Building the IoT: standards and hardware needs
When I was a teenager, or even at the beginning of my professional career, the answers I gave to questions I was asked were yes or no, black or white. Today, when I am being asked the same questions, the answers are more like, well it depends … nothing is black or white. So, if you ask me, does the IoT exist? Well it depends.
Let me explain. Let’s start with a high-level view of IoT infrastructures.
The diagram in Figure 1 is a high-level architecture. Although not exhaustive, it shows the main components comprising an IoT system:
- Embedded hardware
- Embedded software
- Wireless sensor network(s)
- Mobile network(s)
- Mobile operating system(s)
- IP network(s)
- Information technologies (cloud back-end services)
- Enterprise resource planning and related IT
Figure 1: Typical IoT Architecture
In addition to architecture considerations, we must also understand the division of hardware equipment by device category to help evaluate the challenge in deploying a global IoT infrastructure.
Figure 2 shows that there will be many more small-edge devices (sensors/actuators based on microprocessors and microcontrollers) than there will be networking or user devices. Because of the sheer volume of devices, the cost structure of a sensor/actuator device is crucial to establishing a profitable business case. We need to look at all the elements composing such a device and identify where the challenges are.
Figure 2: Intelligent Device Hierarchy
(Source: Harbor Research)
With IoT infrastructure and IoT device categorization broadly defined, let’s look at the developments required in our industries to make IoT a reality.
There are markets where the current state of technology—including sensors/actuators, processors, wireless networking, the internet or use of internet protocol (IP)-based networking technologies, mobile networking (cellular and the smartphones/tablets), and some form of back-end storage and processing/analytic capabilities (i.e., the information technology (IT) portion of the cloud)—are producing products and systems that can sustain a business (i.e., generate profit). The consumer market is a perfect example.
Consumer IoT vs. industrial IoT
The fact is, as consumers, we accept product behavior and usage that we would not accept in other markets. For example, we accept that our smartphones, tablets or PCs need to reboot. We accept a certain level of non-secure processing, communication, or storage. We accept the need to perform regular software upgrades on our consumer products. We embrace changing our smartphones (or wanting to change them because we want the latest/greatest) every 18 months. We accept needing to replace our tablets/smartphones every two years or less, but, for example, we accept none of these conditions for our thermostats. This is why I often refer to the consumer market as the IoT laboratory—ideas, concepts, systems, technologies can be tested here.
However, consumer product rules can’t apply to medical or industrial products. For a given industry installing equipment to automate a process, equipment cannot be replaced every 18 months because the cost/benefit goal for that industry will rule out this behavior. When a manufacturing or building automation equipment is put in place, it is for 10 to 20 years.
Not so long ago, industrial equipment was electromechanical in nature. It had a fixed function to perform, and it was designed to do only that. Electronics and processing capabilities have been added to these components. Now, these devices do something in a larger system and are designed to work in collaboration with other devices and systems. New functions or bug fixes can be uploaded to the devices thanks to the processing capabilities and firmware/software in them.
This raises another issue. If the equipment is to be in place for a minimum of 10 years, how do you ensure the piece of equipment itself has the right hardware resources to allow for upgradability? Do you need to overdesign the resources to plan for a longer operational life?
Another way to look at the differences between consumer IoT and industrial IoT is shown in a table from Patrick Morehead, from his Forbes article titled: Who Wins in the Industrial Internet of Things (IIoT)? (see Figure 3).
Figure 3: Differences between consumer and industrial IoT
It is becoming clear that there are many technologies that need more development and economies of scale for IoT to achieve the size predicted by many business analysts and media. This idea of the IoT under construction is well-represented in a report from Gartner, the IT research company, called: Hype Cycle for Emerging Technologies. In this report, Gartner forecasts the adoption of different technology trends.
Gartner’s opinion about IoT is interesting. In 2015, Gartner believes the IoT to be in the "Peak of Inflated Expectations" phase, and that it will take five to 10 years to reach mainstream adoption, defined as the "Plateau of Productivity." In other words, the IoT will be under construction for the next decade (see Figure 4).
Figure 4: Hype Cycle for Emerging Technologies, 2015
(Source: Gartner )
There are so many predictions about the IoT that sometimes it is difficult to make sense of all this “hype.” I was always under the impression that the predictions of the billions of devices connected and the trillions of dollars injected in the economy were hardware-related numbers. However, a paper from IDC titled Worldwide and Regional Internet of Things (IoT) 2014–2020 Forecast: A Virtuous Circle of Proven Value and Demand offers a practical approach at analyzing the various industry sectors. IDC defines the “internet of things” as a network of networks of uniquely identifiable endpoints (or "things") that communicate without human interaction using IP connectivity—be it "locally" or “globally.” Clearly, we need more than hardware to meet this definition.
In Nelson Hsu’s article, Which ‘Internet of Things’ Forecasts Are Believable? posted on gurufocus.com, he was able to identify that about 50% of the value created by the IoT will come from connectivity services, infrastructure, purpose-built IoT platforms, applications, security, analytics, and professional services.
This illustrates that IoT is a flexible term. It can be used for many business opportunities. It all depends on who is using the technologies that enable the IoT and what one does with them. If the technologies we need to build the IoT all exist, why is the IoT not yet pervasive in 2016? Well, simply, because some important required features to make the IoT scalable and transferrable to non-consumer industries are still missing.
In fact, the technology needs to mature a bit more to pass the Gartner "trough of disillusionment," but when it does, the IoT’s acceleration will be phenomenal. We will see another typical hockey-stick effect in the number of units deployed. Even if a full-scale IoT is not yet in place, we understand enough about its potential to predict that it will outlive the hype. We just need to work on those areas needing further construction.