The Four Horsemen of the IoT Trough

Today, the promise of Internet of Things is moving from science fiction to reality. However, IoT is at the very early stages of maturity. We aren’t deriving huge value from the data coming from one set of devices talking to other sets of devices. And we have yet to build an ecosystem around the IoT. We’re still looking out into the future, focused on the promise, but ignoring the very real strategic and tactical problems that manifest today.

Another perspective on the state of IoT comes from the analysts. Gartner’s Hype Cycle puts IoT at the very zenith of the “peak of inflated expectations,” where early publicity has produced many success stories along with scores of failures. As technologists, we’re aware of the risks and other issues that can really slow down IoT adoption and create disappointment. What can we do to minimize the pain, the costs, and all the other risks related to the IoT? What can we do to bend the adoption curve?

The four things that can cause real risk for the IoT are security, reliability, economics, and expectations. I call these issues the four horsemen of the IoT trough. Overcoming these issues will help IoT quickly transition through the Gartner Hype Cycle’s trough of disillusionment and on to the slope of enlightenment.

Security
Security is the IoT challenge most often discussed, and the challenge that garners most of the attention of the mainstream media — for good reason. Security risks are real, widespread, and have potentially dire consequences. Somewhere around 2005 is when most of us bought our first IoT device — our home wireless router — even though we don’t think of it as an IoT device. By 2010, hackers on stage at DefCon showed that more than half of the wireless routers in the wild had security vulnerabilities in their firmware that exposed them to hacking. Five years later there still isn’t any way to automatically update firmware on these devices, and anyone can probably get into your home network if they try hard enough.

Unfortunately, today, there’s no codified security model for the IoT.

Reliability
The second challenge for IoT is reliability. Like security, reliability risks are real, and have real consequences. Consumers expect their IoT devices to be always-on, always working. If we push bad firmware update or provision a device incorrectly, all of a sudden, all of those devices out there are dead. No one wants to win the Guinness Book of World Records for creating the greatest number of doorstops in the shortest amount of time.

Keeping connected devices updated and working requires a new way to think about the entire chain of connectivity and communication to and from the backend. Industrial and consumer IoT use cases require reliable, resilient connectivity and failover, yet many IoT products today rely on very rudimentary point-to-point design patterns that fail at the slightest hiccup or scale increase.

Economics
The third IoT challenge is economics. Thanks to Moore’s Law, we now have $5 chips with fully capable processers and wireless connectivity. We can embed Internet connectivity into commodities like light bulbs, fire alarms, thermostats, and probably just about anything else.

Yet hardware and device economics are only part of the story. Early IoT developers rushed to build a market, leveraging economies of scale to push down the cost for their devices, while simultaneously ignoring the recurring costs of the backend. Supporting the IoT at scale requires investments in servers, systems, and people. Thus, there is a challenge to match the one-time device revenue against the ongoing costs of operating and collecting data and creating customer value from the IoT.

Expectations
The fourth challenge for IoT is meeting expectations. There’s tremendous excitement and hype about the value of big data and what we can do with the intelligence that comes from the data generated by all of these always-on always-connected devices. Critically, the right data must be shared at the right time, which provides big value across the entire spectrum.

Customer expectations are also driving objections to the IoT. When every light bulb, every camera, every smartwatch, everything’s interconnected, it becomes very easy to figure out where you are, who you’re with, and what you’re doing. Systems may decide that your temperature’s a little bit high and you need to go to the doctor, or that you’re depressed because your partner just left you. While some might find these features beneficial, others see them as an invasion of privacy.

The Repeated Technology Maturation Cycle
The Internet has evolved through four different waves. The first wave was the Internet itself and building static web pages. The first wave was revolutionary: we could bring up a browser and surf to any website to see what was basically a brochure.

Soon after, people envisioned the web as an application platform, and the dynamic web was built. This required app servers, database servers, enterprise web technology, and broadband connections.

As the web as a platform matured, we saw the rapid growth of mobile, creating yet another explosion of infrastructure built for the smartphone as its own application platform.

The latest wave is the move to the streaming web. The streaming web addresses the always-on phenomenon, where every device is consuming and emitting streams of data all of the time, rather than our traditional model of storing long-lived data in a database and retrieving that data on demand.

In each of these technology waves, four phases have happened as the technology matures. The first phase sees development teams building customized, one-off systems based on last generation’s technology. In the second phase, service-focused platforms emerge that deliver one-stop-shop solutions using proprietary technology and fill the gaps with large professional services projects.

The third phase inevitably sees a move from vertical platforms to horizontal stacks, as the market demands, creates, and enforce standards and compatibility across layers. Finally, mature stacks evolve and codify the way software development is done. For example, today’s web developers chose from a variety of equally capable stacks: LAMP, Rails, NET, Java, etc. A breadth of vendors at every level in the stack support many interoperable choices, and project teams chose the right stack for their needs.

The Internet of Things Matures with the IoT Stack
Going back to Gartner’s hype cycle, to minimize the trough of disillusionment and reach the slope of enlightenment, it’s important to help the IoT mature as quickly as possible.

As technologists, we understand that this requires the development and standardization of the IoT technology stack. IoT technologists are only just starting to think about the IoT stack.

What are the things that we’re seeing over and over again, and where are the places where we can start to have a conversation? Obviously, security has to happen in every layer. But we’re also seeing that business logic is happening in every layer. Unlike the world of the web or mobile, where most of the business logic happens solely in the app server, we expect to see business logic becoming distributed on the devices, as well as in the network.

Over the last couple of years, there’s been a huge amount of debate over protocols, both local area and wide area. Yet protocols don’t deliver solutions, and these debates highlight the immaturity of today’s IoT. There is a move toward some standardization and interoperability, but it’s not happening fast enough.

Next is the need for a data stream network. Just as the web needed CDNs to scale, the IoT needs a smart network for reliable delivery of data streams. But unlike a CDN, a DSN requires two-way low latency data flows, built-in reliability layers to support devices on unreliable connections, and new layers of security. A DSN needs to be able to provide multiple ingress locations for the data, provide filtering, smart routing, storage, and device presence detection.

As the IoT further matures, there will be an increasing need to syndicate aggregated data into new data streams that can be syndicated to business partners. The only way to avoid an integration project for every new partnership is to ensure we have standard concepts of data interchange and a well-understood stack for data streams and the IoT. And interconnecting everything is how we derive true value from the IoT.

To achieve this true promise of the Internet of Things requires that we first solve the challenges of the four horsemen — security, reliability, economics, and expectations.

Todd Greene is CEO and founder of PubNub (www.pubnub.com). He’s filling in this month as the author of Making Connections, which is typically written by Rich Tehrani, CEO of Technology Marketing Corp., the parent company of IoT Evolution magazine.