IoT needs a Paradigm Shift

IoT, the Internet of Things, has been a promising concept for over 30 years. It expands the internet from a system that is primarily used by people to communicate and share information, to one where equipment, devices, machines, vehicles, sensors and countless other objects also communicate and share information – with each other and with people. The benefits of connecting “things” to the internet are extensive and significant. IoT has enabled new business models, has completely changed some industries and is one of the main drivers of digital transformation.

In spite of its promise the uptake of IoT consistently lags its potential. IoT falls behind each growth prediction made by reputable organizations. Like every new technology, IoT was overhyped at first and rationalized over time. However, with its long slow track record, there seem to be other factors in play than normal over-enthusiasm would suggest. What are those factors?

?The problem with IoT – Part I: IoT is complex

There are many reasons why technology driven projects and innovations can fail, and that holds for IoT as well. Three factors appear to have specifically added to IoT’s challenges. First, IoT is complex, and its complexities are often underestimated. They include technical, conceptual and implementation aspects. Some of them are:

IoT systems combine various technological solutions and skills

In many applications each of the individual technology solutions that make up the IoT chain needs to be at state-of-the-art level in order to meet the economic requirements. IoT solutions are a combination of at least a field device, data transmission technology, a back-end or cloud application, a user interface and data processing applications. In many instances the sensor or the asset that is being connected is also part of this chain. This means that companies that create end-to-end IoT solutions need have solid expertise in each of these areas.

Very few companies are set-up to cover the entire collection of expertise and skills. To solve this, companies begin forming “eco-systems” of cooperation partners, each with different expertise. This works well for some, but the “give and take” cooperation that this often entails does not always align with the nature of companies that are more focused on the latter and less on the former.

IoT is conceptually difficult????????

Although it has improved significantly over the last 5 years, as IoT is being implemented on a wider scale and data consumption for processes and decision making is starting to take off, the value of IoT is often difficult to see at first. In many IoT projects, the value becomes obvious only after implementation. Sometimes this takes years, for example in predictive maintenance applications where the use case requires a set of historical data in order to train the predictive models. This can be understood intellectually of course, but many recognize only after the use case has actually been applied in practice why it was worth the investment.????????

The creation of an IoT solution takes time??????????????

To complete an IoT solution from initial conception until the start of full-scale implementation typically takes at least two years. That in itself is not a problem, as many new initiatives take time to complete. There are many that take longer and that are in fact very successful. The problem for IoT is that this is often not anticipated. Instead many projects start that expect a lead time of a few months. Although there are examples of solutions that were completed in such a short time, most do not. This means that initiators and champions do not deliver to the expectations of their organization, lose credibility and struggle with keeping budgets for their project. In addition, the team that initiated the project often dissolves over time as members move on to other position to further their professional career, and those that replace them usually do not have the same commitment level to the success of the project.

These three factors have contributed to the sad fact that about two thirds of IoT projects fail or never graduate from the pilot phase. Disappointments from failed projects have set back the progress of IoT in organizations, sometimes for years. In our own practice we encountered a textbook example of a failed IoT project about ten years ago when we were first approached by a company that since has become one of our clients. This company, a multinational, had launched an IoT solution in its home country. This project encountered many problems – all related to those three factors. These were of such significance that the country manager shut down the project and turned away from IoT completely. In the meantime the project team had introduced the solution in some of their other countries as well, who had smaller operations and more patience. They asked us to support them sorting through the issues, make the system’s operation reliable and add functionalities that deliver actionable information for the users. That process itself took about a year, but the system has been in service in those countries ever since and helped them in quality improvements, costs reductions and faster respond times.

It took close to a decade for this company to start looking seriously at IoT again, at significant opportunity costs. It is now on an accelerated path of catching up across its organization, partially because some of the countries that did continue with the original solution started their own, small scale IoT projects for other use cases without much coordination between them. And they are generating value from those.

The problem with IoT – Part II: IoT is for engineers

The story of IoT is told by engineers to engineers, in language they understand. The story is typically not translated to the language of the C-suite. IoT is seen as a concept where “OT meets IT" – in other words where Operations Technology and Information Technology come together. In many companies these are managed by different departments and are silo’s that do not work together well. This in itself is another reason why IoT often experiences problems. If an IoT project is initiated from the perspective of operations, it needs to use and have access to IT technology across the chain, as described above. If the project is initiated by the IT department, they need to embrace “OT thinking” to make the project successful. For both areas this is difficult. We have seen more examples of successful projects that were started from the OT perspective than those that were started by IT specialists, but both paths are challenging.

Against this background, and the wide array of technical expertise and skills that are needed, the discussions about IoT often remain on a very technical level and focus on the added value that it can have for operations. The company’s management however is interested to know what it can do for its business. Only in 2021 I heard for the first time a company made an argument that they initiated an IoT project because it would allow their CFO to have information about the company’s actual performance faster – as the data from the measurement of the liquid levels in their remote storage tanks gave information about the financial performance of the company weeks before the invoicing information in their books and the cash flow through their accounts. Alan Greenspan started tracking movements of railway tankers to improve his macro-economic models when he was a young economist. IoT can do the same for companies, and can allow its leaders to take better decisions. This fact is usually not included in the internal memo that justifies the budget request for an IoT project.?

The problem with IoT – Part III: IoT is more about opportunities than problems

Selling solutions, especially if they are innovative and unfamiliar to customers, works better if the solutions solve problems. IoT solutions can solve many problems. Those problems are not always urgent enough for companies to embark on a path filled with unknowns, and if competitors keep living with the same problem the sense of urgency is even lower. In many cases a problem needs to be large and important before a company will initiate IoT. In our own practice almost all IoT projects started at that point – a company or department wanted to solve a problem that kept the manager up at night.

This does progress IoT but drives bespoke implementations that often deal with one specific objective only. These solutions are usually not prepared to act as infrastructure for other solutions. In “smart buildings”, an area where IoT started taking off about 5 years ago, many developers are required to install multiple gateways as each gateway supports only one or a handful of specific solutions, sensors or use cases.

The real value of IoT is often in its opportunities. IoT can change complete business models and enable new services and propositions. However, there are far fewer IoT projects that are started from a vision to capture new business opportunities. They do happen. In our own experience these initiatives all shared one common characteristic: a person that understood this vision, had the position or power to drive the company to starting the project and had the ability to provide the overall project management needed for the case – based on a thorough understanding of what does and does not work in the specific industry.?

IoT’s traditional approach

The complexity of IoT has led initiators of new IoT projects to structure the IoT solution into different sub-systems. The split into these sub-systems follows closely the elements that are mentioned above – one team develops the IoT device that is connected to the asset, another team the back-end or cloud application, yet another the user interface and nowadays often also a “data team” that works on advanced data processing. As there are very few companies in IoT that can handle this entire chain, these teams often operate in different companies. Looking around the IoT universe of 2021 you will see a plethora of small and medium size companies that specialize in one part of this chain.

To facilitate this cooperation, interfaces are being standardized. Some of those standards have existed for 50 years, like Modbus, and some were created more recently. With these standards the integration can be accelerated. This is especially noticeable in the cloud where the RESTful API is now a commonly used interface that enables different systems to connect quickly, securely and reliably. The ecosystem approach takes advantage of these developments in standardization.

These developments, and other technological advancements, have helped IoT but overall do not eliminate the structural challenges that underly the slow uptake of IoT. To really help IoT take off, a paradigm shift is needed. There are some signs that this is starting to happen. The last section of this article explains this phenomenon, and why it is important to embrace it completely.

Focus on IoT’s weakest link …

The traditional subdivision of IoT solutions as described in the previous section leave the weakest link of the IoT chain exposed. This is the link between the device in the field and the back end or application in the cloud. This is often a wireless connection, where data travels over large distances and through systems that are operated by third parties. In that link problems can occur – from package loss and poor signal quality, to security compromise and complete network access elimination due to expiring roaming agreements. There are protocol, timing, addressing and other technical challenges. In addition to transferring of data (which is of course what it is all about) this link needs to be able to transfer complete new firmware versions “over-the-air”, patches and encryption certificates.

Because of its sensitivity, the implementation of all these technical elements takes time, careful coordination between the firmware team of the device and the software team of the cloud application and extensive testing. It is also the link in the entire IoT chain where the standardization has progressed the least. Structuring the development such that the two teams are in different companies, as is currently often done, makes this process even harder. The development of all the functionalities that are needed in this link, testing and hardening them, such that the IoT solution is ready for roll-out and scaling, takes experienced development teams more than a year. This contrasts the linking of the device to the asset (for example with Modbus, CANopen, BLE, I2C, UART or many other standard protocols) and the cloud application to enterprise systems or front ends (with API’s or webhooks), which can often be realized in hours or days.

It is key for the success of IoT that the paradigm shifts. Rather than breaking the IoT solution between the field device and the cloud, the sub-systems need to be split between de field device and the asset on one end and an “IoT application” and “back-end applications” on the other, in the cloud. The field device and the IoT application together form an IoT-connector. The IoT application is completely integrated with the field device and manages all the functions of that are needed in the IoT link between the device and the cloud. The IoT-connector is a building block of the IoT solution. It can be connected to the asset on one end and to the back-end applications or enterprise system on the other. Both these connections use standard interfaces with standard protocols. BACE is an example of such an IoT-connector, with an embedded device on one end and an IoT application on the other.

Other emerging solutions of this problem are driven by IoT platform providers, who make SDK’s or agents available – applications that developers can place on their IoT device through which the device is automatically integrated with the cloud. It extends their platform to the edge. This does speed up the development process significantly, but works better for Linux devices than for embedded devices and results in a bolt-on solution with weaknesses and limitations.

… and why that matters

Having an IoT building block that creates a connection between the asset in the field and the enterprise system in the cloud, that is fully hardened, secure, already available at scale, future proof, with worldwide coverage and with an IoT portal that enables managing the IoT operations, resolves or reduces many of the challenges of IoT. Examples of these advantages are:

• It is now possible to create a first version of an IoT solution in weeks, instead of a year or longer. This version can be used for feasibility tests or pilots, for demo’s and for gathering support for the project. As the building block is already fully tested and operational, the feasibility can focus on the rest of the IoT solution, where the uncertainties and risks are usually much less significant.
• This first version of the IoT solution will start collecting data rapidly. This means that the first tangible results will be in reach quickly, allowing the team to assess the added value early in the process.

There are many more advantages but, in its core, with this building block the project team is able to focus on its own area of expertise. Eliminating the distractions that come with developing the IoT link between the asset and the cloud helps speed up the project, simplifies it, offers instantaneous scaling and increases the likelihood of success.

The next time you consider an IoT solution and want to “kick the tires and take it for a spin”, to get a good feel for its potential, do it with an IoT-connector like BACE.