Racetrack Memory -- Computer Memory That's 100,000 Times Faster Than Today's -- May Arrive in 5-7 Years
Using a nickel-iron nanowire, says scientists, we may soon have computers that boot up instantly, use less energy, and run much, much faster.
Computer memory and storage have always followed the Law of Closet Space. No matter how much you have, you shortly discover that it isn't enough.
So it's good news that scientists in Switzerland are working on racetrack memory, shock-proof memory that they say is 100,000 times faster and consumes less power than current hard disks. Professor Mathias Kläui at EPFL, Laboratory of Nanomagnetism and Spin Dynamics and SwissFEL, says the new kind of memory may soon be possible — and a market-ready device could be available in as little as 5-7 years.
Every time a computer boots up, two to three minutes are lost while information is transferred from the hard disk into RAM (random access memory), explained Kläui in a statement. "The global cost in terms of lost productivity and energy consumption runs into the hundreds of millions of dollars a day," he says in a statement.
Kläui's solution uses a nickel-iron nanowire. Nothing moves mechanically. "The bits of information stored in the wire are simply pushed around inside the tape using a spin polarized current, attaining the breakneck speed of several hundred meters per second in the process," he says. "It’s like reading an entire VHS cassette in less than a second."
The paper is basic physics, Kläui emphasizes, but it entails important consequences for the racetrack memory as developed by IBM. That is, he explained via e-mail, "Given our results, one could expect very fast access times, faster than what has measured so far for these devices, so it would be a very competitive memory with lots of potential."
For each bit of information to be clearly separated from the next so that data can be read reliably, the scientists use domain walls with magnetic vortices to delineate two adjacent bits. Millions or even billions of nanowires would be embedded in a chip, providing enormous capacity on a shock-proof platform. The technical details (which are far beyond my understanding, though maybe not yours) are published in the journal Physical Review Letters, and the results were confirmed by scientists at the IBM Zurich Research Center in a Viewpoint article, The alphabet of spin in nanostructures.
Uh, cool. But what's that mean in English?
If racetrack memory plays out the way scientists expect, we could have a real breakthrough in data storage and retrieval. "Racetrack-equipped computers would boot up instantly, and their information could be accessed 100,000 times more rapidly than with a traditional hard disk," says the statement. "They would also save energy. RAM needs to be powered every millionth of a second, so an idle computer consumes up to 300 mW just maintaining data in RAM. Because Racetrack memory doesn’t have this constraint, energy consumption could be slashed by nearly a factor of 300, to a few mW while the memory is idle. It’s an important consideration: computing and electronics currently consumes 6% of worldwide electricity, and is forecast to increase to 15% by 2025."
Oh! That's good.
A market-ready device could be available in as little as 5-7 years, says Kläui. I asked if those devices will require a whole new hardware architecture. That is, will it mean an entire new computer design, or will it be an advancement in the manner of, say, USB or Firewire ports — something new that a manufacturer can add on to the design of "regular" computers?
"Both," Kläui responded via e-mail. "In a first step, one could envisage attacking the Flash market and replacing the USB sticks with something much faster, unlimited re-writability, etc. In a more radical step one could then replace all memory with racetracks (maybe apart from the cache SRAM) and this would greatly simplify the computer architecture but would mean that computers have to be redesigned."
One could imagine the technology being useful for both small mobile devices (such as smartphones) and large systems (such as servers). But because of the advantage of low power consumption in standby, Kläui said, mobile devices could be a target market at first.