Industrial IoT (IIoT) vs IoT for industry (IoT4i)

You heard these terms, increasingly used, that are IoT (Internet of Things) and IIoT (Industrial IoT). Terms that seem so close that we tend to think that they are intimately linked, or even that they address the same thing but in a different context, which for the second would be that of industry. But the technicians, or marketers, are teasing or, because they surf the trends, it tends to generate misleading shortcuts.

You understand, behind my words, that there is no more than a fairly distant cousinship between IoT and IIoT. I will try to explain why and what are the consequences of these differences on the technologies used in each of these areas.

What is the Internet of Things in Industry (IoT4i)?

Let’s start with the Internet of Things (IoT), which the usual translation “Connected Objects” does not allow to understand correctly. Indeed, this term gives the illusion of an object to which we would add connectivity, but for what purpose? to speak to whom? The literal translation, which illustrates the purpose much better – “Internet of Things” – invites to connect objects together to generate collective intelligence, interactions, immediacy, proximity, as the Internet allows between humans.

The concept of IoT appeared first, with the desire to capture new data, from the environment, physical data with the aim of integrating them into digital processing chains and thus enriching the digital transformation with new information from of our real world. Indeed, our digital transformation is poor in real-world data, which handicaps many changes in business processes. Let’s take just one example in CRM: how do you maintain your relationship with a customer when he no longer opens your emails or does not answer calls from your call center, which has become the norm? The answer comes from the product itself with which the consumer has a continuous interaction, if he communicates, he maintains this relationship for life. Isn’t that what google, amazon, apple have succeeded perfectly with smartphones, or voice assistants?

The Internet of Things gets along with a massive approach, it is a question of collecting a very large volume of data, of a similar nature, geographically very distributed, with the aim of extracting a stronger value from it than unitary information. . It is a question of extracting from the mass of data behaviors, classifications, patterns capable of piloting decisions and processes.

IoT data is therefore voluminous data overall, but individually very small, without great diversity and rather captured at very low frequency, because the IoT involves a very strong constraint as to its energy consumption, linked to the impossibility of maintaining, over a very wide area a large fleet of objects.

How is Industrial IoT different?

On the other hand, the Industrial Internet of Things is quite different. If it is also a question of data whose purpose will once again be the management of decisions and industrial processes, the means to achieve this are different. This is about improving the manufacturing process through good use of data, making it more efficient, greener, more economical.

To do this, it is a question of collecting data from the production chain, this data is therefore highly localized in the same unit of places, in a controlled context and accessible to maintenance where the number of machines involved (the “objects” of the IIoT) are ultimately few in number. The very positive point of the IIoT is that, despite the fact that this term is more recent than the first, the availability of data is much more mature. Indeed, industry 3.0, which has robotized manufacturing processes, has embedded a large number of sensors and generated a very large volume of data. Industry 4.0 consists, among other things, in exploiting this data more intelligently, what is called “extracting value from it”.

We therefore have, to sum up, in the IIoT, a typology of data that comes from few objects, with great diversity, a very large volume, a high frequency and captured in a unit of location.

Two very different worlds…

You see we basically have an opposition between IoT and IIoT:

  • a massive number of objects against few objects
  • a data unity against a diversity of data
  • a geographic disparity versus a geographic unit
  • low frequency versus high frequency

…Which are however totally complementary

Two things bring these two fields together: the first is that despite all these oppositions, it is fundamentally a question of processing massive data, big data and that the extraction of value will often go through the use of Artificial Intelligence type technology. The second is that the IoT is, nevertheless, a complement to the IIoT because it makes it possible to enrich the data resulting from the manufacturing process with complementary environmental data which today is only slightly integrated, natively, in the machines of production.

We therefore come to the understanding of the technological issues relating to the processing of this data by what is called an IoT platform in one case and an IIoT platform in the other case.

The IoT market has, as we have seen, been pioneering, so many solution providers have implemented solutions capable of collecting IoT data. So data from very many objects whose fleet management is critical, which emit homogeneous data at low frequency but which end up representing a large volume of data that we will access to the object or to the group of objects for individual or aggregate analysis. IoT platforms will therefore favor simple but scalable integration; this brick will be critical because we must not lose the data that is not stored in the objects. They will not seek to create a correlation link between the data because they are homogeneous. The notion of asset (sensor) management will be strong, whether on the model or at the instance because the management of the object fleet is critical but also because it carries the key processes of manufacturing, deployment, maintenance of the large fleet of objects, the management of which is the critical element of the project.

The arrival of the IIoT market, as I said earlier, more mature, has been a pivotal opportunity for IoT players as it continues to lag behind the growth forecasts. Many IIoT players today have therefore simply rebranded their IoT platform but without changing the fundamentals, taking advantage of the confusion effect between IoT and IIoT.

Hence the very specific characteristics of an IIoT platform

However, an IIoT platform meets very different criteria, it must ingest large volumes of data, in real time, process them in the flow, and have the possibility of linking the different types of data received: process the diversity of data and extract value. This is key because the focus is no longer on data integration (which however remains critical), it is on the real-time processing of complex data crossing a model allowing data transformation and linking. This requires advanced development capacity for multiple models and an advanced, scalable representation of data: the same data will address a large number of business use cases, analysis, real time, reporting, optimization, management, etc.

Basically the notion of asset is not central, even if it is useful. The model is not linked to the asset but rather to the manufacturing process because this is what will allow the link between the data. This model is called digital twin, there are multiple digital twins and the one used in manufacturing corresponds to that. I will not go into the details of this concept which may be the subject of the next article in this series. The engine of the IIoT platform is above all an ability to query, in real time, a massive volume of heterogeneous information.

The technical characteristics of an IIoT platform are very different from those of an IoT platform. Provided they are not opposed, an IIoT platform will benefit from being supplemented by a light IoT platform for the integration of complementary IoT data, as we have mentioned. On the other hand, using a platform designed for the IoT in an IIoT use case amounts to missing the heart of the objective which is the valuation of the data and facing difficulties during the integration. This will result in both a much longer implementation time with broader specific developments, but also scalability problems that will lead to extracting data to third-party tools likely to better support growth and whose price will disrupt the business model. Many companies, still in a largely exploratory phase, will only realize the impacts of choosing a platform born from the IoT late in the transformation project. Many projects will end in failure due to too long implementation penalizing the demonstration of value or technical difficulties related to the implementation or modeling unsuited to the specific industrial context.

If it is not the only dimension to take into account in an Industry 4.0 transformation project, the IIoT platform nevertheless remains critical for the success of the project. As we have seen, it must be adapted to the specificities of the IIoT, preferably be native IIoT and not derived from the IoT, then adapted by marketing and at the margins by R&D.

To properly identify whether the solution meets the expectations of an IIoT project, you must focus your attention on how the production process will be modeled. What tools allow modeling? Is it easily scalable? Does it allow interaction between the developer and the business? Is the processing engine scalable? Does it support a wide variety of data? a very large volume? Finally, it will be necessary to look into the capacity of the tool to be usable and modifiable by the profession, by the field: a tool to help with transformation, the IIoT platform must not be a tool for engineers, nor a tool for the central but a tool for the field.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.