Software-defined automation is an evolving strategy to decouple software from hardware to enable smarter virtualization, containerization and truly modern IT architectures. Manufacturers that embrace this approach gain greater flexibility, and can streamline upgrades while accelerating innovation initiatives. 

But they have to understand how to do it first.  

During this Automation World podcast we chat with Stefan Basenach, head of process automation technology at ABB, to discuss what software-defined automation really means, why it's gaining traction and how it might just reshape the future of industrial operations.  

This podcast is sponsored by ABB.

See below for full transcript:

Chris McNamara: Welcome to the Automation World Gets Your Questions Answered podcast, where we connect with industry experts to get the answers you need about industrial automation technologies. My name is Chris McNamara, Automation Group Market Content Director, and today's podcast is about the benefits of software-defined automation as enabled by some core architectural principles. By separating software from hardware and adopting open, portable, and secure technologies, organizations can gain greater flexibility and reduce lifecycle constraints. They can have stronger cybersecurity and a more agile path to modernization. So, we're going to dive into all those details. Our podcast today is sponsored by ABB, and our guest is Stefan Basenach. Stefan is Head of Process Automation Technology with ABB. Welcome to the podcast, Stefan.

Stefan Basenach: Hello, Chris, how are you doing?

Chris McNamara: Wonderful. Thanks for joining us. Tell us a little bit about your role at ABB and how you work with software-defined automation initiatives.

Stefan Basenach: My role in ABB is running the development of our process automation products globally. We have several products in the market and we feel responsible for keeping those systems alive and at the same time driving innovation into those products. I am here since 93, but they will be in different roles. So I have a broad portfolio of experience working mostly in project and now giving this project experience also in the R&D ecosystem to make sure we deliver and produce and develop the right products for the market.

Chris McNamara: Wonderful. So today we're looking at software-defined automation. This is a very hot topic in our space, but it's a very broad topic. So I'd love to get your definition of software-defined automation. What does it mean to you?

Stefan Basenach: Yeah, thanks for framing it like that, because I think that's the strength and the weakness of that term. It's used very often, but very often in a totally different context. So putting the definition at the beginning, I think it's absolutely mandatory. In my view, the headline, the umbrella is the frame is the decoupling software and hardware. So that's, I think, the common term. But we clearly differentiate software-defined automation on the core control level, where we talked about embedded controllers, industrial PCs, connection to IO points, where software-defined automation is totally different when you go compared to the operation system. So the machines where the operator is sitting in front of the MES systems or digital application on top of that. There's also digital software-defined automation happening, but in a totally different way. So on the lower level, virtualization is the name of the game for many for many providers of that solutions of the automation solution, and virtualization can mean you run the application on any hardware. We are still a strong believer that the time of an embedded controller is not fully gone yet for the process automation piece, so we do virtualization, but we keep specific IO modules, we keep embedded controller as part of our portfolio. On the operator side, more the software side, there we clearly take a step away from Windows-based systems, with Windows Server farms, where you run the software in a more monolithic towards the Kubernetes architecture, where we want to make use from the modern IT architectures, software containers, easy to deploy, etc. So those two worlds, both under the headline of software-defined automation.

Chris McNamara: Okay, and you just touched on this a little bit, but I want to dive further into it. How has software-defined automation become the current preferred strategy for monitoring manufacturers, and why is that? What's the motivation?

Stefan Basenach: I think the motivation is in process automation. Bottom line is they want to get away from this large migration jobs every five to six years, which was very often driven by absolute sense of the latest server feature or server release of Microsoft, and that forced you to move to the next step. And when you move to the next step, you wanted to innovate, you wanted to get hold of the latest features, but it always meant a shutdown, a time where you need to potentially do re-engineering, etc. So that's the visible, most visible, and I think that's the biggest difference to the factory automation. The most visible bottleneck of those old times. And with the software-defined automation, we get this lifecycle flexibility where innovation is not waiting for the next big release. Innovation is rather deployment of new software packages at the time when you need it. And having this functionality, you can also offer software on-demand, which wasn't possible before. So let me give you a super simple example. When you have an operator station, a client where the operator is sitting in front of, running on HTML, using a software-defined backend as a server functionality, and you're running into turnaround where you will need 50 to 60 more of those thin clients. In the past, you had to install those thin clients, download the software, buy the hardware, specific hardware, because it's a pretty software. In future, you just buy this 50 additional licenses for two months, You use it and then you give it back. So this kind of flexibility, especially in turnaround situation, I think it's a huge benefit for software-defined automation.

Chris McNamara: Excellent. So flexibility is one of the key selling points. Let's talk about some other benefits, you know, these emerging approaches with software-defined automation are enabling wins in terms of what else? Outside of flexibility, greater efficiency, what?

Stefan Basenach: Yeah, you get lower operational cost. You can innovate when you want, which makes, again, the plant you're running more competitive at the end. And this is important for any technology. Software-defined automation features will be successful if they make the plant of our customers more competitive. If they are able to reduce the lot size, if they're able to change faster recipes in a batch process, if they're able to run the plant on different loads, full loads, medium load, or low load of the plant. So if the customer is able to earn more money or reduce the cost, then those features will be used. So we always try to look from the customer value perspective and far less from the technical features, yeah, you can gain with the technology, yeah.

Chris McNamara: Of course. Well, oftentimes the customer doesn't care about the technology. They just care about the outcome.

Stefan Basenach: True, true. And still also in many of the discussions we do with ourselves, in my view, we still focus too much on the feature set. We need to get better in translating this feature set to the use cases and functionality our customers are asking for.

Chris McNamara: Yeah. Sure. Well, you're not alone in that respect. That's a very common focus for solution providers. Let's talk about an early win. We just talked about some of the capabilities. I always like to entice people to explore new concepts with the earliest, quickest win, the low-hanging fruit. What's an early win that you can achieve embracing software-defined automation?

Stefan Basenach: Yeah. An early win is when you embrace that solution that you get a better flexibility where to deploy the algorithm, the software. So when your software is independent from the hardware, you can do the engineering very generically. You don't even need to know where it's running. And then based on your architecture or existing functionality, you can deploy that automation at the right place for that moment. This is a very low hanging fruit. Another thing in my view is something we discussed with our customers. Some of our customers have already huge data centers. So the hardware already exists, it's maintained, it's paid, it's running. So we have the flexibility to run some of our software in those data centers of the customers. Before, you always had to buy hardware and software from us. The next thing, very quick win, is deployment of new features. You don't need to wait for the next big release. If there's a new trend functionality, you want to ship it to your plant, you can do that with a software container much, much faster than before. And so those are the small pieces when the architecture is right, you can very easily implement.

Chris McNamara: Okay. Excellent. Let's talk cybersecurity ramifications. Does this approach expose enterprises to greater risk?

Stefan Basenach: The cybersecurity is a big topic and most of our customers are working what we call in Europe critical infrastructure. I think you have NOCSIP or something like that in the US. So any society or country has a different name for it. But the main target is to protect the plants. Architecture is several layers. So we introduced with automation extended what we call separation of concern. So we define a core control piece, which should be by definition pretty stable. So the PID loop doesn't change every night. Because you don't change the process plant. There is an actuator, there's a sensor, there's a control algorithm. So this is the stable part. And here you want to have a very safe, secure environment, encapsulate that. And there, you need to make sure that your system is able to always get the latest patches without disturbing operation. So here's the mantra, really, keep the system stable and be able to deploy patches without disrupting operation.

Chris McNamara: Flexibility even in the cybersecurity realm.

Stefan Basenach: Exactly, and then we have the we have the second layer which is connected to this core control, and we're using here the Namo open architecture concept. This is a kind of controlled communication line. We also call it a republisher, so transporting the data to the second digital environment. And here you have a more flexible architecture, because here is exactly where you want to have fast innovation, where you want to introduce new things fast. But using Kubernetes as a shell, as a platform, gives us the opportunity to keep the security level still on a very high level, using all these IT features and functions we have today. And here I think we with the new cybersecurity threats we have through the to the AI algorithms. I think that's really a very interesting part of the game. And it's a it's like a it's like a race. Who is faster? Sometimes the IT suppliers will be faster, sometimes the bad guys will be faster. So that's very interesting and very much debated in the community. We believe bottom line, this race will make our systems more secure than before, because there's an unprecedented and never seen push for making our systems more reliable, more up to date, and secure as possible. Those are the two layers that are important in our view.

Chris McNamara: Wonderful. Industrial automation is maturing incredibly rapidly and mainstreaming, like snowballing, it seems. How is that maturation of industrial automation changing software-defined automation?

Stefan Basenach: Again, software defined automation for me is not, it's not the solution, it's an enabler. And I would differentiate a little bit this kind of innovation. I think in core control, the innovation is not so big. I think the basic stuff; connect the transmitter, make sure the PID loop is working, start up the plant. There is not a lot of innovation. And I think we see some suppliers saying, okay, let's get rid of this 1131 code. We do that all with AI and Nobody wants to go in my view, no customer I'm speaking to wants to have a black box running the plant. So I think in this part, innovation is really how to protect it, how to make software independent from hardware, but not to innovate like crazy. The innovation happens one layer up where the use of data is making the difference. And also here software defined automation is only a small part, the algorithm is only a small part of the solution. I think the main difference, the main innovation comes from using the data differently than before. Using more data from the field than before. And so the data perspective is, in my view, an even more important part of this innovation part, software defined is the algorithm, data handling, contextualization, semantic layer is the second important part and both together give us the innovation.

Chris McNamara: So who's doing it right? Are there any fields or any verticals or any industries that are front runners with this and what good ideas can we steal from them?

Stefan Basenach: A pattern I always do, I look into factory automation. Factory automation is introducing features very often 5 to 10 years earlier to process automation. So if you look at virtualization of PLCs, they're already doing today massively, introducing AI algorithms on top tons of data they're collecting, connecting the plants with the ERP systems, using AI analytics. So all those things are a little bit ahead. So what I'm seeing is that some of those technology adapted to the process automation needs will find a way into our business as well. When I look at processes, where we'll see innovation the fastest, I think it will start with more smaller systems, food and bev, pharma, these kind of technologies. I don't see it coming super, super fast on chemicals, steam crackers, offshore wind platforms, or oil and gas platforms. I think that's something that's coming a bit later.

Chris McNamara: Okay. Interesting. So we're forecasting a little bit now. Last question for you, Stefan. Where is this headed? How does software-defined automation evolve and who benefits most from this growth? Where are we going?

Stefan Basenach: It will evolve step by step because our customers have invested a lot of money in the automation of their plants and they don't want to rip and replace. So I think we as supplier, and that was always our mantra, innovation with continuity, is we want to build a bridge from today to the future. We have done that the last 40 years and we will continue to do so. And as I said before, the value is the important piece. The same for AI. Customers will spend money if they see how their plant is running better than before. The feature alone means nothing for our customers.

Chris McNamara: Yeah, yeah, the outcome.

Stefan Basenach: And if you take that serious, it means who will benefit the most -- hopefully our customers. And I think from the supplier side, The company that has the best migration towards the new future, I think, will be going out as a winner out of that race.

Chris McNamara: Wonderful. Stefan Basenach, thank you for joining us today on the podcast.

Stefan Basenach: Thank you very much.

Chris McNamara: We'd like to thank our sponsor, ABB. Very interesting stuff. This Automation World Podcast was produced by Endeavor Business Media, a division of EndeavorB2B. I would want to thank our loyal listeners. Be sure to subscribe if you're not following already and follow us in your favorite podcast app. And as always, you can find us at Automationworld.com, where we continue to explore this world of automation.