Beware the Internet of Things hype

The "Internet of Things" is a misnomer. Maybe that's why Internet of Things forecasts are often exaggerated.

Illustration of Internet of Things

Now that we’ve connected nearly every person on the planet, the focus has shifted to connecting things.

No doubt more and more things will be connected to the internet. However, as I discussed in a previous post, predictions about the scale and pace of growth are inflated. This is partly because the “Internet of Things” is a misnomer.

Estimates of the number of things already connected to the internet (2016) are all over the map, ranging from around 6 billion to over 15 billion. The only possible explanation for such a large discrepancy is that forecasters don’t agree about the types of things that qualify. For instance, should a wearable device that uploads data to a smartphone once per day count?

Similarly, forecasts for the number of things that will be connected to the internet by 2020 run from 26 billion all of the way up to 50 billion. Analysts become giddy when they realize that there are well over one trillion things that could potentially be connected to the Internet. Though the most optimistic forecasts assume that RFID or NFC tags will be attached to everyday items, and that somehow qualifies them for membership in the Internet of Things.

Note that even the most conservative forecasts would require connecting at least 3 billion new things to the internet each year. That works out to more than 8 million things per day. Connecting that many things isn’t impossible -- consumers worldwide buy well over one million gadgets each day. However, it took the mobile phone industry more than 30 years to reach 6 billion total subscriptions. Remember also that forecasters were predicting tens of billions of things would be connected to the internet within five years back in 2004.

Not surprising, there is a great deal of confusion about what is meant by the Internet of Things. Ten years ago, it was understood that the Internet of Things referred to machines as distinct from consumer devices. Since then, the dividing line between consumer devices and machines has blurred. Today, your phone is as adept at tracking your motion and location as it is at letting you make and receive calls. Consequently, the Internet of Things now includes not only sensors and actuators in oil pipelines, but components in smartphones and televisions that can be queried or instructed to execute commands. Perhaps that’s why companies such as Cisco, Intel, and Qualcomm talk about the “Internet of Everything.”

The Internet of Things lumps together markets at different stages of development that really deserve to be examined separately. It doesn’t help to speak about opportunities in environmental monitoring, health care and transportation as if they are just different facets of one huge market. The buyers, motivations, and application requirements in each of these industries are very different. Understanding and addressing their specific needs is key to success.

There are also some technical issues that challenge widespread assumptions about the Internet of Things. While many types of things can benefit from connectivity, they are not necessarily best served by being connected directly to the internet.

Arguably, every human being should have internet access, because the internet is a fabulous resource for learning, communicating, and shopping. But not every sensor and actuator needs to be directly connected to the internet.

Francis DaCosta, author of Rethinking the Internet of Things, uses the Mars Rover as an illustration. It makes more sense for the Mars Rover to analyze its own sensor data and make on-the-spot navigation decisions than to send the data all the way back to Earth and wait for commands. That doesn’t preclude Earth-based engineers intervening when the Mars Rover isn’t making sufficient progress. However, efficiency, privacy, and security are compelling reasons for keeping as much of the decision-making local as possible.

Similarly, requiring every device to support the internet protocol (IP) is self-defeating. Sensors have come down in size, price and power consumption, enabling them to be used more widely. Insisting on IP support adds back cost and complexity. A better arrangement is to permit simple sensors and actuators to communicate with local hubs possessing the intelligence to filter, reformat and package data as needed.

What’s driving the connection of things to the internet? Things are used to collect information in the field and apply it where it makes the most difference. It’s this rich flow of information that enables digital business models and digital user experiences.

For instance, medical gadgets can be used by patients to detect health problems before hospitalization becomes necessary, to shorten hospital stays, and to reduce readmission rates. Dr. Eric Topol predicts that the continuing advance of medical technology will lead to a future with far fewer hospitals. Hospitals will be mainly for intensive care.

There has been too much chatter about the fantastic number of things that will be connected to the internet. Instead, the focus should be on how individual industries can connect the right things to create unprecedented value for customers.

This article is published as part of the IDG Contributor Network. Want to Join?

The march toward exascale computers
View Comments
Join the discussion
Be the first to comment on this article. Our Commenting Policies