The 3 ways the Internet of things will unfold

The Internet of things is hot. Practically every tech vendor is using the label for some of its products. Cisco Systems and PwC both predict that the market will be worth trillions of dollars. The Internet of things is also the tech industry’s latest overhyped technology — most of what is called IoT is not IoT, and the IoT market will never be worth trillions of dollars unless you declare that IoT includes anything that uses power, a chip, and some communications capability, which is a pointless definition.

Despite the tech industry’s fierce attempts to scrub all meaning from the IoT label, something real and valuable is occurring in the Internet of things. But users and IT organizations can’t take advantage of it without understanding what’s going on, which is what this post explains.

[ What does the “Internet of things” really mean? Glad you asked! InfoWorld provides a sensible definition for IoT. | Liquid computing is the next sea change for both users and IT. | Subscribe to InfoWorld’s Consumerization of IT newsletter today]

Several technologies are making IoT widely possible, mostly from the mobile space.

One is the low-power processor, typically based on the ARM designs already in use by nearly all smartphones and tablets. They’re much cheaper and smaller, as well as more power-efficient, compared to traditional Intel and AMD x86 chips. About 40 percent of them are used in devices you may not expect, says ARM Holding marketing VP Ian Ferguson, such as in-car infotainment systems. Companies like Texas Instruments also make a bevy of chips — some based on ARM, others not — that power everything from sensors to alarm clocks to garage door openers to beacons.

Then there’s Bluetooth and Wi-Fi, including the networkless Wi-Fi Direct (WiDi) variant standard in recent mobile devices and computers. (This is how Apple’s AirDrop works, as well as Windows’ and Android’s Miracast.) The two networking standards are commoditized, so they’re finding their ways into all sorts of devices.

Industry efforts like Thread are trying to develop a constrained communications standard that lets devices communicate over a common protocol (as opposed to a radio technology) without the full computation and energy consumption overhead of the typical IP stack. The Thread effort sees IP at the edge of the stack, so the low-requirements communications eventually can connect to the Internet and heavier-weight systems, but it doesn’t force every component to be able to do so.

Basically, it’s getting cheap and easy to put a chip in it. And it’s getting easy to add coprocessors for everything from motion detection to radio connectivity, from graphics processing to encryption. More devices can compute and connect as a result. Power sources are shaping up as the limiting factor, so there’s lots of research on everything from better batteries to converting radio waves or motion into power.

There are three clusters of real IoT activities, and each is on its own path. Some paths may cross, but understanding the three separately will help you strategize about your own IoT engagement:

• Machine-to-machine is simply about efficiency, not fundamental new opportunity
• The notion of smart systems will gain traction, with Bluetooth peripherals as the first step
• The ad hoc Internet of things is well under way

Machine-to-machine is simply about efficiency, not fundamental new opportunity For decades, we’ve had industrial, medical, and office equipment that could talk to other equipment, such as thermostats that communicate temperature information to normalize HVAC settings, assembly-line sensors that let robots know to stop welding if the line is delayed or stopped, and EKG readers that alert a nursing station if worrisome readings occur. This is known as machine-to-machine (M2M) communications, and it’s really useful.

These established M2M uses are now getting the IoT label, but they are not really changed by IoT. However, they’re cheaper and easier to deploy because of greater technology standardization that is making the larger IoT trend possible. We’ll thus see the “industrial Internet of things” (the new name for M2M) become more widespread, as smaller companies can afford to join in and larger companies can afford to bring the notion outside of expensive manufacturing systems.

It’s like when PCs arrived in business: Suddenly, a computer didn’t cost millions of dollars, so computing could go beyond the data center. Until recently, M2M technology was about at the level of business computing technology in 1980 in terms of cost and reach.

What’s made M2M easier and cheaper to deploy? Bernie Anger, the general manager of General Electric’s Intelligent Platforms division (a big M2M vendor for industrial automation) points to three factors.

• ODBC User Agent adoption: This version of the venerable database connectivity protocol is not Windows-dependent, so devices on all sorts of platforms can now share data through a known protocol, not just PCs or devices running Windows Embedded. Due to the relatively low cost — ODBC UA-capable devices with local computation ability and network access cost just $200 — it’s affordable to have more devices connected.
• Hadoop and similar mass-scale data processing technologies allows analysis of massive data in cost-effective way. When analytics was an expensive, scarce resource, companies limited what data they collected and analyzed to the most critical areas. Now they can apply analytics to more areas, and they’re doing so.
• The ubiquity of the HTML5 Web standard in client devices: That means more than the use of iPads, smartphones, computers, and other off-the-shelf equipment — it also means that specialty devices now use a client UI that’s well understood and compatible with all the computing devices you have. The burden of writing to proprietary user interfaces is greatly reduced, and operator familiarity is greatly improved.

“None of these is a revolution, but they come together now to enable the scale and speed not possible a decade ago in the M2M/SOA worlds, when everything was essentially custom, nonstandard, and heavyweight,” Anger notes.

Over time, the use of standard protocols and technologies will allow the “back end” M2M systems to interact with user-facing technologies, which will provide some white-knuckle moments for the guardians of the core systems while making them more valuable overall.

The notion of smart systems will gain traction, with Bluetooth peripherals as the first step Connecting M2M systems to the rest of the world will scare many IT pros, though “rest of the world” really means “other parts of the world.” But it’s inevitable because it’s so useful. I recently profiled a simple example of a utility company managing its systems via iPads to be able to respond to problems faster. That’s the simple example we’ll see first.

That example is not restricted to M2M systems. It’s basic field service, and it’s happening in all sorts of ways. For example, some oil rig equipment has sensors that a field engineer can tap into via an iPad, then communicate to home base over satellite or other communications systems to get diagnostics and proposed fixes, as well as interactive manuals. The same is true for airplane engine and copier repairs.

This too is not a new area. What is new is using consumer-grade equipment like the iPad, standard communications technologies like Bluetooth and Wi-Fi, and standard application languages like JavaScript and environments like HTML5. Again, it’s old-fashioned industrial computing made easier through modern technologies, then rebranded as IoT.

You’ll see more of it. As an example, Motorola Solutions recently announced a Bluetooth barcode scanner that works with Android and iOS devices. Historically, such scanning equipment is proprietary and expensive. Special training is also required to use and maintain them. By making the scanner a Bluetooth peripheral for common mobile devices changes that equation. Now, your employees can use equipment they likely know how to operate, using Web or native apps that are familiar to them to control the peripheral.

There are many examples of using mobile devices as computing hubs to sensors and specialty peripherals, especially around Bluetooth. This is going to change the specialty-gear industry in profound ways. It’s already altering the consumer sphere, with everything from fitness wearables to ice-fishing aides. But cheaper, easier-to-use equipment running on common devices is merely the first step.

What comes next is what I call smart systems. Because these peripherals are connected to, in essence, mobile computers and those in turn can be connected to the Internet and all available cloud resources, they form a network of both data and operation. This is where the “Internet of things” label rightfully applies.

What do I mean by that kind of network? Think of a delivery driver, who now has a signature pad that collects your signature and lets the driver input status like “addressee not home, package not delivered.” The data could then be transmitted via a radio in the truck so that the shipping info is updated on the tracking website that both the sender and recipient can monitor.

That existing technology has proven quite useful, but imagine if the signature pad were a peripheral or used a tablet’s touchscreen. If the addressee is not home, the driver could take a photo of where the package was left, so the person knows where to look for it — or of the menacing dog preventing delivery or of the person who signed for the package (because they almost never ask for ID). That’s just the camera. If Bluetooth-powered door locks ever take off, they could interact with a back-end service for which the addressee has door-lock access rights temporarily, so the driver could open the door to leave the package in a safe location.

As the day progresses, the status of deliveries could be compared for nearby trucks, allowing transfer of cargo to equalize workload — or even shift packages to a second driver who can revisit an address that day knowing the person is now available, rather than ship the package back to the distribution center and try the whole process the next day.

All the pieces exist in some form today, but their distribution is uneven — both the hardware and the cloud-connected apps. As they become more common, we’ll get smarter interactions that let us improve service in a whole range of fields, not simply package delivery. A quick example: A smart pillbox coupled with the sensors in a wearable or smartphone could allow remote monitoring of patients anywhere and provide a way for the patient to engage back (show a photo of the pill — is it the right one?).

Such technology is already in trial sessions. But these trials focus on the center monitoring the edge: the patient, the home alarm, and so on. They’re not so focused on the reverse, which is letting the edge query the center — the delivery service, the doctor’s office, the alarm company.

That’ll come after the center-to-edge uses are in place. Once a connection is set, exploiting it in two directions becomes much easier. Then we’ll see connections across multiple systems in a federation, in the same vein we’ve seen in the Internet and the cloud.

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.

Apple is using non-network connectivity to make its Macs and iOS devices federate on the fly in what I call “liquid computing,” bringing IoT to more than small devices.

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.