
Photo: rozdemir – shutterstock.com
If the pandemic has brought something positive with it, then it is a digitalization boost through completely new application scenarios. The past few months have influenced the IoT market in such a way that it is now developing in two directions: While existing customers continue to plan and pursue their longer-term projects, the use cases are being pushed with new projects, with which costs can be saved at short notice. However, such quick wins are difficult to implement in an industrial environment, because IoT projects span a period of ten years and more. The reason for this is simple: a smart factory does not grow in the green, but almost always develops out of the brownfield. This mix of old machines with proprietary legacy interfaces and modern systems that can be connected via MQTT or OPC UA can always be seen in the long term. The pandemic will not change that.
Many of those who pursue a strategic approach in their IoT projects have devoted themselves intensively to the topic of connectivity in the past few weeks, in order to be able to emerge from the crisis stronger – albeit with a bit of a delay, since many companies are or have been working short-time – a special route with which Germany seems to have put itself on a leg in an international comparison: Especially now, when many things can be changed in the manufacturing and software area because the belts stand still, there is a lack of human capacity. So while everything in the USA or China is just going on as before, here in Germany you are faced with the challenge of continuing to innovate, maybe starting one or the other project a little faster and discarding it if it doesn’t work. With the joy of experimentation, we are simply lagging behind in Germany, which poses major problems for us in long-term Industry 4.0 projects. Can a large part of this perhaps be solved by using IoT platforms or an IoT out-of-the-box?
- Jürgen Hahnrath, Cisco
Finding use cases that can be implemented quickly is absolutely sensible, because they show both the potential and the limits of the concepts. With the limits, it is very important to find out how to make the whole thing safe and how to scale it in bulk. Connecting thousands of machines is no longer a manual process. And so the cloud automatically plays a role when considering automation, scaling and the question of what data is needed where. And that means that in addition to an overall architecture, you also need an operating concept for it. - Kai Waehner, Confluent
A manufacturing company is not just about connecting data from an OT perspective, but understanding it as a continuous stream of events. These usually have to be combined, correlated and processed continuously – often in real time. The sources can come from ERP or MES systems in the factory, as well as from the CRM system (which may run in the cloud) or in the data warehouse. But even in an ideal world with open standards, you will still have different standards and communication paradigms. This will continue to pose challenges for the industry in the future, because there will never be a platform or solution for it, but it will always be a combination of partners and different types of solutions – both on the OT side and on the IT and Software level. - Dr. Nicole Göckel, CONTACT Software
My best practice tip is: start small, then roll out. It is very important that you already have a clear vision of what goal you are actually pursuing. Digitizing because everyone does it does not make sense, there must be a business idea behind it. For example, that you want to improve production, reduce downtimes, plan maintenance segments or even make the product smart in order to be able to offer customers added value with digital services. But you shouldn’t do everything at once, but start with a project, observe the development and then take the next step. - Marten Schirge, Device Insight
My urgent advice to all companies that have not yet dealt with the topic of IoT: Act now! They are actually already too late, which is why you should definitely invest in corresponding projects now. They should focus on the one hand on the use cases that generate added value for the company and on the other hand on their own core competencies. Because until every German medium-sized company has tried to set up its own IoT and software development team, there is already too much time in the country. It is much more important not to miss picking up the entire organization, otherwise it will always be just individual projects. - Florian Doerr, HPE
I recommend pragmatically getting into IoT projects because the classic approach, planning three years and rolling out three years, doesn’t work in the IoT. Already on the way to planning, changes and new insights can be reached that influence the scope. You should definitely have a big picture, but then start pragmatically quickly and use an agile approach to ensure flexibility in the projects, which also makes it possible to adjust where something did not work. - Sven Koltermann, Telefónica
It is important that you basically start with digitization. With finished IoT solutions, you can already make your first attempts. But for companies that have not yet dealt intensively with the Internet of Things, it is advisable to get external support. Through targeted advice, it can be determined in which areas a company can become smarter and how IoT applications can be integrated step by step into their own systems. - Johannes Kaumanns, Telekom
Successful implementation of IoT projects presents many companies with complex challenges. The main reason is the multitude of heterogeneous ecosystems – with a wide range of different devices, protocols, networks, clouds and applications. This multitude of possibilities makes the compilation of the right IoT solution as well as its provision and operation a mammoth task. What companies need are people in IoT projects who have a shining eye because they recognize the potential of IoT. To be successful, they need an IoT hub with which they can connect the heterogeneous systems with one another and create unlimited possibilities for the best innovative IoT solutions.
The turnkey IoT out-of-the-box variant is well suited for individual use cases, start-ups or companies that want to gain initial experience in the IoT environment. But one thing should be considered: “out-of-the-box” suggests that it can implement a use case using a next-next-finish approach. And that is not the case! This may work in a closed environment, but when it comes to scaling or a rollout, the ecosystem is too complex. Scaling, operation across multiple production lines at different locations and data sovereignty should therefore be considered from the start. In addition, there is the guarantee of reliability, availability at all loca
tions, data security and the consistent configuration of the customer solution – all of this is not trivial and cannot simply be made available out-of-the-box. Another point, why there are difficulties in the box in addition to a solution, is already in the name “IoT”: It is also about things that can not only be done in the cloud. The factory, the consumer cell phone or – for reasons of cost and latency – data processing will always be located in the Edge. Accordingly, there cannot be one solution for everything, which is why partnerships are extremely important to combine all of this.
There is no general answer to whether a platform is the better approach. If, for example, the cloud solutions have already been set by the company, there is no longer the option of dealing with the selection of platform providers – a major disadvantage, because dealing with a platform and the technology requires that you already have one has dealt with the added value and the use cases. Therefore, the first step should always be to clarify why you want to implement these projects and with whom.
Information on the partner packages of the study ‘Internet of Things 2021’
One of the central points in the IoT is that the customer is in control of his data and his decision about how the data flows. From a strategic perspective, every company must see it that way so that it does not experience a vendor lock-in and remains in a position to change providers. This is where hyperscalers have to rethink so that the checkpoint lies with the customer. As long as the arm wrestling takes place, the customer is always the victim in the end. If you want to prevent lock-in, the full range of Hyperscalers is not available. Certain SaaS products or options that enable added value such as more cost-effective operation have to be dispensed with. In practice, it has been shown time and again that customers who enter into a lock-in are more likely to do so with the large providers because they are already bound to them in part by their IT infrastructure.
Open interfaces or container technology already make it possible to create a certain degree of independence via the clouds. Nevertheless, the big hyperscaler will always be part of an architecture, because from a commercial point of view it doesn’t make sense to do some things yourself. However, the core of a strategy should be to make business-critical applications more platform-independent – for example in the form of using technologies that can be used to move functionality and logic from one platform to another. It is undisputed that system integrators will become more important in the future. They keep an overview of the hardware and software market, in which players are becoming more and more diverse, and provide advice on the choice of partners with their industry know-how.

Photo: panuwat phimpha – shutterstock.com
Study “IoT 2021”: You can still participate! |
COMPUTERWOCHE is currently conducting a multi-client study among IT decision-makers on the Internet of Things. If you have any questions about this study or if you want to become a partner, Ms. Regina Hermann (rhermann@idgbusiness.de, phone: 089 36086 384), Mr. René Krießan (rkriessan@idg.de, phone: 089 36086 322) and Mr. Bastian Wehner will help you (bwehner@idg.de, phone: 089 36086 169) gladly further. Information on the Internet of Things study can also be downloaded here (PDF). |
The experts disagree which role GAIA-X will play in this future. For some, it has the same effect as short-time work: While China and the United States are already implementing everything, Europe is delayed by many years in trying to regain data sovereignty. Cloud providers already offer enough options to encrypt data – in such a way that the encryption is implemented by the end user and the cloud provider itself has no way of decrypting the data.

For the others, the European project is pointing in the right direction, as it is not just about encrypting data, but about a distributed network in which every cell phone or computer can be a GAIA-X node with the capacity where Services are running, expanded – a different approach to the Internet than the central approach that we now have. However, what we now have is enough expertise to implement data sovereignty without having to wait for the GAIA-X architecture. Because if GAIA-X is to be seen as another offer for the top of the pyramid in the sense of cloud connectivity, then its architecture must also enable independence from cloud services. In addition, the project will only be able to assert itself if it is also competitive in other dimensions such as costs and functionality.
Cost is one of the main reasons why data processing will continue to take place in the Edge. Sending all the terabytes of data created in a smart factory to the cloud not only costs a lot of money, it doesn’t bring added value everywhere. The second reason is latency: it takes a few milliseconds for the data to arrive in the cloud – too much if you want to use it in real time for quality assurance or predictive maintenance. And finally, security is an argument for not giving data to the cloud. In the manufacturing sector in particular, many people guard their quality data because they belong to their trade secret.
If, for these reasons, the IoT environment will always remain hybrid, do you need additional interfaces or are the previously available ones sufficient? MQTT and OPC UA will certainly become the standard in the industrial environment at some point. The challenge in the next few years will be to connect the fieldbus systems of the brownfield machines, to tap the essential data and then to transfer them into a kind of data hub so that, for example, a data scientist can use them sensibly.
Interfaces at the IT / OT corner are just as indispensable as a shift in the mindset of those who design such systems. You don’t necessarily need more technical APIs, you just have to use the existing ones more often. Thinking more about data sharing when it comes to business models is also recommended. If you build the IoT solution just for yourself, there is no way to perceive the added value that goes beyond it. (Hi)
Information on
the partner packages of the study ‘Internet of Things 2021’