In his book The Gorilla Game, management consultant Geoffrey Moore identifies a tendency for some technology sectors to develop winner-take-all outcomes. The Internet of Things (IoT) market looks like a textbook example, as many industrial customers (original equipment manufacturers) are looking to consolidate around a single platform, a slice of enabling technology. It is surprising, therefore, that no IoT gorilla has yet emerged from the mist.
+ More on Network World: Most powerful Internet of Things companies +
IoT is an enabling technology that many OEMs will rely on, but they will likely not want to develop themselves. It’s complex: The components of an IoT platform span several disciplines requiring close integration. And not only are there economies of scale, but there’s a network effect: Customers will be drawn to the most popular platforms, making them even more successful. All of these factors favor the emergence of a gorilla IoT platform.
Components of an IoT platform
An IoT platform consists of sensor-control modules in “things” connected via the internet to a cloud service that collects their sensor data and transmits control signals. The cloud service then provides interfaces to mobile apps for consumer products, or it provides analytics engines and management dashboards for enterprises.
For example, consumers wishing to unlock their front door will use a mobile app connecting to the IoT cloud service, which in turn queries the sensor-control module in the door lock. There is no direct connection to the lock; all messages transit the IoT cloud service. It is inevitable that many enterprise IoT systems will follow the same architecture: We will become accustomed to over-the-top cloud services in what have, up until now, been inside-the-firewall enterprise networks.
Thus an IoT platform includes a number of components.
First, the sensor-control module, using one or more of a menu of wireless standards (Wi-Fi, Bluetooth, Z-wave, ZigBee, cellular data). It should be easily incorporated into appliances and devices: A complete solution will include both off-the-shelf modules and hardware-software reference designs, as a refrigerator design engineer doesn’t want to learn IoT but would rather drop in a module and link the control and sense lines.
Second, the cloud platform must be scalable and distributed to support millions of sensor modules. Development and maintenance of this software is a considerable effort. Security must—of course—be designed in, and customers’ traffic must be segregated. APIs from the cloud allow analytics engines to pull and crunch data, and a simple analytics dashboard should be built in.
Third, the complete platform offering includes templated management functions for “standard” IoT modules and development kits for mobile apps, as IoT customers may not have expertise in this area.
Why no IoT platform standards?
One could ask, very sensibly, why the industry cannot establish standards that allow IoT platforms to be dismembered into their components with recognized interfaces between them. If this were the case, sensor-control modules could be built separately from cloud services, and analytics engines or mobile apps could be easily connected to multiple services.
There are two answers:
First, so many layers of protocols are required that even standards-of-standards bodies cannot help the industry agree on a limited combination. Today, systems integrators’ omnibus IoT architecture charts attempt to list all combinations, and they show so many protocols that no one can understand them.
Second, many of the large players see the potential to become a gorilla in this market and have developed proprietary twists and subjugated standards forums to protect their ecosystems.
The complexity of the system, and the tight integration between components, coupled with the lack of generally adopted standards are driving IoT vendors towards comprehensive, wide-ranging, semi-proprietary platforms. These platforms will not easily interoperate: A sensor module or mobile app built for one platform will not work on another. But they offer the convenience of a ready-made system with properly defined interfaces. OEMs and sensor application specialists with limited resources will be drawn to these established platforms rather than building a complete system from scratch.
No real path to IoT domination
The platform specialists are making progress. But the path to domination of IoT by a small number of winning platforms is not inevitable. While the architecture described above—module to cloud service to analytics-control-management APIs and apps—is universal, there are fault lines through wireless technologies, data patterns and equipment segments.
For instance, some of the most successful IoT platforms to date connect cars over cellular data networks, while others have focused on home security over Wi-Fi and Bluetooth. IoT covers such broad markets that each segment may develop specialist leaders.
Time will tell whether IoT platform providers grow into a large, fragmented group of chimps or a smaller group with a leader for each market segment—or whether a single gorilla platform emerges. When many powerful players see the same opportunity, several years before the market takes off, they can compete with one another to a stalemate. It is most likely that, with standards-of-standards work in gridlock, the IoT market will spawn a broad range of semi-proprietary platforms with a leader for each wireless technology and end-product segment, but no dominant gorilla.
This article is published as part of the IDG Contributor Network. Want to Join?