 Welcome back to Boston Mass Choose. It's Red Hat Summit 2023 here in Boston. I'm Paul Gillen joined by Rob Screcci. A wall-to-wall coverage of Red Hat Summit and the Ansible Fest, which is integrated this year, all about automation. And today it was all about the Edge. The whole keynote this morning was really focused on the Edge. And one of the speakers was Bernard Eshman, who Chief Technology Officer at the ABB Process Automation. ABB, of course, the industrial automation giant. And Edge was all the talk today. What was your, talk about the nature of your relationship with Red Hat? Yeah, obviously, if we look at the Edge, it's the place where all of the data from the industrial plants that we automate are. And typically the automation system that we develop and that we provide to our customers is dealing with all of these tons of data that are produced in millisecond cycle times for operational needs. The Edge is actually the point where we get the data securely from the automation system to bring it together with IT data. And that way enable smart decision-making by customers about things that actually reduce energy consumption, improve productivity, improve yield, optimize maintenance schedules and things like that. So the Edge is really the point where the IT world and the operations world from the automation system come together. And you have, ABB has a product, Edge Genius, that uses Red Hat technology. Tell us about what that does. Yeah, so if you look at the Edge solution that we provide, it, number one, securely and safely extracts data from the automation system because you don't want to operate, you don't want to create a disturbance to the real-time deterministic operation that the automation system has. So if you've got the data out from the automation system, then you can visualize it, you can actually extract certain types of data, you can actually summarize data, and you can run applications either directly on the Edge or in the cloud or on an on-premise platform to do something useful with it and to help the operators in these industrial plants either to optimize their plant or the whole enterprise. What types of industries are mainly or are really looking into Edge Genius as you start to roll that out? Yeah, the industries that we are serving are the so-called process industries that typically operate continuous processes. So think about power plants or chemical plants or cement plants, but you can also think about things like ships moving around that actually continuously operate electrical equipment and things like that or offshore wind parks is an example that I used in the keynote this morning that is also using that technology. Yeah, so reliability, availability and security have to be at the forefront of everything you're thinking about because these are pretty significant systems that you're integrating with. Right, and many of our applications are in critical infrastructures like energy or water where you don't want to have any shutdowns but even in plants that are not that critical for infrastructure, shutdowns might be extremely expensive. If you think about a petrochemical plant, a cracker is supposed to run eight years in a row without interruption. So you certainly don't want to be the reason why they have to shut it down. I myself in the previous role, I was running a semiconductor production in ABB and it's clear if you actually disrupt the production process and send people out, there's a lot of silicon in the diffusion ovens that you just throw away. So a lot of these applications make it very expensive to be interrupted. On the other hand, there's also a high need for safety and security because if you think about an exothermic reaction in a chemical plant that is controlled by the automation system, you really want to make sure that the automation system doesn't create an explosion creating a hazard for either the environment or the people or the plant. So I mean, you've got to outline a lot of scary scenarios there. What challenges does that present to process automation, which is your focus? Yeah, if you look at process automation, typically customers want to combine two things. They want to have something that is rock solid that they can 100% trust on doing the things that we talked about. But at the same time, there's a need to bring in new technology to actually evolve the automation according to the market needs. Many of these plants are built for very long lifetimes. 50 years is not uncommon. But during this timeframe, the markets change completely. And so the automation system that actually needs to cater to these market needs need to change along. So the challenge that our customers have is how do we actually reconcile these obviously somewhat opposing needs? Stability on one hand and evolution on the other hand. Now we do that by actually separating the inner core of the automation, which hopefully is not changed very often from the parts that are actually taking data out and allowed to optimize, allowed to change things that are not directly related to the inner control loop, where you can even deploy machine learning, artificial intelligence and other new technologies in order to have an outer environment where you're able to change things. This is exactly the environment that we would like to use these containerized microservices that we can deploy flexibly either on the edge or in the cloud. I think it's really interesting. I actually worked for Petrotrin and I actually could tell from the smokestack, you know, it was the cracking of being efficient or not efficient based on the color of the gases coming out and burning. So I definitely understand I think where you're coming from. I think what's really interesting is you look at a lot of these manufacturing facilities, they're building software, but not a lot of software in many cases. They're doing a lot of off the shelf bringing in your automation. How do you really, because a major theme that's been going on for the last two days is really simplicity. How does what you're doing with Red Hat really help them from a simplicity standpoint with edge? Yeah, the simplicity mainly comes in the form of managing the applications on the edge or the related applications that are running on the cloud. Because if we want to evolve them in an environment that is not used to frequent changes, we have to bring in automation and we have to bring in the capability to efficiently evolve the microservices. And this is exactly where the Kubernetes environment where OpenShift and Microshift play a major role. Is it a fully managed service that you offer with that? Do you manage the Red Hat components or does the customer actually manage that for them? Well, typically we manage it for the customers, but you also have to see that customers in this space are very diverse. Yeah, and so you've got big companies that have a huge competence in doing things on their own, but you also have smaller players that actually depend on service providers doing a lot of that. Real time is an important part of process automation and of course a lot of new technology has been coming forward in that area. What are some of the challenges of enabling decision-making using real-time data? Yeah, so if you look at real time, you've got real, real time and real time. And if I look at what is in the inner loop of the control, that's typically on a microsecond, millisecond basis where you want to have very deterministic behavior. So you don't want to have a lot of things that are not 100% controlled. But you're creating time series data that you can then extract from that environment and then the real time becomes less of that real, real time. But it's still having lots of data in a short timeframe that needs to be analyzed. So while the loopback influencing the process might not be within milliseconds, still you need to deal with the data that are coming with that frequency. So yesterday we were talking about ecosystems and ecosystem partners like yourself. How has it been to work with Red Hat on this solution? Yeah, the interesting part is if you've got a cooperation between two partners, number one, the best base for a good partnership is if the two partners are complementary. And this is really something that plays out extremely well here because while we've been the number one automation supplier for the process industries for the last decades and know the customer requirements and these specific verticals very well, actually Red Hat comes with all of the horizontal knowledge from the IT environment and bringing these two things together in the appropriate way to adapt what is available in terms of technology, managing containerized microservices in a flexible way in the diverse hardware environment in the specific environment that we deal with in our industries is actually a good match. The second component I feel is that once you start something new, you have to invest. And both parties actually have to be ready to invest. And maybe the success doesn't come within days. And I felt that there's really a good willingness on both sides for creating a win-win and do the necessary investments to finally succeed. And it would seem that it also has to be long-term given the life cycle of what we're talking about with these facilities. So, and I would say that this partnership has to outlast almost the technology to that certain extent. Do you see that AI is creeping in to the process automation? Because there's a big, I mean, there was the colonial pipeline hack and all of that fun stuff that I'm not fun. But now with everybody in looking at, okay, how do we get the data out to use it in different ways? I would assume that that's part of what you're looking at with ABB as part of the, to provide to the customers. But they're also looking for it for other reasons. Maintenance and facility management and staffing and things of that nature. There are actually two interesting aspects in your question. One is, and two buzzwords, cybersecurity and AI. So let's start with cybersecurity. Conventionally, obviously systems that were built 20 years ago could not even consider the kind of risks that they are facing today. So the only way to deal with current cybersecurity risks in those systems is to as much separate them from the internet as possible, which in some of our customers actually creates a complete block to connect anything to the internet. Exactly, agaping the system. But of course the future systems that will deploy in the automation plans, they will have to be secured by design. They will have to be zero trust so that each component can only work in that system if it's actually properly authenticated and has the proper mechanisms to make sure it's securely connectable. Now, independent of that, the edge solution that we were talking about is exactly the means to securely extract data from the automation system even if it's a legacy system. And it's no problem as long as you can make sure you've got basically an information diode. Information is only going in one direction. It becomes more complicated and more involved once you allow the link back into operating something in the system. So that's the next level of complexity. Now, coming to the AI, that's something that you obviously typically can't run very well on constraint devices directly in the process. So especially learning, that's something that you need to do outside. That exactly requires to extract the data from the system in a secure way and bring it to a place like the cloud where you can actually apply these machine learning mechanisms. And that's coming more and more. And the interesting thing will be how we will be able to bring the two worlds together. Yeah, I'll give you one example. One of the things that we are working on, especially if you look at the fact that the current generation of operators in these industrial plants is getting into an age where more of them retire and a lot of competence moves out of the door. So one of the things that we would like to do in the operation of these plant is to provide as much support to the new operators to be able to deal with particularly unforeseen situations. Because all of the usual stuff is handled by the automation system without human intervention. The humans are needed for the unforeseen stuff. So helping them to understand if something happens, oh, this has already happened three months ago. And another case where something similar happened was six months ago. And actually what the people did at that point in time was this. And then, oh, you can actually simulate what you would like to do now and what we would propose you to do now based on the digital twin of the plant. And only then apply the change. That's a chain that typically requires data processing, artificial intelligence, machine learning that is going on outside of the core automation system. But that has a very tight interface of it. Incredibly complex set of tasks and challenges that you face. But we're glad that ABB has you to drive this process. Bernard Eschermann, Chief Technology Officer at ABB Process Automation. Thanks so much for sharing your expertise with us here and teaching us so much about process automation. Thanks a lot for having me. We're going to take a short break. This is Paul Gillan here with Rob Shea at Red Hat Summit 2023. We'll be back in a moment.