What’s Next for the IoT?

The Internet of Things continues to evolve, attempting to overcome its poor reputation for cybersecurity and making the case for wider adoption, especially by enterprises. Consumer IoT, largely represented in smart-home automation, remains a market being targeted by Amazon, Apple, Google, LG Electronics, Samsung Electronics, and other technology titans.

The big bucks are in Industrial IoT, though. That market has attracted AT&T, IBM, Microsoft, Oracle, Verizon Communications, and hundreds of startups. Some of those startups, such as C3 IoT and Uptake Technologies, have achieved “unicorn” status and attracted significant investments. The Chicago-based Uptake is a shining example of the industry transition from platform-as-a-service business models to software-as-a-service.

Many of the savvier startups are adding artificial intelligence and machine learning to their technology portfolios, complementing their IoT focus.

Ron Lowman, strategic marketing manager for IoT at Synopsys, points to the development of the Industrial IoT, machine learning, audio technology, and voice recognition/processing as important trends in IoT. He notes that connectivity is the big trend in the IoT market and in IoT technology.

“Industrial IoT has been a big conversation piece,” he says. “It’s really moving out of what was smart home last year. There’s now a lot more smart farming, smart industry, smart streetlights – things that have been around, but now they’re becoming a reality. We’re starting to see expanded installation of those with respect to chip design. Obviously, one of the big things is machine learning. This is a pretty wide topic. There are all kinds of things, from chatbots to object identification, and so on. Voice recognition is part of that in some form or fashion. And the last thing that is changing is maturity — maturation of standards. Of course, some consolidation we’ve seen over the past couple of years has provided a little bit better path for customers to focus on. People were going in a hundred different directions, at 100 miles per hour, so now that’s starting to be more of a complete picture.”

Standards and issues
Complete does not mean perfect, though. The Industrial IoT platform markets are seeing “clashes” among information technology, operational technology, and IoT-focused providers, notes Emil Berthesen, a principal analyst at Machina Research/Gartner. “End-users are experiencing the challenge of moving from proof-of-concept and pilots to scaled implementations.”

Across many IoT markets, particularly those with battery-powered edge devices, power remains an issue. Energy harvesting is improving, but progress has been slow. The same is true for battery technology. This has spurred growing interest in low-power wide-area networks and narrowband IoT technology because connectivity—especially wireless connectivity—is so central to the Internet of Things, Berthesen says.

Fig. 1: Vibration-based energy harvesting unit. Source: Mouser Electronics

It also has fostered some new ways of thinking about designing chips for these markets. One of the most striking changes is the move to more customized solutions. One size no longer fits all.

“You can no longer design a transistor and assume it will work under all circumstances,” said Sudhir Sharma, global marketing and strategy director at Ansys. “Some of these devices are expected to last for 25 years. If you put sensors across a die, you may find that you do not want to turn off parts of a chip anymore because there are long-term implications for temperature gradients. We’re paying much more attention to thermal issues and signal integrity.”

New standards are a reflection of this trend, as well. The Bluetooth 5 wireless protocol has had a significant impact on IoT uptake, and it has done so without cannibalizing use of other protocols. There also is a new USB standard that addresses audio, with more reliance on voice processing and voice recognition.

Security
One of the biggest problems involves security. When devices are standalone, security requires a physical breach. Once they are connected to the Internet, they can be attacked remotely.

“The IoT is scary wherever you look at it because of the damage that can be done,” said Asaf Ashkenazi, senior director of product management in Rambus‘ Security Division. “Everything is about scale. So if you attack one washing machine, it’s not the end of the world. But you can have a massive attack on the same model of washing machine, where you turn all of them on at the same time and take out the power grid. A million devices that are compromised can be recruited into a botnet. You can cause havoc with a sensor to measure air quality, too. You can create a pattern that looks like a leak and turns all the alarms on. When you think about security in the IoT, it’s not about one device. It’s an army of clones, because you make many devices exactly the same.”

And this is just the beginning, because more appliances and devices are being sold now with connectivity built in. “Security remains a challenge in IoT, which brings into play a wide range of physical and cyber security issues,” says Berthelsen. “Many companies are starting to address the needs of enterprises, but equally so threats are becoming more and more prevalent. We need to begin to look at combining security frameworks from IT and OT into IoT.”

What has changed, though, is awareness of security is increasing.

“Security is one of the biggest trends in technology,” observes Synopsys’ Lowman. “People are asking detailed questions about what types of security do we need, how much is good enough, how does this align to standards or the recent breaches that they’ve seen, and how do I protect myself? We’re actually seeing people get serious about it. And that’s just in software.”

For better cybersecurity, developers of IoT need a true random number generator, Lowman says. It’s good to incorporate Arm’s TrustZone security technology in chip designs, but there’s more needed to build a root of trust in hardware, software, and systems.

Segmenting markets
When the IoT first began making headlines, most analysts and companies described the IoT as a single market. That’s no longer the case.

Berthesen sees three distinct IoT platform markets emerging—industrial, commercial, and consumer. There are also new architectural developments with edge computing/networking technologies. And there are new ways of using technology that never existed before.

Within those three segments, though, there are a lot of variations. Beecham Research CEO Robin Duke-Woodley estimates there are more than 450 platforms now on the market, compared with 14 in 2008. “Everybody’s got a platform,” he says. “It’s been increasing almost exponentially over the last several years.”

Duke-Woodley doesn’t see any rationalization of the platform market in the near future. And that may be the point of the IoT in the first place. But it also is serving as the foundation for some major innovations in business, such as asset avatars, device shadows and digital twins in the Industrial IoT arena.

Digital twins are a step toward predictive maintenance, says Ansys’ Sharma. “Basically what you’re doing is creating a physics model, which can be used to measure efficiency and uptime. The goal here is zero unplanned downtime. Physics plus analytics equals a digital twin.”

Fig. 2: GE’s digital twin airplane jet turbine. Source: GE

The benefits to businesses don’t stop there, either. “Industrial IoT remains one of the more interesting opportunity areas due to the positive market response to predictive maintenance solutions,” says Berthelsen. “Equally, other segments, such as agriculture, health care, and intelligent buildings, have started to leverage the benefits of IoT.”

And as enterprises transition from product-centric business models to data-driven organizations, they want to own and manage their own data in a clear and transparent way. The challenge of scale, the utility approach to data, and value of data are driving these changes, he says.

AI, machine learning
AI and machine learning startups are drawing the most interest from venture capitalists, and that potentially will have a big impact on the IoT.

“How much of this stuff is real, and how much of this stuff is still basically a giant science project?” asks Jeff Smith, CEO of Arago US. “The real question is how much of this is real in addressing a business problem in a way that today we can’t. There are two ways to think about this. One is AI is a science project, the other is a game-changer. The question for the customer is, ‘Which do I want? Am I interested in a science project to learn, but I don’t really have an appetite or understanding or a risk profile that lets me take on doing something with this if it works? Or am I looking for a game-changer because I have a significant economic or operational business opportunity? I need to fix that opportunity and to address it at speed. And I need something that will commit differently from what my competition can do. Can AI help me with that?’”

The answer appears to be yes. “Why should we pay attention to AI? In part, it’s because of the data explosion,” Smith says. “The practical reality is the difference between the amount of data that we produce and the way that we can and should be interacting with. Leverage—knowledge leverage, and the ability to have some leverage as humans in handling the data volume so that we can get to the stuff that matters, versus having to deal with everything before we know what matters or not, is a very powerful, compelling reason to pay attention to artificial intelligence.”

There is also the continued development of AI/machine learning to enhance business outcomes, and the move from PaaS to SaaS in IoT architectures, he notes.

What’s also becoming apparent is that technologies adopted in one or more IoT segments can completely change adoption rates in other segments. Consider Long-Term Evolution (LTE) connectivity, which is a foundation technology for connected devices.

“It’s all LTE,” says Joel Young, Digi International’s senior vice president of research and development and chief technical officer. “We see a push away from a lot of local networks and landline connected to a cellular-based connection, and that’s universal. In the market of digital signage, we see a great push in Cat 1, actually. When you think about it, maybe you don’t need the bandwidth of a Category 4 or a Category 8, but you want just enough for digital signage. Moving down the line, we’re seeing a big transition now from short-range RF to LTE-M and also narrowband IoT for a global company. In the U.S., it’s all LTE-M for everything from monitoring all those systems, like whether it’s streetlights or sensors that used to be short-range about gateways, right, because they want to have simplicity. And in Europe, you’re seeing a big push on narrowband IoT for those same sets of applications.”

Others agree. Bill Kramer, KORE’s vice president of advanced technology adoption, adds, “We see very similar things in our connect product, which has really been guiding most people up to this point in time. We’re really talking Cat 4, Cat M-type solutions. They’re used for digital signs, they’re using it for business continuity, or a quick startup kind of a solution replacing those wirelines, where it’s a much more economical play. The demand for M and V is tremendous. It’s opening up a lot of use cases. More importantly, for those customers in North America, it really is the alternative to those legacy 2G, 3G devices, on to LTE, that isn’t cost-prohibitive.”

The future
What still isn’t clear with the IoT is how all of the platforms will mesh, though. All of the big companies have their own platforms, and startups generally begin from scratch in designing their IoT platforms. There are similarities, but few industry standards in this area, although the Industrial Internet Consortium is now partnering with the oneM2M IoT standards initiative.

“Of course, not all of them will survive,” says Synopsys’ Lowman. “The thing about these platforms is everybody wanted to have these to give customers time-to-market. But what people are starting to identify is that these things need to do much more than identify quicker time-to-market. They actually have to provide some differentiating features in their end-product.”

Most market observers say it’s still early days for the Internet of Things, and the challenges in the field, such as cybersecurity and power will be ironed out over time. But at least for now, there is still enough churn under the IoT umbrella to make international growth opportunities difficult to assess.

At some point that will change. “Consolidation has started to take place, and will continue to do so in the next one to three years,” Berthelsen says.

But until then, expect a level of confusion to persist alongside the march toward order. “Everyone says IoT is simple,” notes Lisa-Anne Uhrmacher, Vodafone’s interim head of IoT sales. “I’ve been in IoT a long time. It’s just not that simple.”

Related Stories
IIoT Grows, But So Do Risks
Things are coming together for the Industrial Internet of Things, but security is a huge and growing issue.
Data Leakage And The IIoT
Connecting industrial equipment to the Internet offers big improvements in uptime and efficiency, but it adds security issues.
Industrial IoT: Where Is Silicon Valley?
The geography of the industrial automation supply chain.
The Rising Value Of Data
Race begins to figure out what else can be done with data. But not all data is useful, and some of it is faulty.
Safety Plus Security: Solutions And Methodologies
Connecting to the Internet adds new demands for safety-critical markets.