Internet of Things

Securing smart factories: How Schneider Electric connects devices and prevents outages

In the industrial IoT, factories and data centers face two main barriers: connectivity and poor electrical infrastructure.

TechRepublic's Dan Patterson spoke with Andrew Bennett, the SVP of IoT EcoStruxure at Schneider Electric, about EcoStruxure's role in the Internet of Things.

Watch the video, or read the full transcript of their conversation below.

Patterson: The Internet of Things has a profoundly powerful impact on the digital transformation of almost every industry. Andrew, thank you very much for your time today. I wonder if we could start with first, what is EcoStruxure, and what role does the Internet of Things play?

Bennett: Yeah, great, thank you, Dan. Schneider Electric of course, is a company that works with a lot of industrial clients, so you can think about us in industrial space and buildings, in the grid, and of course, what we do for places like data centers. And one of the things that we've been doing for a long time is thinking about connecting devices in those settings to enhance analytics. That's really kind of the core of what we do as a company. We take a very OT, operational technology perspective on that. And so EcoStruxure, EcoStruxure really is our set of architectures that allows us to think about all those connectable devices, and then edge control, how do you actually control those devices.

A lot of our clients are running mission-critical operational technology, so this could be a nuclear power plant, or a petrochemical facility, or the grid. And so, edge control of those devices, actually having intelligence, not just embedded at the nodes but at the edge, is really important as they can't risk being disconnected from the internet.

SEE: The Power of IoT and Big Data (Tech Pro Research)

And then at the top of EcoStruxure, at the top of that stack, it starts with those connected devices and then edge control, really is where we build a lot of our software. So we build software for control, for analytics, of where we're starting to build out some of the AI and machine learning, and where we store a lot of that data of course, in the cloud, so that we start to do more sophisticated views of what's going on in those manufacturing sites, or petrochemical sties, or whatever they happen to be.

And so that's, I would say at the crux of it, that's what EcoStruxure is. It's an architecture that allows all three of those layers to interoperate, and it's really designed to be completely open and interoperable, because while we think our customers should only buy Schneider Electric stuff, that is probably not realistic and so, we need to be very open in the way we work with other vendors as well.

Patterson: I'm glad about the nuclear power plants, there are all types of industrial facilities that are adopting Internet of Things, but generally when I talk to people, even in business technology, they tend to think of IoT as a smart fridge or your Alexa. Can you give us an example of a device that might be IoT and deployed at large scale?

Bennett: Yeah, for sure. I think, think about just a breaker. Just your electrical infrastructure, and really the most important thing, because it's pretty hard to run a manufacturing process or a plant without power. And so today, because of really our heritage and our legacy with power, and delivering clean reliable energy, everything starts with that electrical infrastructure. So at plants today we have customers that are connecting those assets. They want to know about their energy consumption. They of course want to reduce energy consumption 'cause that's good to the bottom line.

It's also good from a sustainability perspective. So if you think about all that electrical infrastructure, then you can start to kind of work your way up the stack in terms of whether it's a motor that's running a conveyor belt, or part of your process. Whether it's a pump. These are all devices that historically were not all connected, right. They didn't historically see the need to necessarily connect those. Over time, more and more of those devices became connected, primarily so that those companies could control.

SEE: Green tech initiatives: Best practices and breakthroughs (free PDF) (TechRepublic)

So when we talk about OT, as juxtaposed to IT for a second, we're really talking about operational technology where a SCADA, which is the software for controlling that infrastructure, is actually opening and closing devices. It could be a breaker, it could be a pump, and what's happened over time is, of course that data is hugely valuable to start to think about broader trends. And so a lot of times we talk about this IT/OT convergence, which probably gets a little overplayed out there, but that's really those two worlds, right, where you're taking all these connected devices from an OT perspective, and starting to think about all the fantastic things you can do with IT from an AI, machine learning, and all the great things that are coming out there in the industry.

Patterson: You said the magic words there. What role does automation and artificial intelligence play with this connected infrastructure?

Bennett: Yeah, listen I think there's an endless set of applications in the IoT and industrial space of the Internet of Things. You know I also think we have to be careful, because I do think there's a tendency and I'm certainly guilty of it, of using a lot of the buzzwords and attracting folks to the pure machine learning side of the equation. But what we're observing with clients today, is first of all, a lot of them are still moving through that journey of actually connecting to those devices, right? Controlling those devices. So kind of the, I would say the ability to actually aggregate data.

But what's happening over time for sure is, as you aggregate that data, as you can start to look at broader trends, you could start to bring in things like machine learning, and the thing that I think that we're seeing today that is the most pronounced, is that you still need quite a bit of human interaction when it comes to machine learning or AI. You need to identify patterns, and then you need to feed those back into machine learning so that you know what that pattern recognition looks like, and then you can start to take proactive measures, and so, just one example.

SEE: What is the IoT? Everything you need to know about the Internet of Things right now (ZDNet)

You know a lot of outages or problems that happen in industrial setting, often start, you can actually look at things like partial discharge, or electrical partial discharge that happens in equipment. And if that happens over a long period of time, eventually that piece of equipment's probably going to fail. And so today, if you kind of looked at the signatures of what that looks like, a human being can look at that, you know we have thousands of electrical engineers in our company. Incredibly intelligent about what they do. You might not necessarily want to go out drinking with them but, they're a lot of fun too, to actually identify these problems. These guys can look at that, they can look at those signatures, they can instantly say, "You're going to have a problem here."

Now to codify that, and actually take that and build AI around that, that process is, I would say we're in the early stages. So you saying this, this is what that pattern looks like, now you can feed that into an AI system, and you can start to create a set of patterns that you can look for, and then obviously do what we ultimately wanna do, which is not just predict, but avoid outages and equipment failures in an industrial setting.

Patterson: Andrew, last question for you, what about the security? All of us adopting Internet of Things, see security as an integral component of the entire infrastructure. What role does cybersecurity and cyber defense play in adoption of IoT devices?

Bennett: Yeah look, the software, the hardware that we deliver for our clients, you know we are working with folks that are driving mission critical OT out there in the industry. So that's the good news. The bad news is that those clients are always gonna be targets, you know really aggressive nation state type attacks, not just your run of the mill hacker.

So for us, as a company, when you kind of go back to thinking about EcoStruxure and those three stacks, I think the important thing is to think, not just about how well hardened is that connected product, 'cause some of them aren't. They weren't historically, they were just maybe a thermometer, right, and so you wouldn't think about how you had to protect just a little device.

But you have to think about cybersecurity comprehensively across that entire stack. Not just the products, not just the edge control, but the software layer at the top. So, for us as a company, that's really our approach to think about all three of those aspects of EcoStruxure, how we not just bring new devices and add new IO to that environment, but how we think about cybersecurity from the product design level all the way up to that top stack of software, so, it's a big challenge, and you know we take it pretty seriously as a company, as you can imagine.

Also see

Robot factory automation
Thomas-Soellner, Getty Images/iStockphoto

About Dan Patterson

Dan is a Senior Writer for TechRepublic. He covers cybersecurity and the intersection of technology, politics and government.

Editor's Picks

Free Newsletters, In your Inbox