Why IBM's Racetrack Memory Could Bring Terabyte-Level Storage to Mobile Devices

IBM believes its development will provide a quantum boost in read/write performance over conventional memory-and add up to 100 times storage capacity as a side benefit.

Just when one starts believing that processors can't possibly get much smaller and more powerful, storage can't get much more capacious and networks can't move data much faster, somebody always comes up with an IT breakthrough that reboots all that thinking.
IBM Almaden Labs in San Jose, Calif., this week revealed that it has made significant progress in the development of one of its pet initiatives, the Racetrack Memory (RM) Project.
Since it's a radically new type of storage and access memory, IBM believes RM will offer a quantum boost in read/write performance over conventional memory-and add up to 100 times the storage capacity as a side benefit.
If this new product works as well as it looks on paper and in the lab, this huge boost in memory speed and capacity could eventually mean that smartphones, tablet PCs, laptops, cameras and other mobile devices will be able to offer terabyte-scale RAM and Tier 1 data storage.
Because of its efficiencies, RM also requires less power than standard RAM, easing the strain on PC batteries-which are also being improved. iPad batteries, for example, now routinely last 10 to 12 hours between charge-ups.
Ultimately, IBM reports in a new paper published in Science Magazine, this could all add up to a single laptop or tablet PC carrying terabytes of data with weeks' worth of battery life.
The Racetrack Memory's design consists of two magnetic silicon domains connected by so-called "racetrack" nanowires, which handle the transport of data. Because these tiny wires connect the data in the silicon stores directly to the computer's processor, the processor doesn't have to spend time and energy seeking out the data requested; it's already right there in front of the processor in a solid-state storage container.
Major back-and-forth steps are eliminated from the conventional process, enabling much higher data speeds. This would be analogous to the difference between direct-attached storage and a storage area network; in direct-attached storage, there are fewer processes to complete and fewer channels for the data to traverse.
Magnetic domains are not new in the field of data storage, but accessing data from them hasn't been improved much until now, IBM said.
In its scientific paper, IBM claims that its RM process now can control the arrangement of these magnetic domains, so that electrical pulses sent along the nanowires can stop rapidly moving data packages on a dime-as fast as a billionth of a second-and exactly in the nanospot of the domain where the data package is required to be.
"We discovered that domain walls don't hit peak acceleration as soon as the current is turned on, and that it takes them exactly the same time and distance to hit peak acceleration as it does to decelerate and eventually come to a stop," RM team member Stuart Parkin said.
"This was previously undiscovered in part because it was not clear whether the domain walls actually had mass, and how the effects of acceleration and deceleration could exactly compensate one another. Now we know domain walls can be positioned precisely along the racetracks simply by varying the length of the current pulses, even though the walls have mass," Parkin added.
IBM did not disclose when RM might be ready for delivery to markets.

Chris Preimesberger

Chris J. Preimesberger

Chris J. Preimesberger is Editor-in-Chief of eWEEK and responsible for all the publication's coverage. In his 13 years and more than 4,000 articles at eWEEK, he has distinguished himself in reporting...