In episode 25, we explore the world of the Data-Driven Factory with Jouko Koskinen, CTO at Fujitsu Finland to find out how this model is reshaping digital transformation in the industrial sector and beyond! 📈🏭
Jouko is noteworthy for his extensive experience in business development to drive change, having over two decades of involvement, between IT and Manufacturing industries.
Fujitsu Finland is part of the leading Technology group Fujitsu pioneering transformative digital products and IT Solutions globally.
We begin the episode, breaking down what we mean by the Data-Driven Factory and exploring the operations, benefits and potential challenges. We then move on to how Data Harmonisation is revolutionising digital transformation, what this means and the processes involved. Finishing the episode, we talk about security, the role Operation Technologies play within securing industrial networks and why OT technologies need to be considered, as well as, the most transformative predictions for IIoT.
Tune in and discover the secrets to the Data-Driven Factory:
- Jouko’s background in IIoT 01:10-02:52
- What Data-Driven manufacturing means for the smart Factory 02:52- 05:11
- The benefits and challenges of Data-Driven Factory operations 05:11- 07:34
- What is Data Harmonisation and how it is revolutionising digital transformation? 07:34- 11:00
- How much of a role does Machine Learning play in Data Harmonisation? 11:00- 12:33
- Operation Technologies importance in securing today’s industrial networks 12:33- 15:19
- The biggest IoT trends set to change the industrial landscape 15:19- 22:24
Welcome to the IoT Podcast Show. Today, I’m joined by Jouko Koskinen. Jouko is the CTO of Fujitsu Finland. Jouko is noteworthy for his extensive experience in business development to drive change and having over two decades of involvement between IT and manufacturing industries. Jouko, thank you so much for joining the podcast show.
Thank you Tom. Thank you very much for having me here.
You’re very welcome. You’re very welcome. And for those that don’t know, this image behind Jouko, we’ve spoken in the past, is a beautiful image of a lake that’s near your house. Is that right?
Yes, correct. I’m on a Lake. I have a sauna on our lakeside there. All the Fins has one, you know.
<h2>What is your background in IIoT?</h2>
Yeah. It’s really beautiful. Maybe you could just start by just introducing yourself and your role at Fujitsu. How you got to this position and your interest in IoT, I suppose?
Yeah. Thank you. Yeah, I’m a CTO for private sector here in Finland, Fujitsu. And it means that I am responsible for developing solutions and concepts for our private sector customers based on our Fujitsu IP and our partner, technology and solutions.
Pretty simple. But we are focusing very much on smart manufacturing, because we strongly believe that every enterprise, goodness or badness are born at the factory floor level, and they are based to combine operational technology and IT technology, which I would like to say it is all a technology.
My history is that, say the past two decades I have worked a half in manufacturing and processing industry in various places and various countries. And second half of that in ICT. Starting with Nokia Networks and then through three, four major IT players in Finland and also I have spent two years in Russia. And as a hobby, I’m studying based on a knowledge management service and dynamic capabilities theories about digital manufacturing as I am BSD candidate. We’ll see what happens in future.
<h2>What does the Data-Driven Factory mean?</h2>
I’m sure positive things. Thank you very much for that Jouko. Manufacturing is a rapidly growing industry in IoT solutions. We hear a lot about IoT solutions in manufacturing. I’m interested to explain what Fujitsu in your area is doing behind data-driven manufacturing and the operations behind that. Could you explain a little bit about some of the solutions that you are putting together for some of your partners?
Gladly. The history is that every factory is more or less like a silo-driven, because in past, based on ISA 95 protocol or standards, they were very higher to put up, and in past there were no need for cybersecurity or digitalization. It’s just focus on availability and products and performance.
This is the current situation. And now, many companies have realized that while enterprise and end-user services are digitalized and everybody’s pressing about those, and what should we do for factories? So, that is the thing that we need to transfer current-old fashioned factories to be modern microservice-capable factories.
So we are focusing here at Fujitsu that how to make this happen, how we can break down the silos and enable factories be effective and data-driven and real-time driven. But the whole basis, the quality of data. And that’s why we are focusing very much on data management, harmonizing the data models, which is space for digital gains, and also how to collect data, how to store this factory architecture. And that we are focusing and helping.
And one very good principle is that we need to have for every customer, certain used case, what they are aiming for against what we are taking this old landscape and turn it into a new one.
<h2>What are the benefits and challenges of Data-Driven?</h2>
Sure, sure. Thank you very much for that. In terms of data-driven factory operations, this is a fairly new term to some of our listeners and people that are subscribed to the podcast. Could you explain, and just go into a bit more depth if you wouldn’t mind Jouko, about the biggest benefits and potential challenges of data-driven and factory operations?
Yeah. The challenge is mostly in people’s heads, mental-driven challenges, that we need to learn from our old-fashioned way of thinking and so. That is the one bit. Because this is always people business, and we need to get the business cases behind our decisions. So we shouldn’t digitalize because of digitalization. We should digitalize for business benefits, and against used cases. So, that is more or less.
But there is challenge that we need to have this data quality to be correct and harmonized and available. Because currently, data is not always available. We may need add on certain additional sensors or something in order to get access to the data.
But, even though, you would have a massive amount of data, it’s useless, unless you know where to use it. And that’s why we need to have those used cases. It means what’s the model the factory data sources, realize are they available or not, what we ought to do, how to bridge stalls together. So, that is were challenges there, but they are very much solvable.
Yeah. I think that’s clear to see, in a lot of conversations that we’ve had recently with people, is the shift from the traditional factory and the traditional way of doing things and the mental challenge that they have in bringing this online to a more connected and data-driven model.
And it’s interesting, you should say that actually, because I think sometimes, we’re our own worst enemy in the onward steps that we can make towards this and the digitalization of this.
<h2>What is Data Harmonisation?</h2>
Tom White: Something that you mentioned actually, just while you’re explaining that Jouko, was data harmonization. Could you talk a little bit more about data harmonization, the process and how it’s revolutionizing digital transformation, and specifically within Fujitsu, how you’re using this and implementing this model?
Yeah. Okay. All right. So starting point is these factories used case. For example, if they have a used case to have a better troubleshooting, if they have a warrant claim, how to drill down and see what causes troubleshoot.
Then we have that, and then we have a product, then we start to model with unified machine language, that case, and that way of modeling contains a lot of technical data, which are relevant to the product and in each section of productions.
In this way, we have value chain and data chain with technical values, dependencies between one to one or one to many. And then we have a model, how the production goes, and how the product is born. This is in translated with the digital tools to the data warehouse to be modeled and then collecting data’s instructional manner.
And then identify what data is needed, and then we take only the needed data into a warehouse and all the others close to the data lake. And we, at Fujitsu, would recommend that we have a data lake already on factory level and data warehouse and presentation layer, which after the software-driven van is the van.
We transfer internet or MPLS to the cloud, in cybersecurity manner to the enterprise data lake and for machine learning capabilities and environment.
And this is basically a very simple architecture and how to go there, is that we go to the customer, we asked, what are your challenges, used cases, if they don’t know yet, we advised them to build one.
And which after that, what is your current landscape, what technology you have in use, what can we reuse, then what kind of preferences you have or AWS, Microsoft, or wherever, if at all, or open source. I have dental building blocks. At that forties used case, you need to have these and these functionalities, starting from factory floor level zero, because there is this level zero, one, two, three, four, which after goes to the enterprise and so on. So, this is like the technological stack.
In this way we built the use case based on that framework of which we have modeled with all those leading technology vendors plus open source methodologists, and it’s like a Lego system, modular and scalable. But always against the piece that’s benefit. And it’s very a nice challenge.
<h2>Is Artificial Intelligence and Machine Learning integral to Data Harmonisation?</h2>
Yeah, I can see, I think the analogy of it being a Lego is set right, is great. Because it’s modular, isn’t it? So you can just add things together and there’s varying aspects as particularly the machine learning. I mean, it’s interesting you say the machine learning. How much does that play in data harmonization machine learning? We see a lot of businesses looking at machine learning or the advancement of AI. Is that quite integral to the data harmonization process?
It is like a one a module on the top of it. Because once we have a data management platform for the company and factory, on top of it, you can pull to all kinds of applications, and machine learning is a part of it. I see the great need for machine learning already on lowest level for quality management, because we have at the Fujitsu, a computer visual inspection so whatever we can turn to the picture, we can analyze and detect failures there.
And immediately, with our infrastructure and architects we can turn it through PLC back to the production and say that [inaudible 00:12:00] detection, quality is not good or bad. So, that is one. But then, also I see the process optimization, great need for that. First of the data science may, as a person evaluate the traits or values for good or badness, which after we can predict, but once those are manually set, later machine learning can start to automatically adjust it to a special value of use. That is very much doable and applied already.
<h2>How can Operation Technologies be used to enhance security in Factories?</h2>
Okay. Fantastic. One of the things that I wanted to learn about today, for me personally, with OT security, it’s becoming widely implemented across the industrial landscape at the moment. Could you talk a little bit about operation technologies and why it’s important for security in today’s industrial networks?
Absolutely. Currently, we have faced already many activities around the factories. One was in [inaudible 00:12:58] countries, one steel mill. They lost the production capabilities because of hacking, and based on public known information it cost 75 or plus million for them to do all, overall. There are many others in worldwide.
So, it’s always people process technology, what you need to consider wherever you have a security issue, because your own employee can be one big threat, but anyway you need to go… As we have now during the COVID time, we have done all the cybersecurity assessments, very successfully. And there is lots to do. I’m afraid to say. But the idea is that we need to walk through the cases. There is certain do model, how we can layer it and segment the factories, and then we define who can talk to whom and how to monitor that systems and others.
So this is the very key in Fujitsu’s operations. And we also always want to say that whatever we are doing, we need to do it cybersecurity manner. So we have services for assessment. And then the transformation, because old factory turned to the new type of factories, and cybersecurity manner is long process because you cannot shut down the factories wherever you wish. There is a certain time per year when it is shut down, and then you need to apply money from management to do something. That’s why we do the assessment of service to do over a long time or fix them fast whatever is needed, and then to operate it for them.
So it is the growing issue and not that easy to solve. But with good technology vendors, what we have a within Fujitsu, we do close cooperation and we have a now under development, very new way of thinking all of the cybersecurity solutions. So hopefully, to say something more, end of a year or something.
<h2>Where will IIoT be in the coming years?</h2>
Yeah. Okay, fantastic. We’ll be looking out for that. And Jouko, we often ask people in positions, such as yours, really the biggest trends that they wish to see moving forward and that they think is going to change the landscape. And, I’d like to know your insights, but specifically within industrial internet of things. Where do you think we’re going to be in the next few years? How transformative can this be from a CTO standpoint? What do you expect us to see in the next five years?
Well, if I say that this year ’21, is the year when the variance is more and more there. Certain proof of concepts are done, and people are thinking as a company is meeting, that what is our digital strategy enterprise, but also for factories.
Big issues are, because many customers, at least here in Finland, big customers, they have SAP as an ERP. And there is to come S/4HANA, S/4, which means that this is exactly the good moment to think about what is based on capabilities, the role of ERP future than it is today, because of we need to have a dynamic capabilities continuously updating change.
We need to take the complexity away from ERP to watch the cloud-based solutions in order to have the capabilities for continuous development. That is one thing there, and so on. So it is then how the factories are linked. Because too often, people are developing enterprise level IT and end customers and forget what are the factories. I see that now this is the learning curve, realizing planning, doing some proof of concept ’22, ’23, it start to grow this cybersecurity.
Data management platforms becomes more and more common because they enabling your quality operations. And then, factories start to gradually go for predictive maintenance more and more. And that is the next thing to come. We have talked about it for many years already, but still not many implemented.
Why is that?
I believe it is too much. Enterprises have focused for IT and end-user facing solutions, and forgotten this factory. But now, it’s a year for factories, let’s say call it… Nominate that this is a year of factories. So it will come, and somebody has done it already or always but major part of the companies has not. That’s why it is still to come. And I see a great need that this quality management is one to go and come. Because of at the factory, we need to secure the availability.
Whatever helps companies, factories, secure availability is key. That’s why cybersecurity is one. That’s why the continuous operation is another. That their production is performing. Availability, performance, quality. That is overall equipment efficiency.
So I wouldn’t say that at least in coming two years, nothing… It’s data management platforms, priority maintenance. But then four or five years time, AI is more and more there, because we have now first two to three years learned about the digital. We have a data management platform, which is foundation. We have some operations there, which after is… Not to rally. We can put some AI more and more. So they need to spend that. Maybe I didn’t explain it clearly enough, but or [crosstalk 00:19:33] whatever, but…
No. I think you explained really well. Thank you. Yeah. I think it’s interesting to know it because in simple terms, if you ask people involved in IoT, what they predict to happen in the future with factories and various elements of industry, you can often draw a mean from this and then start to look at trends and pattern analysis. And I think it’s very interesting what you’ve mentioned, and also the aspect of perhaps people knowing about this, but not yet really implementing it because it’s not been a priority. They’ve looked at other priorities. It’d be really useful to see what happens in the future when dumped. And no doubt, you’ll be there. You’ll be involved somehow you cope, right?
Yeah, absolutely. But also I would like to emphasize that, one is what is kind of holding it back that in many companies, factory directors are somehow too strong persons and too influencers. And that’s why maybe that companies, once they get the business and factory managers come more and more same thing where this is to make it better.
Now factories they said, we cannot do this because they are afraid availability or performance will go down. It’s like in every development, that we believe that it should happen tomorrow, but people are slow movers. And if you need to guarantee sales and profit, that holds you back to do radical changes. So many factors to be considered, but it is to come. We are already on the way. And I see great positive things to happen.
Fantastic. Well, I think that’s a great point to end the show today, Jouko, but I really appreciate your time. Thank you for coming on and sharing your insights, specifically from your standpoint, the CTO at Fujitsu, Finland, where you’ve been, where you’re going within IoT. It’s really insightful, and I’m sure all the listeners have really enjoyed this episode. So thank you so much for your time.
Thank you, Tom. And we strongly believe that now is the time to move, and we are here to help.
Okay. And I think that’s a perfect closing statement. Thank you. Thank you, Youko.
Thank you. Bye, bye