Utilizing Tech - Season 7: AI Data Infrastructure Presented by Solidigm - 05x16: The Convergence and Standardization of IT and OT with Andy Foster of IOTech

Episode Date: August 21, 2023

Industrial control and operational technology systems face the same forces as IT applications and are rapidly being integrated into comprehensive edge infrastructure. This episode of Utilizing Edge brings Andy Foster of IOTech to discuss standardization of industrial IoT with Allyson Klein and Stephen Foskett. There are many more similarities between industrial control systems and information technology systems than many people realize. Both are impacted by standards in commodity hardware, networking and communications protocols, and cloud integration. And OT and IT are colliding at the edge, with a massive opportunity to leverage data. Hosts: Stephen Foskett: https://www.twitter.com/SFoskett Allyson Klein: https://www.twitter.com/TechAllyson Guest: Andy Foster, Product Director and Cofounder of IOTech https://www.linkedin.com/in/andrew-foster-673b268/ https://www.iotechsys.com/ Follow Gestalt IT Website: https://www.GestaltIT.com/https://www.UtilizingTech.com/ Twitter: https://www.twitter.com/GestaltIT LinkedIn: https://www.linkedin.com/company/Gestalt-IT Tags: #EdgeComputing, #IoT, #UtilizingEdge, @UtilizingTech, @SFoskett, @TechAllyson, @IOTechSystems, @GestaltIT,

Welcome to Utilizing Tech, the podcast about emerging technology from Gestalt IT. This season of Utilizing Tech focuses on edge computing, which demands a new approach to compute, storage, networking, and more. I'm your host, Stephen Foskett, organizer of Tech Field Day and publisher of Gestalt IT. Joining me today as my co-host is the fabulous Allison Klein. Thank you so much for joining us on the show today. So happy to be back, Stephen. Another topic on IoT and Edge. I can't wait. Exactly. And that's been the interesting thing that there is such a crossover, isn't there, between IoT and Edge? Because, I mean, truly, IoT and Edge are kind of, well, dancing around the same, dancing to the same music in a way,

in that you've got devices all over the place, you've got sensors, you've got all sorts of stuff happening. And that's especially true in industrial settings, where you've got industrial control systems and so-called operational technology, or OT. And we haven't talked about that much on the podcast, though we have brought it up again and again. Yeah, you know, it's funny. I was preparing for this podcast, and one of the thoughts that came into my mind was, it's weird that IoT sounds like the old school technology when we talk about where we're going with this,

because I think that IoT still has so much potential in terms of where it can go and what it can do for organizations and for society. But yeah, the opportunity to integrate core capabilities and grow the edge into that distributed computing model that we've been talking about and looking at those industrial settings and thinking about what needs to change. I can't wait to get into the topic. Yeah, absolutely. And indeed, that's one of those things where we've been kind of, like I said, we've been talking about this a bit. So we decided today to invite an expert in industrial IoT and control systems onto the show. So Andy Foster is product director and co-founder of IoTech. And as you can tell from the name of the company, it's right there.

They know a heck of a lot of this about this stuff. Andy, welcome to the show. Thanks, Stephen. Thanks, Alison. It's a pleasure to talk to you guys today. Yeah, as Stephen said, I'm Andy Foster. I'm the product director at IoTech Systems Limited. We're a software company. We develop software platform technologies for the industrial edge and also management solutions to support industrial environments as well. And it's interesting that those systems are starting to become interoperable and intermanageable and integrated into this whole global compute world because, I mean, traditionally industrial control systems were pretty standalone, weren't they?

Absolutely. I mean, you know, edge computing in the OT world is not particularly new. I think what is developing is the convergence between the OT world or the edge OT world in industrial systems and also the IT world. And that's where that's the direction of travel i think a lot of the big companies and organizations that we support across different verticals who are deploying um edge and industrial iot solutions they they have both you know typically a system will have both edge elements and cloud components. So interoperability between those different, which have been segmented environments traditionally, is becoming more and more important. Andy, I was so excited to talk to you because of your breadth of experience in this space. You talk about the convergence of OT and IT, that represents not only a change in the way infrastructure is deployed,

but also the cultures associated with that infrastructure and how it's managed and the thought processes about what you can do with the information on these systems. Can you tell me a little bit about how you've seen this evolve with your client base and what you've learned in terms of the combination of OT and IT? Yeah, as I said, I think one of the, this is something that's been evolving, you know, probably over the last six or seven years. You know, as I mentioned before, one of the key challenges is that traditionally those worlds have been, if you like, siloed. The IT world is fairly well understood.

A lot of the technologies that are used within there are fairly well tried and tested. The OT world is very, very heterogeneous. There's a lot of different technologies and devices and different types of standards that are used in the OT world. And actually to bring those things together is a significant challenge. There's a cultural challenge and there's a technical challenge. The IT guys don't necessarily understand the OT world. And then technically, because you're dealing in the OT world with a myriad of different technologies,

being able to, particularly when we're focused on the industrial IoT world, what we're talking about is acquiring large volumes of data from different systems. Some of the data gets processed on the edge. A lot of the data gets shipped up into the cloud, into the IT world for big data storage, further analytics and things like that. So having the IT guys understand what's going on in the OT world and then supporting the actual technical challenges of moving that, supporting a hybrid system where the data has to be processed both on the edge and in cloud has, you know, is something that is evolving. The technologies initially were pretty immature, quite fragmented.

Building these systems was quite an integration nightmare. But as we've moved forward, technology vendors, the application of open standards is facilitating a much better convergence. And I think now the systems are, you know, the systems are being deployed successfully. A lot of, you know, a lot of organizations are deploying industrial IT at scale. So things have moved along a lot. And, you know, that's a combination of maturity

of the technology, the adoption of open standards, and then the vendor community has stepped up to the mark as well in terms of the enabling technology, the platform technology, the management technology that can be deployed to facilitate or enable this end-to-end system. It's interesting because everything you're saying, if you had changed, I don't know, the subject of the sentence would sound entirely familiar to people in data center IT, in cloud, in edge, in networking and security. It's like, you know, they say software is eating the world. Open standards and interoperability is eating the world.

Sensors are eating the world. You know, IoT technology is everywhere. It's like all of the same, all the same forces, technological forces that are impacting all of these other areas are impacting the world of industrial control systems. And the industries are responding in similar ways. But I think one thing that's different about the industrial space is that you're dealing with different kinds of companies and people with different kinds of backgrounds. And they're approaching things very differently because they don't have that sort of IT mindset that you'll find in the enterprise technology space, whether it's storage or networking. And I wonder if you can talk a little bit about how are these standards emerging? You know, what direction are they going in? And really, what's driving standardization

in the industrial space, as opposed to the IT space? Yeah, well, that's a good question. So standards in the industrial space are nothing new. So, you know, that's one of the challenges. There's a myriad of different standards for communicating with equipment in the industrial space. And actually supporting all those myriad of different types of protocol is one of the key challenges. How do you actually connect when you have so many different standards? And then how do you basically, if you're having to connect to equipment which has many different languages, if you like, and data formats, how do you take that data and present it in a normalized format, both to the applications that you want on the edge?

It might be your edge data processing or your filtering algorithms, et cetera, et cetera. And also that data, at least a subset of that data is typically going can actually support, if you like, the connectivity to many different data sources, particularly in parallel, can acquire the data at high volumes and possibly very low latencies, can take that data, normalize it, perhaps into a common standard representation. So this is where new standards are being applied. Things like, for example, OPCA, it's used heavily within manufacturing and industrial space.

So you may be taking data feeds from lots of different protocol standards, normalize it into, if you like, a common standard like OPCA or perhaps MQTT or JSON, and then provide that data to the IT guys or the cloud service environment in a format that they understand. So again, applying standards so that they can get that data in a representation that they can manage in a more efficient way. They can't just tap in to all of the different data feeds and all the different protocols. So again, if you like, from the edge environment, the standards being applied there to

normalize the data. And then from the edge environment, the standards being applied there to normalize the data. And then from the communication, if you like, northbound up into the IT environment, the cloud environment, the standards being applied there to support communications, to support providing the data in the formats that the cloud environment or the IT guys expect. And that can be, you know, that can be, you can have, a lot of times when you're interoperating between the edge and cloud, there's common standards that are being applied now. As I said, things like MQTT or REST technologies or even

PCOA as a standard communication fabric. But then on top of that, if you're working in different verticals, on top of the core communication fabric, if you like, there's data standards that have been applied across the verticals so that you're providing the data in the format that your domain expects. So things like in the renewable energy space, we've got things like SunSpec. In the building management space, you've got standards like Haystack, which are basically data models that you can apply on top of, if you like, the core communication standards. You know, one thing that I was thinking about as you were talking, Andy, is that coming from a data center space myself, I've seen the struggle for true interoperability in multi-cloud environments. And, you know, maybe some of the limitations of how cloud has evolved has come from standards not quite getting the customer there.

You are very deeply involved in the standards efforts in this space. What do you think is different here that is going to ensure that companies can really take advantage of true interoperability across infrastructure providers, across different manufacturing or edge environments to ensure that a diversity of solutions can be deployed and managed and be effective in terms of true interoperability? Well, I think over time, as I said, the underlying technologies are maturing. And also, there are a number of key standards, if you like, that are emerging as some of the, I guess, the winning technologies. Because, you know, there's potentially, there's lots of different mechanisms, if you like, if you want to communicate between your edge environments and your data center or your cloud service environment. But there are a number of technologies that are emerging, which I think are standards that are sort of becoming the dominant standards. And that's I think that's a good

thing. Typically, the standards that are successful are the things that get deployed in real environments the most, and then the standards actually evolve to support the deployments and the usage. And I'm talking about, as I said, I mentioned things like standards like OPCA in the

industrial and manufacturing world as a core communications standard that everybody uses as the base interfaces into the OT world. Or things like MQTT has evolved as one of the key standards for asynchronous communications. And if you like, it's one of the key mechanisms that many hybrid cloud edge environments used to communicate between the edge, so MQTT or secure MQTT. And then there's some others, but there are some standards that are clearly evolving where you see them again and again are being used as the common standards within the industrial space.

It's interesting that you bring up MQTT specifically, because I think that that's one that's really going to resonate with those who've been trying to deploy edge applications. There is a lot of MQTT underlying the, well, a lot of the edge systems that people are out there. And I think that it sort of exemplifies the collision between these traditionally segregated worlds. I mean, back in the day, it was pretty common, I think, for people to look at a lot of the systems that you're talking about,

these SCADA and stuff like that, and say, that's something completely different. That's something unrelated to what we're doing in I.T. There's that old saying that operational technology is I.T. for the places that don't have carpet. And I think that that that's really changing, especially in the edge environment, because, well, a lot of quick serve restaurants don't have carpet. You know, a lot of same for a lot of energy production and spaces. Basically, IT is stepping into those areas as well. And we're starting to find common ground. And as I said before, too, the technology, the protocols, they're all coming together as well.

I know that there's a lot of power over Ethernet, for example, that's being talked about in those spaces. And there's a lot of IP-based systems and cloud. There's a lot of cloud integration coming in the industrial control and OT world that may surprise people. Why don't you talk about that a little bit, Andy? How are sort of traditionally IT things coming in and being integrated with OT? Well, that's a great question. So it's actually at multiple levels. So for one of the things we haven't touched on so far is that the OT world is becoming, has tended to be hardware centric. So there's a lot of, you know, PLCs, microcontrollers, a lot of hardware based solutions solutions to support that environment. But the trend is now that the OT world is becoming more software-defined.

So there's much greater use of, if you like, standard commodity hardware, so industrial PCs, lightweight controllers. And basically, the traditional hardware-centric control systems are being redeveloped as software workloads that can run on an industrial PC on is, you know, lots of the OT environment now, the networks are supporting virtualized compute systems, things like containerization for providing, like, application portability. You're seeing a lot of OT applications which are containerized because it just makes it much easier. They're portable. It's much easier to deploy.

And you can run it, you know, as I said, you can run them on lots of different devices. You can orchestrate them from the cloud and stuff like that. So the OT world is becoming, adopting a lot of technologies top-down, if you like, that are coming from the IT world. Like I said, I mentioned virtualization, containerization, the use of commodity hardware and not bespoke hardware devices like PLCs. You can actually run the control systems on PCs as virtualized workloads.

The other thing we touched on before as well is you mentioned things like MQTT. Now, one of the key things that, you know, MQTT as a technology is very popular because, you know, what people are looking for as well and from our customer base as well is they want flexibility and choice and they want, you know, they want their system to be future-proofed, you know, from the day it's deployed and they may want to extend it to bring in new types of use case on their core solution over time. Technologies like MQTT, which is like an asynchronous message backbone, if you like, so it supports loosely coupled applications, it allows plug and play. So you can bring in, you know, you can be publishing, if you'd like, your data feeds across MQTT. They could be picked up by your edge applications.

You could be pushing the same data or subset of the data up into the cloud. But if you want to bring new elements into your solution, into your environment, whether in the cloud or in the, they can tap into that same data feed because MQTT as a technology makes it very easy to do. And MQTT is used, you know, as I said, a lot, not only on the data plane. So for actually the data that's coming off your sensors and devices, but it's used as an integration fabric on the control plane as well. So it's used to actually manage the systems because it's so flexible, you know, in terms of allowing plug and play and choice,

you can bring new ingredients into the system very, very easily. Andy, I think that we've been talking about edge on this podcast and elsewhere quite a bit lately. And one of the things that seems like the dominant topic is what is the opportunity from data at the edge? You know, when you think about OT with, you know, a tradition around siloed solutions that are, you know, primarily focused at risk mitigation of data escape, we're changing a bit to the opportunity statement of what this data provides and what it can mean from the broader organization to give access

more holistically. Where do you see the opportunities that your customers are looking for in terms of that data sharing from manufacturing and more industrial applications? And do you see any particular trends that extend across customer bases in terms of data opportunity? Yeah, absolutely. I think in terms of general terms, what the edge is facilitating is that, as you said, traditional, particularly in the industrial world, you had an organization may deploy a range of different siloed systems

and they have their own connectivity and they're talking to their own pieces of equipment and stuff like that. What people are trying to do now is actually, if you like, get a much broader, if you like, get a much broader, if you like, digitize their operations, get a much better digital picture of what's going on within an organization. And not just, you know, so they may need to do that. They may need to tap into many more different sources of data. And they may be coming out of their siloed systems as well

and be able to have a holistic view of what's going on in their environment, across the environment. So the edge is a key enabler, if you like, of facilitating that digital transformation of organizations, particularly in the industrial space where we work. It's the point where the edge technologies allowing the organizations to connect to all of the different systems, bring that data into the edge environment.

Sometimes they're processing it directly on the edge compute. They may be deploying new use cases which are running different analytics for perhaps doing things like predictive maintenance, determine whether your manufacturing line is running correctly or whether there's possibly going to be a fault on some sort of production process. The other thing it might be doing is normalizing the data and then allowing that normalized data to be sent northbound up into the cloud environment

so that you may be looking at long-term trends in terms of how your factory environment is running or storing the data for other types of offline analytics, etc. But the edge is key to facilitating that, and particularly the platform technologies that are being used at the edge. The ability to basically connect to any piece of equipment, independently of what protocols they're speaking, do that easily. It's typically quite a, traditionally that's been a, you know, connecting to your equipment

has been a time-consuming, costly integration activity with a lot of bespoke code. A lot of the platform new technologies are out there that allow the customers to basically connect to those systems through configuration or using modern techniques like dynamic discovery. You can discover endpoints. You can onboard those devices. The platform is running at the edge,

allowing you to acquire that data at whatever volume and whatever latency you need to get it. Take it, normalize it, independent of what type of data stream it's coming from. So as you said, you might have in a factory, you might have Modbus devices,

Profinet devices, Siemens F7 devices, bring that data independently from what the equipment is you're talking to, provide it into a normalized format. And that's where some of the, I said, some of these new standards are being applied very successfully, like normalize it into OPC UA or normalize it into, you know, J M Q T T Jason, and then supply that data up to either the other, other applications that are running the edge environment or possibly push that data up to the, to your it environment for, for other types of processing.

Yeah. That whole description is going to sound very familiar to those of us in the IT space who've been involved in observability specifically, because that's exactly what happens with a lot of the enterprise observability trend right now. It's the same thing, just different protocols where they're collecting a lot of data at the endpoints, they're organizing it, they're filtering it, they're recategorizing it, and then they're sending it on for processing and analysis. And I think that that's actually another thing that I'd like to call out here is that whole idea of locality of data and locality of processing.

I know that that's something that was very important as well to you, Andy. The idea that there's a flood of data at the edge, and there's going to be more because there's more and more sensors. There's video as well. There's all sorts of different kinds of data collection that are really exploding in industrial IoT, in building IoT, and in edge systems. All of that data has to be processed, has to be categorized, filtered, organized, and then sent onward for processing. And I think that that's exactly what's happening in your world as well, right? That's exactly. And as I said, that's one of the key areas where the edge is actually bringing value

to that. There's many more, if you like, sources of data. You know, the organizers, as I said, there's a trend to connect to lots more systems than they were traditionally connected to to take that data potentially for operational purposes. It may be the data that's coming out of an existing control system, but it may be valuable data in terms of

how the overall operation of a factory is running. But because of the deluged data, typically, you know, because of things like cost, latency, security, perhaps, you don't want to be shipping all of that data up into your cloud environment to be processed. The edge can provide a lot of value there. And basically, as I said, processing it locally, making some decisions on that data which you want to do locally for you know might be for for latency reasons or just makes much more sense to process it. It can filter that data, it can cleanse the data and then it

can present the data in a in the format that data in the format that the northbound services in a much more consumable format, in a more standards-based way, make it much easier to consume that data. It makes, as I said, it makes the integration easier. The edge can actually, by processing that data and providing it to your northbound services in your IT environment or in your cloud in the format that they want and only the information that they're really interested in, it means that all that filtering of the data doesn't have to go on in the cloud, which is traditionally what happened before, you know, before the trend to do a lot of that processing down on your local edge nodes,

a lot of the data was simply shipped up into the cloud where it was filtered and transformed and stored in databases and stuff. But the volume of data and the more and more systems that are being connected up is just becoming prohibitive. So the edge has a crucial role to play, even if it's just a source of where it's filtering the data and, you know, transforming the data before it gets pushed northbound. So I think that one of the things that really comes together for me with this conversation is we've talked about edge and the core capabilities across edge, but in the industrial space, edge to me represents

even more than just the opportunity around data. It also represents what I see as the culmination of the convergence of OT and IT. And that to me is really interesting. I've learned a lot from this conversation, Andy, thank you so much for sharing your views. And I guess the real question to you, Stephen, is where do we go next on this podcast in terms of learning more about the industrial space of edge? Yeah, absolutely. I would love to continue this conversation because I see so much commonality and so much similarity between what we've been talking about when it comes to IT systems and what Andy's talking about when it comes to OT systems. Yeah, I know, just to say, you know, Alison and Stephen, it's been a pleasure talking to you today. I think that

this subject area is, you know, obviously close to what we do as talking to you today. I think that this subject area is obviously close to what we do as an IoTeX systems. I think it's a very interesting area which is becoming very important. If anybody wants to find out any more about how we manage edge systems and how we support

customers that are converging their IT in their own team environments, they can come to our website, www.iotexist.com. We've got lots of information there. There's ways to contact us through the website. Or if you want to, you can contact me directly at andrew.foster at iotexist.com. And I'd be more than happy to answer any questions that you have. Thanks so much, Andy. I really appreciate getting a little bit of a glimpse into the world of OT. Alison, how about yourself? What are you

looking forward to lately? I've been following Edge and writing and interviewing on Edge on my platform, The Tech Arena, which is thetecharena.net. And there you can check out a number of different interviews from folks about the edge as well as download my 2023 edge computing report. And I just can't wait to learn more. I think this is such a vastly changing landscape as technology solutions hit the market and customers get more sophisticated in their deployment. Let's go. Yeah, absolutely. And that's, I mean, that was why we started Gestalt IT 15 years ago to get a bigger picture look and to understand more of the world of IT. And for me, conversations like this are really a continuation, a straight line continuation of that, because those of us

in the, who grew up in the data center space, then we learned about cloud, and now we're learning about edge. We're learning that there's a lot more to discover. So it's really a fascinating topic. Like I said, so much of the same technology that's being deployed in multiple areas, and so many familiar challenges that are being approached. We, for our part, are going to be actually this week at VMware Explore. Allison, I know you'll be there as well. We'll see you there. And of course, we'll be doing more Edge podcasts coming up. One thing I'm planning on doing is recording an episode on site, on premises, as it were, at VMware Explorer. So hopefully, Alistair and I can corner you and Alistair and

get that recorded next week. Also, of course, we've got more companies looking at our Edge Field Day event. And so we'll be inviting them to come on the Utilizing Edge podcast as well. Thank you so much, everyone, for listening to Utilizing Edge. This is part of the Utilizing Edge podcast as well. Thank you so much, everyone, for listening to Utilizing Edge. This is part of the Utilizing Tech podcast series. Previous seasons have focused on AI and CXL technology, and we're already starting to look at what we're going to do next time. If you enjoyed this discussion, please do subscribe in your favorite podcast application.

Give us a rating, give us a review. We'd love to hear that. We would also love for you to subscribe to the videos. You can find them on the Gestalt IT video channel on YouTube. And as I said, we're looking toward the future of utilizing tech. So if you've got an idea, maybe reach out. I'd love to hear from you. This podcast is brought to you from gestaltit.com, your home for IT coverage from across the enterprise. For show notes and more episodes, though, as well as contact information,

head over to our dedicated website, utilizingtech.com, or find us on Twitter or Mastodon at Utilizing Tech. Thanks for joining us, and we will see you next week.