The ad hoc Internet of things is well under way
The third IoT segment is the least controlled. That's both good and bad.
Think of your home. If you're a techie, you probably have several IoT devices: an Apple TV or Google Chromecast, a Nest or Honeywell Lync thermostat, an Internet-connected Liftmaster garage door opener, your car's Bluetooth ignition lock, and so on. Some interact, some don't. Some should, some shouldn't. When they interact, and when they should, is often a personal choice for the user's context.
It's ad hoc and, thus, messy. If there's an integration point it's usually a smartphone or tablet, running apps for each service and collecting alerts into email or messaging apps. The user is the integrator.
What's changing here is the growing collection of ecosystems. Apple is the furthest along in this effort, with its three ecosystem integration APIs debuting this fall in iOS 8 and OS X Yosemite: HomeKit for home automation devices, CloudKit for cloud storage and sync, and HealthKit for fitness and medical devices. On iPhones, the new Health app in iOS 8 acts as a central repository, managed by the user, of health information provided by the compatible devices and cloud services that users choose. It can also be a conduit to other systems, whether an electronic health records (EHR) system or a weight-loss service.
This integration is still ad hoc, but it's organized by a specific ecosystem. It's not so much a closed system as it is a compatible system, sort of like Windows was in the PC era for software and hardware. This native compatibility allows easier interoperability, which lets people create a custom IoT. That's powerful.
It's also why so many providers want to be the ecosystem of choice. Apple is the furthest along, but Google is also pushing hard on several fronts including its Nest purchase, its Chromecast effort, and its active participation in the Thread effort. Samsung has talked about similar efforts, but its execution has been, to be polite, uneven.
The carriers all want to be the hub for such an ecosystem, too, so they can charge even more monthly access fees. Most large telephone and cable companies are trialing such subscription-based home-automation hubs, but they're likely to fail. One reason is that users hate these companies and their log track record of arrogantly bad customer service. The other is that their vision is quite narrow, covering a small collection of things and in a way that is about user lock-in rather than empowerment. Only a fool would trap themselves in a carrier-based offering.
Putting the carriers aside, we will end up with several large ecosystems for such individual-user devices. Some features will be proprietary to the ecosystem, such as Apple's restriction of AirDrop and AirPlay to its own hardware. Some will be partially open, such as Apple's iCloud, which offers iOS and OS X users full compatibility, Windows users limited compatibility, and none to Android users; or iBeacons, a protocol available to any beacons but whose client software runs only on Apple hardware.
Apple is the furthest along in this third IoT category, with its HomeKit APIs for home automation, CloudKit APIs for cloud storage and syncing, and HealthKit APIs for medical and fitness gear. Again, any service or device can use them. But only Apple devices run the client software, so they become the hub for all such devices. Developers can of course use others' APIs as well, such as Samsung's promised health APIs. (Assuming Samsung actually delivers; it's made lots of such promises for home automation but not executed well.) Beacons are a great example, with many equipment makers using Apple's APIs as well as their own, and waiting for Google to come up with its own.
Speaking of Google, the giant behind Android and Chrome OS hasn't come up with a cohesive set of technology like Apple's. But it's been investing in many of the pieces, including robotics and the Nest thermostat, and has been pushing Android for embedded systems as an alternative to Oracle's Java.
Then there's Microsoft, which keeps pushing its old Windows Compact Embedded, a full OS better suited to old-school computing devices and not IoT. Microsoft seems the least engaged in developing an IoT strategy, perhaps because its mobile position is so weak. Struggling mobile provider BlackBerry is also positioning its QNX kernel as an IoT basis. But it's in a better position because QNX is widely used for in-car infotainment systems and has the virtue of letting user-facing technologies like Apple's CarPlay ride on top -- so it can coexist with multiple user ecosystems.
Keep the three IoTs separate
Where all this leaves us is a set of distinct but overlapping markets all sharing the "Internet of things" label. They may share some technology underpinnings and some basic characteristics, but that's like thinking of PCs, networking, and databases as all the same because they are all computer technologies.
The Internet of things is many things. Understand the IoT that matters to you.
This article, "The 3 ways the Internet of things will unfold," was originally published at InfoWorld.com. Read more of Galen Gruman's Smart User blog. For the latest business technology news, follow InfoWorld.com on Twitter.