With so much buzz around the Internet of Things, many are wondering what will be the platform on which the IoT ecosystem will be built. Stijn Schuermans reviews several interesting developments in IoT hardware modules that make getting ‘Things’ done a breeze, and adds some question marks with the widespread focus on devices.
Five years ago, at the onset of the smartphone revolution, no-one predicted a significant latent demand for apps such as flashlights or “Draw Something”. And who would have thought that a game like Angry Birds could create billions of dollars in value; not just for the publisher, Rovio, but also for Apple as the platform owner?
So it is in the Internet of Things (IoT). We don’t know which IoT applications will become successful; indeed, it is fundamentally unpredictable. As we argued in our June report “The M2M Ecosystem Recipe”, trying to deterministically control the innovation process is at best ineffective, at worst futile. (To anyone who’s attempting to map the IoT ecosystem and predict which market segments will be winners: good luck!)
Luckily, the app economy clearly showed that when you enable people to experiment, they will find ways to create value, often in unexpected places. Paraphrasing Amazon’s Jeff Bezos: empowering others to pursue their dreams and to boldly experiment creates a diversity of improbable, but often successful ideas. Gatekeepers that insist “that will never work”, even when they are well-meaning experts in their field, ultimately slow down innovation. Developers, in other words, are a crucial ingredient of any IoT ecosystem
Where will IoT developers come from?
Before an Internet of Things platform can reap the benefits of a large developer community, however, it first needs to convince them that the platform is worth their time and effort. After all, every platform is ultimately competing for the attention and mindshare of developers. As it happens, lots of exciting technology is being built that enables developers to experiment faster
New hardware and operating systems are allowing developers to dramatically cut the costs of bringing IoT products to market. These modules are low-cost, easy to master and allow for quick iterations in the market, even if they are not entirely optimized for large-scale production. This is exactly what’s needed for developers to uncover those killer apps. The first priority is figuring out which products people value; the hardware can always be redesigned and optimized once an application becomes successful.
Let’s explore the prowess of five technologies to reduce barriers for developers and, equally if not more important, appeal to their sense of cool
- The status quo of hardware modules doesn’t look too good in terms of low barriers and coolness. Development modules offered by mobile operators (see e.g. here) or by chipset vendors (mbed, WICED) might be affordable for engineering companies, but certainly aren’t for hobbyists and tinkerers. Mostly, they seem to assume that their users are hardware engineers solving connectivity problems. With all respect for the people solving those difficult technical problems, this is not what’s blocking wide adoption of connected things
- A bit easier to digest are those modules that appeal to computer programmers. Windows Embedded and variants of Linux have long been established in the world of connected things, e.g. in network security cameras. They are now joined by more accessible components like the Raspberry Pi (a $35 computer originally intended to help kids learn how to program) or the BeagleBone. Raspberry Pi’s and the likes can be great to connect appliances, but they are often bulky and power-hungry, making them a no-no for battery-operated devices
- The king of low-barrier microcontrollers is without a doubt the Arduino platform (although there are other players in this space as well, like the Electric Imp or the Spark). It is affordable, real-time, easy enough for hobbyists, powerful enough to power satellites, extendible with so-called shields and it exists in literally hundreds of sizes and shapes, due to its open license
- If familiarity with the development environment can lower barriers for adoption, then what about Android? Can the largest smartphone platform leverage its app developer base to gain the IoT market? While Android certainly solves a lot of the hardware and connectivity headaches that a stock Linux might not, its main power lies in user interface technologies. These are much less relevant in IoT devices and might make the operating system too bloated for many devices. However, its open license might make it appealing not to application developers, but to hardware makers. Android’s popularity means that a lot of hardware components already have Android drivers available for them. This makes getting Android up and running on some new bit of hardware relatively cheap and easy. Let’s not write it off just yet, in particular for more complex IoT devices.
Closing thoughts – maybe the device doesn’t matter?
Those who want to enter the IoT market, software developers and electronics enthusiasts alike, certainly have a lot more choice of technologies than they did even a few years ago. Cheaper and simpler early-stage devices are surely crucial to enable the “perfect storm” of the Internet of Things.
To play with IoT applications, you might not even be a developer anymore. The flurry of user-level IoT platforms is testimony to this: Thinking Things, Makey Makey, ATOMS, Ubi and Twine are just some of the non-techie IoT building blocks that come to mind. Is this a sixth entry-level way to attract IoT solution makers?
In the end, IoT devices differ from smartphones in a crucial aspect. They are not the main interface to the human that uses them. Much more than the hardware technology, the value in the Internet of Things will be created by combining and presenting the data that these devices generate. “Connecting things” will be about the connections, not about the things
- Stijn (@stijnschuermans)