Country missing? Please select your nearest region...
Published on 15th January 2013 by Peter Hannaford
How many atoms does it take to store a bit of information?
I’m a sucker for Wired magazine. Together with my iPad (and the Sunday Times digital edition), it’s one of the few things that I’m rarely without. I like the way that Wired provides a slightly leftfield view of the people and ideas that could change the way we live, work and play. If you work in the technology industry or data center sector, it’s a must read. (What’s more, and I don’t get paid for saying this, you can subscribe to Wired for a very reasonable £28 per year in the UK).
A story caught my eye in a recent edition which got me thinking about innovation in our sector. “IBM Brains Turn 12 Atoms Into World’s Smallest Storage Bit” concerns an IBM researcher called Andreas Heinrich who has found a way to store a single bit of information using just 12 atoms. Of-course (and I know you’d expect this), he’s had to use an anti-ferromagnetic structure to do so, as opposed to the ferromagnetic structures which are the norm for hard drives. For comparison, hard disk manufacturer, Hitachi, reckon their storage devices require around 800, 000 atoms to achieve the same end for storing one bit.
According to Wired, it’s unlikely that Heinrich will solve all the problems of uncertainty associated with particle physics in the immediate future. So it’s also pretty unlikely that we’ll see this technology in next year’s USB memory sticks. Atom scale storage apparently operates at 1 degree Kelvin (which is cold even for an industry that likes to keep their rooms cold). To date no-one, Heinrich included, has the simplest clue how something so small can be built outside a laboratory, let alone how it can be done cheaply.
Nonetheless, the principle has been demonstrated, and that gets a lot of points in my book. If it can be made to work economically, we can only start to imagine the impact of atomic storage on power, cooling and space in current-state facilities. It could be a real game changer. But what really makes me wonder is whether there is anyone sitting in a lab somewhere (or even a garden shed) trying to change the game in the data center?
It feels like years since we last saw something radical – perhaps it was when APC took cooling units off the data center wall and moved them into the racks. Even then, you could argue that this was more a matter of arranging things better than actual innovation. As for other innovations we’ve seen in recent years (free cooling, economizers, eco-mode UPS and the like), well I’m sorry to say that a lot of this stuff has simply been around for years. Sure, we’ve seen advances; things have got smaller, more efficient (cheaper?). But where are the real breakthroughs?
Obviously I apologise if you’ve been working on something transformational like a cellulose microprocessor that can be grown on a tree. But if you think I’m being harsh, I’d ask you to cast your mind back and let me know your contenders for the best five data center innovations of the last decade. I promise I’ll compile them into a list and publish them on this blog.