The flaw in the ‘Internet of Things’

The “Internet of Things” is back, with a sudden explosion of commentary about how it’s The Next Big Thing. It was The Next Big Thing about a decade ago as well, but turned out not to be. Maybe it’s back in the air because people are excited again about mobile tech innovations such as the iPad and are looking for more things to make exciting. Maybe it’s a secret plan to get people to finally care about IPv6 and IP address shortages; after all, if everything is connected to the Internet, they’ll need IP addresses, of which we’re running out in the traditional IPv4 form.

For probably the same reason the pundits’ view of the future always has flying cars and talking computers, the examples for the Internet of Things are household appliances: refigerators that order a new carton of milk and laundry machines that call for service when needed. If you think about it, though, these are silly examples that reflect lack of thinking: How will that carton of milk actually find it way into the fridge, and how often does the washer need servicing? And would you spend several thousand dollars to replace these items for the alleged convenience?

Empowerment doesn’t come through automation
My InfoWorld colleague Lisa Schmeiser suggested the key insight that these people are missing: The effective service providers of the last decade haven’t removed the user from the equation but gotten users to do the work of the providers — and enjoy it. That is, the user feels (and often is) empowered even though he or she is doing more of the work.

Think Amazon.com and pretty much all e-commerce. What these providers do is give users a lot of information and control acccessed over the Internet, and they create a relationship (as digital and fake as it may be) with the user that keeps him or her as a happy customer through a sense of command and direct fit.

That’s a better way to think of the Internet of Things: a medium for user-specific self-service. If your fridge is connected to the Internet, it won’t be to order milk but to help you compare your dinner ideas with what you have in stock. If your washer is connected to the Internet, it won’t be to call a repair technician (who knows who it will pick or how that person will get access to the house) but to let you know your washing patterns for insight into possible ways to save water and power. That’s the promise of those so-called smart power meters, though in practice they seem to be secretly designed to increase customers’ bills, not empower them to use energy better; instead, they serve as a cautionary tale of how not to do the Internet of Things.

I’m also dubious about the robo-nanny form of the Internet of Things, which today keeps coming up as the example of medication that has embedded RFID chips so a “smart,” Internet-connected pill bottle can monitor whether patients take their meds. (It turns out half of patients don’t take their medicine regularly, which is thought to increase overall medical care costs as they get or stay sicker than they should.) If you miss a dose, you get a call or text message. Reminders can be great, but this smacks of Big Brother — and it’s a slippery slope I’d rather not build. Why not just issue these people a time-release patch instead?

Forget the vision of anything with a power cord being connected to the Internet so that it can act autonomically. If that’s your goal, it appears that you really want to remove people from having a role in day-to-day life (as caricatured in the movie “Wall-E”).

The Internet of Things that works
Instead, the Internet of Things should be more akin to the benign aspects of the always-available Internet portrayed by science fiction writers such as William Gibson and Samuel Delaney. Information from lots of sources — websites, media creators, libraries, universities, satellites, road sensors, red-light cameras, and so on — is accessible for querying and correlation. You also have control over your own devices, such as activating your home alarm from the airport or catching up on your DVR’s recordings. This is already happening, in ad hoc ways — and that’s good.

We’re just at the beginning of this. The big wave still to come involves sensor data (called telematics) embedded in lots of devices. Some we have: traffic flow sensors, weather satellites, and GPS locaton tracking. As the technology gets cheaper and wireless broadband gets more ubiquitous (and less expensive — which still needs to happen), you’ll see sensors pop up in all sorts of places. For example, delivery trucks could serve as sensors for pollution, radiation, weather, chemicals, and so on as part of a broad net to aid regulators, police, and others improve quality of life.

Some businesses already do this; delivery firms such as FedEx and UPS have sensors throughout their delivery trucks so that they can monitor the vehicles for preventive maintenance. In addition, the drivers can micromanage how they drive to optimize schedules, fuel consumption, and the like. Many factories have similar sensors embedded throughout their manufacturing equipment for similar purposes.

One key lesson from these uses is that not every device has to be wirelessly connected to the Internet. They can connect to a central reporting device that uploads their pooled data through a single connection. This is key for both not wasting IP addresses and keeping the radio spectrum from getting clogged or too pricey.

A new opportunity in support of this form of the Internet of Things is presenting itself. Mobile devices such as smartphones and slates typically have several sensors in them — location detectors, accelerometers, compasses, ambient light sensors, audio recorders, and cameras — that could provide contextual information to make apps smarter about the user (she’s driving, so switch to a voice interface) or about local details (check the photographed bar code and see who sells that product nearby). These personal sensors could also provide environmental context for use in overall quality-of-life tasks, such as for transit optimization, though issues of spying and privacy will be barriers. People will need to trust any data is anonymized at the source.

Unlike a decade ago, there’s more ability today to create a useful Internet of Things. Let’s hope the smart ideas that get follow-up attention, not the milk-ordering refrigerators.

This article, “The flaw with the ‘Internet of Things,'” was originally published at InfoWorld.com.