How Big Is IBM’s Biggest Storage Array in History?

  • Share
  • Read Later

Everyone know what a petabyte is? No, not a petting zoo for ancient computer chips, or copies of the once influential 1980s computer mag. We’re talking about one quadrillion bytes, or 1,000 terabytes. For context, I’m the (proud) owner of a 1TB external USB hard drive, and yes, it wasn’t cheap. Last I checked, the biggest consumer hard drives money can buy today—internal or external—top out around 8TB and cost well over $1,000.

So when IBM says it’s developing a 120 petabyte storage array for an incredibly storage-hungry client, be impressed (very impressed). That’s a record-snapping 120 million gigabytes, after all.

(PHOTOS: 10 Questions for IBM’s Watson)

You know how you’re always hearing about water-cooled CPUs, usually in accord with gamers hoping to crank the clock frequency up so high they could oven-cook a steak in the chassis (or more likely just deep six the chip) without a special cool-down solution? IBM’s new storage array’s going to need that in abundance, because we’re talking an array made up of a whopping 200,000 standard hard drives crunching along in mass-tandem.

What do you store on a 120 petabyte hard drive? How about 2.4 million dual-layer (50GB) Blu-rays discs, or as MIT’s Technology Review reports, about one trillion files.

“This 120 petabyte system is on the lunatic fringe now, but in a few years it may be that all cloud computing systems are like it,” IBM’s Bruce Hillsberg told TR, adding that keeping track of the file system alone will consume two petabytes.

What happens when drives go bad? IBM’s replicating the data several times across the array, as in most multi-disk arrays, but it’s employing a fine-tuned version of the technique that allows the computer to keep going, full out, even if disks within the array fail. When one disk fails, it writes the lost data back to replacement drives slowly (thus not onerously taxing the processors), and only speeds that process up if several more disks are unexpectedly lost in the same timeframe.

Very cool IBM, though I don’t supposed we ought to expect a mobile iDevice-sized version any time soon.

MORE: How Long Until the Cloud Renders Hard Drives Obsolete?

Matt Peckham is a reporter at TIME. Find him on Twitter at @mattpeckham or on Facebook. You can also continue the discussion on TIME’s Facebook page and on Twitter at @TIME.