IBM racetrack memory could replace hard disk drives
Nobody shouted “and down the stretch they come,” but IBM’s potentially revolutionary racetrack memory may have entered the homestretch toward implementation after a recent announcement from the company.
IBM has created a prototype and announced that it has developed a method that allows it to construct racetrack memory using current chip-manufacturing tools and procedures. By simplifying the process of creating the memory, IBM has allowed racetrack memory to move from a “nice idea” to a realistic option for future memory technology.
Racetrack memory had been a project for the IBM development team for several years, before the company announced the technology in 2008. This new type of memory claims to combine the large storage capacities of traditional magnetic hard drives (HDDs) with the speed and durability of flash memory in solid state drives (SSDs).
Racetrack memory consists of nanowires that are extremely small. When the nanowires are etched onto a silicon wafer, they can hold the zeros and ones that make up computer data by changing the magnetic orientation of each section of the nanowire. Each nanowire is created in a “U” shape, and the memory bits move around the U, similar to moving around a racetrack. The U extends vertically from the wafer. The write and read process requires less than one nanosecond. With such small parts and with the vertical configuration, this type of memory should yield greater storage densities than other types of memories currently in use.
With the prototype that was showcased at the International Electronic Devices Meeting in late 2011, IBM is now able to place all of the components and circuits for racetrack memory onto a single piece of silicon. This design change makes building racetrack memory cost-effective. Without this breakthrough in design, racetrack memory would have headed for the long list of technologies that appeared to be a great idea early on, but that never were able to be commercially feasible because of expensive manufacturing requirements.
So, with that potential problem out of the way, how does racetrack memory affect the computing market?
Even with the prototype now created, it still looks to be several years before racetrack memory could seriously challenge HDDs or SSDs.
The problem – if you can call it that – is that HDDs just work so well and at such a low price. Hard disks certainly have some problems with failures, primarily from drops, but they still tend to work well overall with acceptable reliability, especially when the price is considered.
HDDs have set the bar pretty high for competing memories. SSDs don’t have the problems with destroyed data if they’re dropped, and they work faster than HDDs. However, HDDs are so inexpensive comparatively, that SSDs continue to struggle to gain a significant market share in the computing market, despite obvious advantages. SSDs are slowly gaining a foothold, but it’s been a tough market for SSD manufacturers to crack.
The memory market is a bit different from other high-tech markets, where unseating the older, established technology typically is pretty easy for the new technology. In display technology, for example, LCDs overtook CRTs in relatively short order. LCDs had such a large advantage in usefulness and quality that consumers didn’t mind paying the higher price for LCDs. CRTs also had some design limitations that make it tough for them to innovate further to maintain their market lead.
With HDDs, the value remains high, even though they’ve been a must-have component of PCs since the mid-1980s. You receive a lot of storage space for a low price, and the speed is adequate for many users. Manufacturers have been able to continue to add more storage space and more speed to HDDs, helping them remain comfortably in the market lead. SSDs have a lot of advantages, but consumers just can’t yet justify spending that much more money for less storage space, despite the performance advantages. SSDs remain more useful in niche computing and technology products.
Racetrack memory not only has to overcome what HDDs have been able to establish over the past three decades, but it also will have to overcome the growing SSD market. That’s a tall order, no matter how good the technology behind racetrack memory is.
The gap between HDDs and SSDs in terms of value is closing, but it still exists. That gap is significantly wider for racetrack memory and the more established HDD and SSD technologies, so it appears that it will take a long time for racetrack memory to make a serious dent in the market, despite the recent announcement of the prototype. Racetrack memory is going to have to gallop hard to win the Memory Derby.