Facebook’s Open Hardware Project – Does it compute?

“Open Source Hardware” is not a new concept, but it’s certainly a poorer cousin to the Open Source Software movement in terms of the sheer volume of commentary and general awareness.

Despite this, we have noticed an increasing buzz about this licensing approach in the past 12 months or so and, with the recent announcement that Facebook is ‘open sourcing’ the specifications to its new data centre, we thought it was time to do our bit to increase awareness of this alternative licensing model.

Like open source software, the principle focus of the open source hardware model is to grant others the freedom to use, modify and (hopefully) improve hardware designs with the ultimate aim of creating a collaborative and sustainable community of developers.

Facebook’s “Open Compute” project certainly has some lofty aims and it seems to have delivered some excellent initial results with figures estimating that their new servers, which they describe as “vanity free” (one can’t help wondering how many Facebook profile pages could be described in similar terms), are 38% more efficient and 24% less expensive to build & run than other state-of-the-art data centres.

As with many open source projects, the natural reaction is to question why all this is being given away for free.  In Facebook’s case, there would seem to be a number of good reasons:

1. Facebook is not in the market of manufacturing and supplying data centres.  Without an obvious route to commercially exploit these new designs, there would seem to be limited commercial advantage in keeping them proprietary.  It will be interesting to see if the big hardware manufacturers will be as keen to release new designs under a similar model!
2.  The design of hardware is inherently more ‘open’ than software.  This is simply a question of ease of access.  If I want to see what hardware components make up my desktop computer, they only skill I need is the ability to operate a Phillips screwdriver.  Indeed, more often than not, a gloriously detailed descriptions of a ‘teardown’ can be found online (perfect for anyone who might, for example, want to know how many magnets are inside a new iPad2 Smart cover, but can’t bring themselves to break out the stanley knife and find out).  By comparison, reverse engineering of software requires considerably more expertise, not to mention that such act is likely to put the user in breach of any associated software licence.
3. Higher volumes lead to lower costs.  A number of the components used by Facebook need to be custom made – it is possible that encouraging more people to use this design will lead to an increase in availability and a decrease in the cost of such components;
4. It’s good publicity.  It may be a cynical view, but this is certainly great publicity for a company that may be seeking to counter-balance its increasingly ‘corporate’ image by reaching out to a large community of developers and IT enthusiasts who believe that the web should be and remain ‘free’.

There are, of course, a number of drawbacks inherent to Open Source Hardware projects compared to Open Source Software development, not least that the barriers to join the community are normally higher (testing a new hardware design will usually require components and/or tooling equipment to be purchased at some point) and from a licensing perspective, there is not yet a set of ‘community approved’ standard licences to draw upon that are comparable to those approved by the Open Source Initiative.  We watch with interest whether the Open Compute Project will be able to overcome these issues and create an engaged and thriving community of developers.

Oh… and if you’re still wondering about that teardown – 21 magnets in total (6 for the ‘hinges’, 14 to keep the cover closed / in the stand position and 1 to activate the ‘sleep’ sensor).