Facebook is building its own storage hardware and aims to publish details so other companies can make their datacentres more efficient.
The designs should become available in May via Facebook-spinoff the Open Compute Project, the company confirmed to ZDNet UK on Friday. The move will come a year after it started publishing the design specifications of its own ultra-efficient servers.
"We're taking the same approach we took with servers — eliminate anything that's not directly adding value," Frank Frankovsky, Facebook's director of technical operations, told Wired. "The really valuable part of storage is the disk drive itself and the software that controls how the data gets distributed to and recovered from those drives. We want to eliminate any ancillary components around the drive — and make it more serviceable."
Frankovsky gave few details, but did say the company was looking at ways to simplify the way drives are held inside storage gear so they are easier to swap in and out as they fail or are replaced.
But further clues can be found in the research areas of some of the company's 2012-2013 Facebook Fellows — PhD students that it funds for a year.
Rashmi Korlakai Vinayak was given a fellowship to fund her work on "new encoding mechanisms for distributed storage systems with a goal of significantly improving their reliability and elasticity."
Another fellow, Tyler Harter, will focus on modelling the Hadoop File System to try and make it more efficient.