Google Motorola’s announcement of Project Ara – their modular phone project, fresh off the back of a series of major announcements from mobile manufacturers (Nokia, Apple and Microsoft) signals their intent to make Motorola a byword for innovation again, but can they deliver on the promise of a truly modular mobile device?
To start, it’s worth pointing out that although Project Ara, one of the more truly surprising pieces of tech news this holiday season may be taking social media trending by storm, the concept of modular mobile devices are nothing new.
Phones with modular screens have been around for a while, and modular cameras are becoming increasingly widespread, especially from manufacturers with their own optics division to leverage their mobile side with.
Dave Hakkens, a designer by trade, first mooted the possibility of a completely modular device a month or so ago. Hakkens’ has since stated he’s partnering up with Motorola though questions remain about the timing of last month’s video and the extent of his involvement prior, was it all a big marketing drive by Google?
Either way, the fundamentals of what Motorola are planning with Project Ara is now established. The phone itself will be an endoskeleton, supported by upgradeable “modules” that can be swapped in and out as the user sees fit and depending on what they need. Even more interestingly the hardware itself will be an open platform, with the development of modules open to hardware manufacturers of all kinds through an MDK.
The extent to what you can actually “swap out” in a system-on-chip age is up for debate, but it looks like Project Ara is moving away from some of the more extravagant claims about swappable CPUs and RAM that initially accompanied its announcement and towards more modest claims about daisy-chaining the removable camera, battery and speaker modules. This is significant because a lot of the initial scepticism that went hand in hand with the buzz around Phonebloks were questions about the ability to upgrade more fundamental hardware components. It’s also significant that Phonebloks itself will continue as an independent venture.
The promise of Ara and the inherent limitations of mobile devices
Google obviously intend to keep the traction generated by the initial buzz going with the announcement that the first Ara MDKs (Modular Development Kits) will be sent out “sometime in the Winter”. This is understandable given Motorola’s increasingly flagging fortunes over the past few years, Google evidently want to “do something” with it as one of their weaker subsidiaries and the modular phone concept, whilst certainly inviting of a fair degree of initial scepticism, has gotten their brand name back into the headlines in a positive way. This is certainly the most exciting thing we’ve seen from Motorola in years.
It’s when we start discussing the actual hardware behind Ara, that significant questions start rearing their head. As discussed earlier, Motorola seem to be avoiding making commitments about swapping out CPU and GPU parts, for the simple reason that most modern smartphones are SoC, that is, a system on a chip, with everything from the wifi, CPU, GPU and other components all soldered onto a single chip.
Assuming that Motorola adopt some kind of SoC-based design, the question is what exactly will the consumer be able to swap? If there’s no ability to swap out the most system critical parts of the hardware then it’s debatable what actual value Ara will hold.
To follow on from this, pursuing a modular design carries a number of inherent negatives. There’s the lack of optimization that most hardware heterogeneity precludes and the evidently bulkier size that a modular device will entail when compared to purpose-built smartphones with one SoC. More problematically than all this is that any sort of swappable SoC module would have to have more pin connectors than is currently feasible for a smartphone sized device to possess. The modular design of your average PC is a good illustration of this. It requires high speed differential signalling connectors between components that force a limit on how small high-end PCs can actually become. This means that either Motorola are developing a small form-factor connector that is previously unheard of, or, again, the modularity of the device will be fairly limited.
The final question is whether there is a mass market for this sort of device to begin with. Smartphones already possess arguably more processing power than any mobile apps really need and achieve this with pre-built OEM designs. On top of this, the pricing structure for smartphones themselves is incestuously linked to the contracts offered by network operators. This is a structure most consumers are fairly happy with as it affords them high-end phones that they otherwise wouldn’t be prepared to pay out of pocket for and regular upgrades that may render the idea of an “upgradeable phone” as fairly pointless amongst consumers as a whole.
Uniform hardware architecture and standardized firmware: Ultimately more important than modularity
On a broader and more long-term level it’s ultimately unimportant if Motorola fail to take any significant market share from this endeavour. Google were always going to absorb Motorola losses simply for the luxury of having their own player in the consumer electronics industry and the patent portfolio that comes with it. Moreover, the marketing behind this announcement has successfully associated Motorola with the word “innovation” for perhaps the first time in near a decade and Google are quite happy to employ them in this way as a sort of test-bed for new ideas.
The most important thing about this announcement isn’t the modularity, but the possibilities it signals about standardized mobile hardware, this is something that has gone largely unsaid but is worth explaining.
You may ask: “But PCs are modular and simple to put together, why can’t smartphones adopt the same architecture?” Good question. PCs are modular because they operate across a certain uniformity. This (general) consistency of firmware and OS is what allows you to easily plug GPUs, CPUs, NICs in/out.
By contrast, the current mobile environment does not support this. This is because, in the PC example, the aforementioned firmware builds a device tree of hardware devices attached to a computer, which is then passed on to the OS (say, Windows) to query and build a generic or specific driver list for.
ARM and mobile OSes see no need for this. Crucially, mobile firmware is compiled different for each device. A good illustration of what this entails is how a particular manufacturer (for example, Samsung) may ascribe different pins for the input and output of a particular component. When this fragmentation occurs, a phone’s firmware has to be bespoke.
Moving towards firmware standards means standardizing mobile devices alongside them. This ultimately means that you we could be reaching a point where you aren’t tied down by a particular phone’s firmware. You could buy a phone from a network provider, format it and install a completely different OS. If Ara is any indication, we could be heading in this direction sooner than we think.