Automation without boundaries

It’s been some three decades since the Purdue model first formalized the seven-layer pyramid for “computer-integrated manufacturing." While the model’s intent was primarily functional, form often follows function, and the intervening decades have left us with complex strata and siloes of isolated data that is difficult to integrate and contextualize. Peter Zornio, chief technology officer for Emerson’s Automation Solutions business, joins Keith Larson to discuss a new vision of a flatter, software-based architecture for process automation that democratizes data and enables optimization without all the former hurdles.

Transcript

Keith Larson: It's been some three decades now since the Purdue model first formalized the seven-layer pyramid for computer-integrated manufacturing, that those of us in the industrial automation space have all come to know, if not love all the time. And while the model's intent was primarily functional, form often follows function, and the intervening decades have left us with complex strata and silos of isolated data that is often difficult to integrate and contextualize.

Hello, my name is Keith Larson, publisher of Control magazine ControlGlobal.com, and you're listening to a Solution Spotlight edition of our Control Amplified podcast, sponsored this week by Emerson. Here to discuss a new vision of a flatter, software-based architecture for process automation that democratizes data and enables optimization without all those former hurdles, I'm happy to be joined by Peter Zornio, chief technology officer for Emerson's Automation Solutions business.

Welcome, Peter, and let me first start off with congratulating you on the return of Emerson Exchange as a live event after a three-year, pandemic-induced hiatus. It's really great to get together in person and witness nearly 3,000 automation professionals gathered together in one place to learn and share best practices. Welcome.

Zornio: Hi. Thanks, Keith. I totally agree. It was really almost a palpable energy. I thought you could tell everybody was just really happy to be back collaborating together in person.

Larson: Yeah, absolutely. It was a real pleasure. I guess just to jump on in, there were lots of compelling innovations on display at Exchange, it's hard to even make a list. But I was particularly intrigued by the future vision of Boundless Automation that was really presented for the first time there at Exchange. Can you just take a minute, and we're gonna dig into this in the course of the podcast, but try to walk our listeners through the essential aspects of this new architecture and how it's different from the automation systems we've seen up to now?

Zornio: Sure, Boundless Automation is our term for the vision of the future of operations technology architecture. And by that I mean, everything ranging from the intelligent field devices to automation to OT software, the complete envelope of things that make production OT functions, operations technology functions, operate today, with some specific examples of new products that we have that illustrate those concepts.

We got there pretty simply by looking at some of the limitations that you already mentioned of today's architecture and then looking at all the technology advances that exist in the IT world and in cloud and this new set of technologies that we call edge. And I think digital transformation programs really highlighted a lot of today's limitations, those things you mentioned.

The fact that we have very layered striated architectures with software sometimes tied to very specific hardware, it becomes difficult to move data up and down the layers seamlessly. We've built the layers largely as a security construct, or, as you already mentioned, as form follows function. And then, we also have seen with digital transformation, as people went beyond automation in production and went into areas like reliability or sustainability or quality, that they actually built an entire other, let's say, vertical silo of data with the applications that work in those areas. I mean, I worked with a number of folks specifically on moving sensor data directly to the cloud, for instance, for reliability programs.

We looked at those limits, and then we looked at all the technologies that exist that we probably haven't taken advantage of in the OT world as much as we should, technologies from wireless cloud edge, and we said, okay, really the thing that makes sense going forward in the future isn't this kind of silos or layers, it's thinking of three integrated computing domains—the intelligent field devices, the edge technologies, and the cloud—combined as peers from a computing environment perspective, but with a cohesive software environment that ties the applications in those three environments together. So, much more those three compute environment as peers, rather than the layers we've traditionally come to think of them as. And that's really the core concept that we discussed as Boundless Automation.

Larson: Yeah, you've used the phrase integrated by design to describe how those various levels or those platforms, I guess those are those pure level platforms, as you just described, is it more than just common communication standards, API's and the basic information models that we already use? What's different about the communications now than it than it has been in the past?

Zornio: Yeah, so although I've talked about these three compute environments as peers, each one of them will be supporting software. Software is what's actually going to deliver the value applications. But the concept is that they need to be integrated together from a software perspective, much more beyond what standards typically do. Standards generally will give you what I call a lowest common denominator kind of functionality.

Now, standards are absolutely good and essential for some core infrastructure. For instance, it wouldn't be great if we were changing the voltage and frequency of the power coming out of the walls every day. Or even looking at a more digital example, if we look at the cellular technology, a lot of people believe that the US got behind and got more expensive because they didn't kind of standardize on one communication tech right from the beginning. But I don't think anybody would have said it would have been a good idea to try to standardize what the actual mobile devices and phones on that infrastructure work like.

So, you need to very selectively decide where you're going to have standards, because by definition, you can run into the problem of limiting innovation. It's not always a purely technical problem. You can define very complex standards that might deliver a very high level of functionality, but then you're going to get into a commercial and testing issue around actually having a multi-vendor environment and who's going to be in charge of making more complex pieces of software work together.

So, I think what we've seen is that the software market, over 20-30 years, has proven that people want and will pay for a unified software ecosystem that provides much more ease of use and provides for usability of the data, what we call data democratization, inside that software ecosystem. You know, Microsoft originally, when you look at the direction ERP systems, went into integrated suites, when you look at what I already mentioned, the mobile and phone ecosystems and the apps that went on those, and in our space the distributed control system (DCS). Frankly, it's a pet peeve of mine when people talk about DCS kind of like its hardware. A distributed control system is an integrated, cohesive suite of automation software that's been designed from beginning to work together, contrasted with individual pieces. As I've mentioned, that concept has obvious benefits that we've seen people be willing to pay for.

Larson: Yes, definitely. Certainly, the Apple environment and the ecosystem, many people are happy with that. And it's just harder to be totally open and integrated. That's a tough road to hoe. You can't always have both even in an ideal world, I think.

Zornio: Exactly.

Larson: So, providing more effective and efficient data management is one aspect of that, because when you're dealing with multiple environments, and that contextualization of data has been really been a key push for Emerson for several years now, thinking back even to 2019 and before the pandemic. The recent controlling interest acquisition of AspenTech inmation technology has promised to advance those capabilities even further. Can you maybe talk a little bit to the importance of effective data management in this boundless automation system of the future? I mean, we're outside of the traditional constraints of the DCS, which did a pretty good job, actually, as long as you were in that domain. How do we manage that in this broader ecosystem?

Zornio: Yeah, first, I would say that effective data management and making data usable is probably the key driver for any next-gen architecture, at least in our eyes, okay. We have a lot of good software that provides a lot of specific functionality. We have a lot of analytics tools. If we've learned anything from customers running digital transformation programs, it's this concept of data democratization or usable data is what's key. Everybody figured out they could use standards, they could move data back and forth, or put it together in a data lake in a cloud. But actually having the context in that consistent data model around that data is what makes it actually usable when you go to put it in an actual application.

In inmation, which you already mentioned is a technology that Aspen just purchased and was a technology Emerson was using, it does a great job of taking what you have today in terms of disparate software, all of which with its own data model, and taking that and layering on a cohesive S95 overall data model on top of that. But what we discussed in the Boundless Automation vision is that should be inherently designed in this integrated portfolio of software, that you shouldn't have to do something like you do with inmation to superimpose another data model, the application should share that data model right from the beginning. And of course, now with Emerson and Aspen, we have the broadest suite of OT software in the industry, and we have a great opportunity, a great, frankly, selfish commercial opportunity to bring them together into that sort of unified software environment that will deliver unparalleled ease of use and data usability, or data democratization, for our customers to differentiate that extended suite of capability.

Larson: That makes a lot of sense, makes a lot of sense. Next question, the process automation industry has long been pretty conservative bunch, and for good reason when it comes to adopting new technologies. But the pandemic seems to have only accelerated the widespread acceptance of especially cloud as an important extension of on-premise automation systems. What advantages, from your perspective, does cloud connectivity offer users of Emerson systems? And perhaps more importantly, how does it increase the value you can deliver on their behalf?

Zornio: Well, the first thing I would say is, of course, it has really run to the cloud. I mean, I can tell you from our own IT department, right? So, IT has fully embraced the cloud. And frankly, there's many applications in the OT world that are not operations that are time critical that could move to the cloud today and enjoy those same benefits that the IT folks are enjoying as they move to the cloud: outsourced infrastructure, know software and clients, universal access, elastic scalability, pay-as-you-use consumption-based models. Yes, OT has had a lot of trepidation to do this in the past, but I think that's changing. And I think part of it is because IT has gotten more involved with OT, again, because of things like digital transformation programs. But beyond that, IT will just make integrating OT systems with business functions like ERP easier. Now, it doesn't solve that data integration issue, as already mentioned, but it's a better starting point when you have the various systems all housed together in the cloud. But the cloud will also make for centralizing remote support or leveraging expertise, all important things as people look at remote autonomous control. Cloud isn't going to be for all OT functions, at least not yet. You know, latency, the availability of the connectivity, data privacy, it'll be a long time before everybody's comfortable with critical functions in the cloud. But that's where we see this new generation of what's coming to be called edge technologies, typically based a lot of times around Linux, but also using things like hyper converged infrastructure and other new IT technologies coming in and being the new environment that we could deploy those more critical automation kind of functions, and still get things like fleet management, hardware independence and easy scalability.

Larson: Yeah, I think it's that transparency between the two environments where you can put what makes sense in the edge there and what makes sense in the cloud. And, you don't really have to even know where they're residing, if you see latency issues or whatever put it where it makes sense, but not have to worry about the glue in between.

Zornio: That's exactly the idea. But again, that's where having a cohesive software infrastructure come together that allows deployment anywhere, but it still allows that unified software environment that delivers usability and data democratization

Larson: It takes a lot of work to make things really simple, right, on the outside.

Zornio: That's a great expression. That is very true.

Larson: Well, it's hard to talk cloud without talking about cybersecurity in the same breath. Along with integrated by design, Emerson has used the term secure by design to describe some of those elements of the Boundless Automation implementations. Can you talk more about what that means for users from a practical perspective, the next generation of cyber security and Linux systems?

Zornio: The goal is very easy to understand: more secure, less complex, easier usability of data. I think anyone who's wrestled through setting up those layer networks today and configuring firewalls and doing all the things that you need to do to really properly secure an OT system today understands when I say, it's not easy and it is complex. So, zero-trust security is a concept that's been around for a while that would make that happen. More security, less complexity, where you don't depend on the layered protections like we do today, but grant access purely on an as-needed basis. Sounds great, but it's a concept that really needs to be designed into the architecture from the beginning.

Larson: Yeah, yeah. It's relatively easy to envision how the cloud and increasingly intelligent field of measurement and actuation devices will participate in this Boundless Automation vision, but what about the control functionality itself? I have a harder time getting my head around, how that's contained or corralled so that we still get all that deterministic response and it all works together? I mean, at Exchange, you showed up a package controller that brings together controlling IoT connectivity, but previewed that new DeltaV edge environment that due out in '23. Do these offerings represent a first step towards that cohesive platform at the center of tomorrow's Boundless Automation system?

Zornio: Yeah, those products are great first examples of what we're talking about when we describe bringing modern edge technologies, and bringing them in to that secure control environment. In this particular case, what we've done with both of those products is we've buttressed up the modern, Linux-based edge environment, all the technology that comes with that, the ability to do deployment and monitoring from the cloud, the ability to take advantage of all the new tools that exist in that environment, open software tools, and have them be right there next to the security and control environment, whether it's the one in the DeltaV DCS, or the one that's in the pack, you know, automation controller environment.

First step is saying, you don't have to give up this control environment, but it's absolutely, we see where those control environments become the environments we build upon to move those control functions into including the core control function in something like our DCS controller. As I already mentioned earlier, DCSs are largely integrated suites software. Yes, a piece of them, that control piece and the IO piece, are implemented on purpose built-hardware today, but there's no reason why that's going to continue to be the case as we move into the future.

When you look at the field, we also see technologies like APL increasing the capability of field devices to become edge compute devices all on their own. APL provides increased power, which allows you to have increased CPU and memory, which means you can run more advanced applications, even in intrinsically safe areas. And when we look at leveraging technologies like 5G, we actually see expanding the world of our traditionally process-oriented sensor technology into remote and mobile applications. So, those two products are great examples. We also showed some products, where we see field devices becoming more intelligent and becoming basically little edge servers all on their own.

Larson: Well, I'd be remiss if I didn't bring up one last newsworthy item that just after Exchange, Emerson announced it divested a majority stake in the commercial and residential side of the business, in a deal valued at, what $14 billion? I think Lal Karsenbhai noted that this is the next step in Emerson's ambitions to become a "global automation player." I think that's a pretty nice nest egg when it comes to investing to increase the footprint and capabilities of what I think of as an already significant global force, but what do I know? What can you tell us about developments we might expect in the coming months?

Zornio: Yeah, I think Lal describing it as a next step was a little bit of an understatement. Certainly the goal was and did achieve creating the largest pure-play automation company. And frankly, we're big in all areas whether it's sensing, final control, actual control as I mentioned, in all areas whether it's process, hybrid and discrete manufacturing. And of course, with Aspen, we set up the largest, or certainly one of the largest, industrial software companies for OT software. So, you can expect that, yes, we'll take that nice nest egg, as you described it, and be targeting expansion in basically all those areas, because it's very clear we're now committed to being a pure-play automation company. And that's our clear future. There could be some highly related or highly overlap areas that maybe get invested in, but automation is what Emerson is going to be about.

Larson: All right, great. Well, thanks so much, Peter. Excited to see what those coming months and years bring. Thanks for sharing your perspective with us today and it's been thought provoking and a real pleasure as always. For those of you listening, thanks for tuning in, and thanks also to Emerson for sponsoring this episode. I'm Keith Larson and you've been listening to a Control Amplified podcast and my guest today has been Peters Zornio, chief technology officer for Emerson Automation Solutions. Peter, thanks so much for joining me today.

Zornio: Great talking to you as always over 30 years, Keith.

Larson: I'm afraid to admit that. We've been through some acquisitions ourselves, so I know exactly what all that is about.

So, for you on the listener side, if you've enjoyed this episode, you can subscribe wherever you find your podcasts and can find the full archive of past episodes at ControlGlobal.com signing off until next time.

For more, tune into Control Amplified: The Process Automation Podcast.