera, typically take up several tens of megabytes. Fast forward to present day where disk drives are a few inches in diameter and can store terabytes (TB) of data (1 TB = 1 million MB).
As disk drives decreased in size and the amount of data being stored increased exponentially, researchers became concerned about the ability to accurately recover and successfully read the stored bits of data. Carnegie Mellon’s Data Storage Systems Center (DSSC) is a world-leading academic research institution in data storage that focuses on magnetic data storage technology for hard disk drive applications.
In the early 1990s, José Moura, a professor in Carnegie Mellon’s College of Engineering, along with his then Ph.D. student Aleksandar Kavcic, now a professor at the University of Hawaii, set out to find an innovative way to accurately recover bits from the ever-shrinking storage disk drives of the future.
Kavcic and Moura invented and patented a detector that could safely and accurately extract recorded data from disk drives. When the early 2000s recording technology changed to perpendicular recording, their detector algorithm invention became a must-have technology.
“Alek Kavcic and I chose to target the limitations that were sure to arise in reading bits in magnetic recording in the future,” says Moura. “Instead of looking for quick payoffs from incremental improvements, we invested our efforts in understanding and abstracting the fundamentals. In a moment of serendipity, we were able to develop a simple, fundamental new way to account for the main physical limitations of magnetic recording and to invent a detector that outperformed all others.”
It is estimated that the disk drives in 60% of computers (over 3 billion) made in the last 14 years contain this detector technology enabling users to recover saved data.
“Our patience paid off in the 2000s when
Looking to the future, Moura is developing algorithms that digest the tremendous amount of data being collected from multiple sources in everyday life. Nowadays, cities are covered in sensors that monitor security, weather, traffic patterns, energy consumption, pollution levels, and more. The data from these sensors allow us to understand the normal social behavior in a particular city. By addressing these data-rich environments, Moura hopes to understand the normal—and abnormal—behaviors of cities, from traffic jams to inconspicuous sources of pollution to unbearable sources of noise. Ultimately, Moura’s data analytics may help urban planners rethink the way they envision cities.
“If urban planners better understand city bottlenecks, they could reimagine a city and have major social and environmental impact,” says Moura. “It would give city dwellers a more pleasant day-to-day life. I think we live in interesting times. For the past few decades, society has shaped technology. Moving forward, technology is helping to shape society. And that is pretty amazing.”