With the near-duopoly of Android and iOS well-established in the handset world, it is easy to forget the viciousness of the mobile operating system wars in the early days of smartphones. The traditional cellphone vendors huddled together for warmth around the Symbian platform, but the politics of that effort, and Nokia’s dominance, destroyed it. BlackBerry, Palm and of course Microsoft battled for air.
It was not inevitable that the two new entrants would win out, but it was inevitable that the mobile world would only support one or two platforms at massive scale. The lower the costs of the devices and services, the less the sector could survive fragmentation, so as in the IT and PC worlds before it, the mobile industry made its choices. Now the nascent internet of things (IoT) sector will be gearing up to make similar decisions and there are both old and new candidates to become the Android of ‘connected everything’.
One of the candidates is, of course, Android itself, or at least a Google-driven open source operating system. While Apple may well keep iOS focused on the smartphones and gateways that control the IoT, rather than the ‘things’ themselves, others want to embed their software, and build their developer ecosystems, at every level of the IoT stack. Google, ARM and Huawei are three examples – each targeting somewhat different functions and products, and not necessarily mutually exclusive; but each aiming to be the controlling genius of the IoT.
There will be no simple anointing of one or two winners in the IoT. Unlike the homogeneous smartphone market, it will span an array of devices and sensors, of vertical markets, and of applications and services. It is logical to assume that there will be high level software platforms and developer toolkits targeting specific vertical or horizontal segments (Apple CarPlay, Android Wear and so on). There will also be new OSs from the smartphone world, which will look to take advantage of the fluid state of the IoT to gain a bigger presence in new device categories such as wearables and cars (Sailfish Jolla, Ubuntu, Firefox Mobile and even Google Chrome).
However, beneath all that will be the software needed to control and secure millions of tiny low power devices, which cannot support a full OS but still need to be integrated into a common platform – with developer APIs, security and connectivity – if they are to deliver their data to the apps and the cloud, and if the economics of connecting every device on the planet are to work out.
This low level, grungy layer of operating systems is where the latest battle is being planned. Intel is sure to play a role, given its push into all kinds of chips for connected devices, and its highly strategic ownership of Wind River, which has expertise in real time operating systems (RTOS) and virtualization, among other assets.
The most prominent so far is ARM’s mBed OS, which provides core software functionality for chips based on its Cortex-M microcontroller family and supports partners using its mBed developer platform. Joining that race was Huawei’s LiteOS (Wireless Watch May 21 2015), which the company says is the lightest OS in the world, requiring just 10Kbytes. It is the cornerstone of Huawei’s Agile IoT, which also includes the gateway and controller reference platform, aiming to simplify the task of the product developer with a complete base solution, and so accelerate the creation of a large number of devices and applications.
Many of the stripped-down operating systems are designed to fill the gap between traditional RTOS offerings, for embedded objects that did not require connectivity, and the ‘fatter’ software which has generally driven devices with internet links. Now those devices are not just smartphones and cars, but street lamps and meters, a far thinner OS – but with full developer functionality – is essential.
So it is unsurprising that Google would aim to get in on the act, pushing down to a layer of control deeper than the visible branded Android and Chrome platforms – a layer where the provider of the OS may be invisible to the outside world, but nevertheless sits like a spider at the center of the developer and device web. It is reported to be planning to launch an IoT operating system at its I/O developer conference this week. This would be another piece in a growing number of elements in Google’s bid to shape the IoT, which also include Android Wear and other Android extensions; and the Thread WPAN connectivity protocol driven by its Nest subsidiary, and potentially a challenger to Bluetooth and ZigBee.
Reports surfaced last week, ahead of I/O, that Google would unveil ‘Brillo’, an OS for connected devices, and, according to The Information, would “release the software under the Android brand, as the group developing the software is linked to the company’s Android unit”. That could indicate that Brillo, while targeting lower level functionality and far more basic devices than Android, would still offer accessibility to existing Android developers and a degree of commonality across connected ‘things’ from cars to sensors.
The reports say that Google plans to offer Brillo to OEMs for free, in an attempt to unify an increasingly fragmented ecosystem and stimulate uptake, though it will be interesting to see whether there is an equivalent of the Open Handset Alliance to try to keep the platform harmonized. The open source community would clearly like to see greater openness than Android, in reality, delivers – while the smartphone OS is free to developers, Google requires adopters who want to include one of its applications on their devices to bundle them all. Such restrictions could limit adoption of Brillo, and so this may be the chance for Google to adopt a new approach – which may be justified because it will monetize IoT gadgets primarily via the data they supply for its cloud services, rather than via direct on-device services.
Future licensing agreements aside, Brillo is apparently aiming to cater for devices with 32-64Mbytes of RAM, which is a significant step down from smartphones, but an order of magnitude more powerful than microcontroller-powered units that will be home to Huawei’s LiteOS or mBed OS.
However, memory and other resource restrictions mean it is not possible to port Android apps directly to these more constrained devices, which is why Brillo is needed – and of course, it may pave the way to an even lower-spec platform to compete directly with LiteOS and the others which will no doubt start to emerge from various parts of the software ecosystem. For now, though, Brillo will be targeting more fully-featured connected ‘things’ (likely those with screens), rather than items like switches and actuators that would be better handled by a microncontroller.
The most obvious tie-in for Brillo will be with the Nest-led smart home platform – the company’s smart thermostat, smoke alarm and Dropcam security camera would be prime candidates for Brillo. With the acquisition of Revolv, Nest bought into a fully-fledged smart home hub, which would likely run a more fully-featured version of Android than Brillo – but Google/Nest will also be able to leverage its leadership of Thread, whose low-power, 802.15.4/LoWPAN/IPv6 mesh networking protocol is due to appear in June or July.
For now, Brillo and Huawei’s LiteOS appear to be fulfilling different functions, but the Chinese company’s operating system decisions in the coming year or two will have a significant impact on Google’s ability to achieve its ambitions. On the one hand, Huawei uses Android as its primary mobile OS and is talking about wanting a close alliance with Google in the IoT too. On the other, it is reportedly readying its own mobile OS, which would reduce its reliance on Android and enable it to integrate tightly with popular services in its home market – the Baidu search engine or Alibaba ecommerce apps, for instance – to compete more effectively with Xiaomi and others.
That would fit with the rising determination of Chinese mobile giants to lead the development of patents, standards and platforms for the IoT and 5G, to increase their own international success and to support government calls to make China technologically self-sufficient. In that context, while any new smartphone OS is unlikely to take significant share from Android (even in China, local attempts like Baidu OS and Alibaba’s Aliyun have made little headway against de-Googled implementations of Android), it is hard to imagine that Huawei will not want its own software platform for the emerging IoT.
In an interview with EETimes, Shao Yang, president of Huawei’s strategy marketing department in the consumer business group, insisted Google was more friend than rival, saying: “Google is a good place to start” when moving into the IoT. Yang outlined phase two of the company’s mobile strategy, which involves connecting all kinds of devices, not just handsets and tablets.
To accelerate progress and build well targeted, implementable platforms, the firm intends to create several horizontal or vertical business units, including smart home, smart cars, wearables and other. In each one, it will create a ‘round table’ consisting of vendors and developers with expertise in that segment.
These groups will have limited membership but will focus on pooling their technology contributions in order to accelerate the process of “getting everything connected” and building big data pipes. The focus will be on the implementation level rather than on designing new standards. Each round table will be structured in multiple levels, including R&D and C-level, and once one group makes a decision, it will share and implement its ideas.
Yang was clear he would like Google to participate in these exclusive round tables and claimed Huawei would not necessarily take the lead role in each group. “If people asked Huawei to lead at the round table, of course, we’d be glad to serve. But we can be one of the ‘knights’ at the table,” he said in the interview.
Huawei knows it needs specific expertise to master the particular demands of industries which have, until now, had few connections to one another – from the auto sector to white goods to smart factories. Its own power will come from joining the dots, providing core technologies to “connect the round tables” as Yang puts it, and supplying common underpinnings such as the OS. And this is why Google may be welcomed as a partner in an area like smart home, but will surely not be tapped for the underlying OS, which must be Huawei’s preserve.
Apple may be doing U-turn on HomeKit:
Apple seems to have backtracked on its HomeKit stance. Several sources are reporting that the new version of its mobile operating system will have a dedicated Home application – which contradicts the previous narrative that HomeKit would simply be part of the software framework within iOS.
So if iOS 9 does indeed feature a Home app, which would presumably govern the connected devices that are badged with the smart home equivalent of the Made for iPhone (MFi) certificate that guarantees compatibility, the next question to ask will be whether Apple will make the new app a feature of its upcoming Worldwide Developers Conference (WWDC) in June.
That date coincides with Apple’s announcement that the first HomeKit devices will be available in June, and conveniently matches with Apple’s tradition of announcing the latest version of iOS at WWDC and then releasing it to the public within a couple of weeks. Consequently, it seems likely that Home will be announced alongside iOS 9 at WWDC, with the news that Apple Stores will soon be stocked with the first fully HomeKit-compatible smart home devices.
The anonymous source that instigated the rumors says that Home can be found embedded inside the current iOS 9 developer builds. The source adds that the app will offer “fairly basic” functionality, including wireless device discovery and activation, creating virtual representations of rooms and the devices inside them, and using the Apple TV as a hub.
The Apple TV inclusion has seemed apparent for some time, although Apple refuses to comment on its role in Apple’s smart home ambitions. Currently, it appears that a controller devices will need to be inside the home network in order to allow iOS to control the devices.
When the iPhone or iPad is at home, the mobile devices will be able to do this, but if a user takes that device with them away from home, they will have to use the iOS unit to communicate with an Apple TV inside the home – which can then convey the commands issued from the mobile device to the connected devices within the home, such as lights and garage door openers. Consequently, this means that Apple’s radio protocol strategy is entirely focused around WiFi and Bluetooth, although WWDC might throw up a new medium such as ZigBee, Z-Wave or even Thread.