The road to Hell, they say, is paved with good intentions. Google has updated its OS for Internet of Things (IoT) – now called Android Things – with the release of a developer preview, but I fear that if the OS is released to open source, the same chaos and insecurity that hampers Android will prevent this from becoming successful.
Releasing software to open source in the fixed world works well and all benefit from it, but in mobile devices it has caused nothing but problems for the last 15 years.
Android Things is an update of Brillo (launched at Google i/o in 2015) which to date has seen very little traction. What has been more successful is Weave which is a communications layer that enables all of these devices to talk to each other as well as interact and integrate with Google services such as Google Assistant. This has fared considerably better and is currently being implemented by Phillips, Samsung, Belkin, TP-Link, Honeywell, Wink and a number of others.
Using Android on IoT devices is fraught with problems:
- First: Most IoT devices today are not required to do much other than turn things on or off or relay the data from the sensors to another unit. Consequently, using a smart operating system such as Android even when it has been stripped down appears to be overkill. This is because a completely proprietary real time operating system (RTOS) will be easier and cheaper to deploy and will probably result in much longer battery life. It also has the benefit of giving its owners complete control which is something that they wont have using Android Things.
- Second: If the software is open source then those that use it are likely to pick and choose the elements and the APIs that they need for their device and drop the rest. This means that every device will be running a slightly different version of Android Things making updates, security and software management almost impossible. It will also make it much more difficult to include devices as part of a wider ecosystem as each device will need to be assessed to see what it has and whether it will work with other devices and services.
This is why I think that it is important that Android Things is like Android Wear and Android Auto, which are not open source and remain tightly controlled by Google. That way the software will be much more easily managed with timely updates and consistent APIs.
Despite this, I think that the piece of this puzzle that really matters to Alphabet is Weave. Weave connects all of the devices together as well as connects them to Google services such as Google Assistant.
Furthermore, Weave sits mostly on a server which is fully under Google’s control and which it can update at anytime. It is this piece that allows all of the devices to be integrated together (like HomeKit) and controlled from one place such as Google Assistant.
This is critical, as it is this piece that will pass all of the data back to Google to help it improve its AI as well as monetise the usage in the normal way.
Consequently, I suspect that the best option for IoT device makers will be to do their own thing on the device but then ensure that the device can integrate with Weave such that they can benefit from being part of the wider ecosystem.
HomeKit and Weave are just two of a myriad of solutions that are available for IoT devices which in itself is a big problem. This is because it is very difficult to decide which one to support and as a small company this could easily be an existential choice.
I think that these problems will keep IoT as a theme with a lot of promise but very little substance in 2017. The two sub-segments of smart home and e-health are likely to emerge first, but it is going to take far longer than the press releases would have us believe.
Alphabet remains on my indifferent list with most of the good news and none of the bad already priced into the shares. I prefer Tencent, Baidu and Microsoft.
This article was first published on RadioFreeMobile.