Harnessing the Power of Edge-to-Cloud Architecture

59 min video  /  55 minute read Download PDF
 

Speakers

Travis Cox

Chief Technology Evangelist

Inductive Automation

Cole Wangsness

Partnerships Leader

OnLogic

Mike Walsh

Senior Product Manager

OnLogic

Cloud-native applications have supercharged industrial systems with previously unthinkable levels of storage space and computing power.

But how do you take full advantage of the cloud when your facility relies on dispersed networks of lightweight edge devices? What are the software and hardware requirements? Where is the best place to start?

Having the right systems in place to connect the edge to the cloud will help you ensure scalability, efficiency, and stability, all without requiring a dedicated server room. Join industry experts as they outline the best methodologies for effectively collecting data and gaining insight from every part of your network.

  • Learn how to start leveraging the true power of the cloud
  • Discover the benefits of putting industrial hardware closer to data
  • Understand how software and hardware decisions influence each other
  • See what real-world edge-to-cloud success looks like

Webinar Transcript

00:00
Travis Cox: Hello, everybody, and welcome to today's webinar, “Harnessing the Power of Edge-to-Cloud Architectures.” My name is Travis Cox, and I'm the Chief Technology Evangelist at Inductive Automation. And a lot of what I do in my role is to help educate and provide and get awareness of the Ignition platform and technologies that are out there today. And I'm really, really excited about today's webinar. And we have a couple of amazing speakers presenting with me. First, we have Mike Walsh, the Senior Product Manager at OnLogic. And then secondly, we have Cole Wangsness, the Partnerships Leader at OnLogic. Hey, guys, how’re you doing? Welcome to the webinar. Thanks for being here.

00:40
Mike Walsh: Hey, thanks, Travis. Good to be here. Looking forward to this.

00:43
Cole Wangsness: Yeah, absolutely. We're looking forward to this.

00:45
Travis: So why don't you guys real quickly give a little bit of introduction to yourself and what you guys do at OnLogic. And Cole, I'll start with you.

00:52
Cole: Yeah, thanks, Travis. So as Travis mentioned, my name is Cole Wangsness. I'm the Partnership Leader here at OnLogic. So really, my goal is to manage our partnership program and work with customers to understand what, together with our partners, including Inductive, can do to build a better solution together and make it easier for you guys to actually deploy software and the hardware together. So I've worked with customers here for a number of years as our resident expert in verticals and software. So really, I work with a lot of folks similar to probably what you guys do on a day-to-day basis, deploying thousands of different types of solutions to thousands of different customers. So I've seen and done a lot here and aided by my colleague, Mike Walsh, who I'll let him go next.

01:35
Mike: All right, thanks, Cole. So Mike Walsh here. I'm Senior Product Manager here at OnLogic. And I own a lot of the portfolio products that we have here, and in particular, systems like Inductive. So it's really been great to work with Travis and the team at Inductive to understand what customer needs are. And then Cole and I will put our heads together for portfolio evolution, adding features and functionality, that kind of stuff to make sure that we're keeping up with all of the Inductive needs that are out there. So again, thanks for having us today, Travis.

02:06
Travis: Yeah, thanks for being here, guys. I know that OnLogic's hardware is great hardware that can run the Ignition software product, especially at the edge. A lot of what we're gonna be talking about here today. So let's take a look at today's agenda. First, I'm gonna briefly introduce Inductive Automation's Ignition platform for those of you who may not be familiar with it. Then we're gonna define what we mean when we talk, edge, and what the benefits are of an edge-of-network architecture. And that leads us into hardware and software considerations that we need to have when building a robust architecture. Then we're gonna talk about how to scale and bring the edge all the way to the cloud. And finally, we'll look ahead to the future and finish the webinar off with an audience Q&A. If you have any questions during the presentation, please type them into the questions area of the GoToWebinar control panel, and we will answer as many of those as we can at the end of the webinar. If we can't get to all your questions in time, we encourage you to reach out to one of our talented executives or with OnLogic to help answer those questions.

03:05
Travis: And then, to answer one question right now, a recording of this webinar will be made available within the next couple of days, so you can watch it on demand and share it with other folks. So first and foremost, let's get a little background on the Ignition software platform. Ignition is a universal industrial application platform. It's basically an enablement platform. It allows you to build HMI, SCADA, MES, and IIoT solutions. And it is used by 44% of Fortune 500 companies and 57% of Fortune 100 across all industries around the world. And it has an unlimited licensing model. So you get a server license that is completely unlimited in terms of tags, screens, clients, device connections, and more. It is fully cross-platform to run on Linux, Windows, and Mac OS. And it's really perfect, of course, when we're looking at talking about the edge. It is built on IT-standard technologies and uses modern technologies. It's a scalable server client architecture that is completely web-based and web-managed with designer and clients that you can launch anywhere. It is completely modular, so you can pick and choose the functions that you want, and you can easily expand onto Ignition as you go forward.

04:13
Travis: And it has a very rapid development environment and easy to deploy that out to clients. So that's a little bit of an introduction to the Ignition software platform. For those of you who may not be super familiar with OnLogic, Cole, will you wanna give some background on OnLogic?

04:31
Cole: Yeah, absolutely. And I'll just say, one of the benefits that's great with Ignition is the fact that it's so easy to use. Even someone with a business degree like myself can learn to use it and deploy it for customers. So if you haven't heard of OnLogic, really to give you guys a high-level overview, we are a hardware manufacturer in the industrial PC space. So what that really means is we're creating reliable industrial computers to meet the needs of your applications, offering a wide variety of products with preloaded with Ignition Edge as part of the Onboard Program, making it easier for you to deploy. Or if you're deploying Perspective or Vision or a full gateway, we can offer the system with an OS of your choice, bare bones, or customize that to your needs as well. So we work with many, many joint customers with Inductive deploying Ignition all the time, every day.

05:23
Cole: And so when we look at the products, what we have is a variety to meet your different use cases. Each project is gonna be a little bit unique. So our model is such that we've got industrial computers, fanless, ultra reliable, the vast majority of our systems are fanless, but those are designed for indoor manufacturing type environments. If your environment's a little rougher, and we'll talk a little bit about indoor versus more of a remote applications, you might want something in the rugged line, wider operating temperature, interesting power setups, impact resistance. Those are all things we do as well. But one of the big things we do jointly is HMI. So having a modular HMI where you can customize the screen size, resistive or capacitive based on whether your operator is using gloves or not.

06:10
Cole: Resistive being that manual, similar to if you used to like a... This will date me, but like an old ATM screen where you have to press really hard. The technology has come a lot farther, but that's the way most people know resistive screens versus capacitive is like the phone. So that's another key area for us. And then lastly, edge servers, when you need more compute... When you need something more towards a true server we can help with that as well. And what you'll see is when you look on our website, everything's built to order, and you configure it exactly how you want it. We can give you a real price and tell you exactly what you're gonna pay. And so we just try to make it super easy to work with us.

06:47
Cole: Really, our objective is to make it easy for us to be your hardware partner, so that you can focus on your software, your application. And so, that consultative model is for anybody. If you called in today, we're not gonna put you on hold or anything. You can work with a salesperson or for your larger project, one of our sales engineers, figure out what you need, get a prototype assembled quickly in days instead of months. And when you wanna deploy that mass, we can provide digital services like pre-imaging systems with custom images, BIOS customization, branding to really make it on your own. Or if you'd rather get something that just doesn't have any customization off the shelf, we always offer lifecycle management because we understand it's not always that people have the same lifecycle requirements as an IT group where it's, “We're just gonna rip and replace every three years.”

07:35
Cole: They want something that's gonna last a while. I said assemble too, we’re a US founded and headquartered company but with a global reach. So systems deployed for US and many international customers are actually built right here in Vermont, on nearby our production facility here. And we also have sites that will do manufacturing in the EU as well for our EU and UK customers. So really, at the end of the day, speak to one of our salespeople and our support techs, you're speaking to a real person in your time zone, and we really value that. So at the end of the day as an IPC manufacturer, we're always talking about the edge or talking about computing and challenging environments. So this is a type of challenge and conversation we have constantly with customers when trying to solve these problems. So we're edge-native through and through. I don't have an edge tattoo on my chest, but yeah, maybe one day. Maybe one day.

08:26
Mike: Yeah, I just really wanna reinforce what Cole said. OnLogic computers live at the edge. This is what we talk about all day long. If you ring up one of our sales folks, they can get a sales engineer on the phone. All of them are very familiar with Inductive, but especially the edge. So the portfolio, we've got a very wide portfolio, we've got a very deep portfolio. It can be a little bit much to get through, and we wanna get you to the right PC. So again, call us early, bring us in, let us advise. If you don't go with us, hey, that's great. Let us see if we can help you along in your project because this is really... What we really enjoy is seeing this stuff out there in the field.

09:06
Travis: Yeah, thanks guys. And that's really what we're here today to talk about and to look at is the edge and to understand how we can take advantage of it and what considerations we need to look at when going to the edge. A lot more customers are excited about the opportunities of a more distributed architecture and be able to get access to more data. But I think what we first need to do for everybody is start our presentation with the what and why of edge. What is edge, right? And for us, really edge refers to edge-of-network nodes that collect data and transfer it to a central location. And it also refers to functionality that is right there at the edge, such as an HMI, a local HMI, or maybe running machine learning tools or analytics at the edge. So it's really that compute that allows us to either move data or provide some local functionality that's required when we look at these critical systems that we have out there. We can't rely on connections to the internet. We have to have things that are close to that source of data. And as systems grow larger and more complex, we see the edge-of-networking architectures require as much consideration as any other part of the system.

10:16
Travis: And I think it's what you guys were saying is that you have these conversations every single day with folks around these considerations. And I think it's what we're hoping here to solve today.

10:28
Mike: Yeah, absolutely. And especially that last point, Travis, my very first job was working in manufacturing. I started out in the late '80s. And I mean, data collection seriously for us was a guy going down there on the floor about once a week with a clipboard and recording gauge readings and with a stopwatch, taking timings and stuff like that. And I mean, that's insane today with the way that manufacturing moves, the speed at which it moves. The defect rates that we're trying to get to. Being able to get that data quickly, easily, and at a very affordable rate today is completely changing the game. And that to me is what's really exciting is we've never had this opportunity before. We're finally hitting that point where the value of the data is so much more than the cost of acquiring it. And it just completely changes everything.

11:17
Travis: Yeah, absolutely. And we are seeing that, right? In the Digital Transformation landscape or in this... Within that fourth revolution that we're in, it's all about data. Data is so critical to our systems. And there's a lot of surveys out there that say that 80% of the data is straight in the field. And what we wanna do is unlock that. We wanna bring it in, and we wanna do stuff with it. But of course, that means that we have to think about it. There's gonna be investment that goes into it. But like, as you said, Mike, it's getting more affordable, and it's the power that we have that we can put at the edge is unlike anything that we've had before. So that's what we really wanna get into more today. And so I kind of mentioned my thoughts on what I think edge means. But what about you guys? Are you guys in line with that? What does edge mean to you?

12:05
Cole: Yeah, absolutely. Edge is a really broad topic. I like to say it's all around us. It's like the force in Star Wars or something. But what it really means is anywhere you have network endpoints, connected devices, that's all edge. So the plant floor, remote sites, whether that's monitoring at a pipeline, whether that's remote telco infrastructure, whether that's a lumber yard in the middle of nowhere, all edge, vehicles connected back via 4G to 5G. But the interesting thing that this presents to kind of expand a little bit on an earlier point you made is, the edge does introduce a new level of complexity. And what I mean by that is think about like an office, deploying in a typical, let's say client PC world where people are using Excel, video conferencing applications, first off. And really, the complication you have is the IT stack, software stack, networking, all the stuff you typically have to deal with in setting up an office. But at the end of the day, the device, whether you're at a small company or a company with 50,000 employees, is basically going to look exactly the same.

13:14
Cole: Whereas at the edge, you have so much more complications around all the different things you need to consider from connectivity, what are you connecting to, environment. We'll talk a little bit about that later, but it's really important to understand that you have this new layer of not just IT and OT hardware or software considerations, but now the hardware conversation is a much, much larger part of that. But at the end of the day you need to have this conversation because the start of this is the data that you're generating or that you could be generating is so valuable. And we can start, and we'll talk a little bit about this, but we can start with local views and reporting, but we quickly can scale these to really business-changing applications, whether it's quality control machine vision, whether it's predictive maintenance, whether it's big data analysis. The starting place is getting something out there at the edge and starting to collect data.

14:11
Travis: Yeah, and you...

14:11
Mike: Absolutely. And I love that whole point about operational technology. The world of operational technology, OT, has grown so much. I mean, back in the day, it used to be, “Hey, what of our IT capabilities can we take and extend out to the OT people?” Now, it's the OT people that are saying, "Look, this is exactly what I need, and I need it in a very turnkey manner, in a reliable manner, something that I can deploy and I can count on." And that's put a lot of challenges back on IT. So this is where folks like our companies are kind of living in this world. It's really exciting because IT has done their thing in that corporate office virtually the same way for how many years, decades, right? OT, it's all new, it's all exciting.

15:00
Travis: Yeah, and you guys made a point about, talking about some complexities at the edge. And I think... We see that. We live and breathe the edge in terms of looking at getting data. And we know that that world, I mean, the brownfield world, there's lots of devices that are already out there that have been in place for a long time that will still have to be running for another 10 to 20 years. And you can't just replace these systems, that's impossible, right? We have to be able to interface with, get that data, and bring it up. And so we do live in a brownfield world. And we talk about edge, it's really to unleash what's in that brownfield world to a more modern era. While at the same time, we are looking at bringing in new sensors, new equipment, that is also still edge, right? And we all this plays together into the broader infrastructure and into looking at more scalable solutions, looking at getting data or democratizing that data and making it available. And that's what we're gonna talk about here. We look at the cloud as we go forward.

15:56
Travis: So just a lot of momentum around the edge. And I just wanted to make sure to say that it was really about both brownfield and greenfield. We're looking at that together. But when it comes to deploying Ignition, I can definitely speak to this. We work with a lot of customers every single day on looking at putting an HMI on the plant floor. That is their guarantee no matter what at a critical asset, or maybe that's a remote site where nobody's typically out there, but if they're out there they need to be able to look at that data. Or it could just be a headless data collection system where they put it in the panel, it's DIN rail, it's ready to go. And they're just collecting that data, bringing it up automatically. And there's a lot of power to that. It could be a control room setup. It could be an onsite server that is running the full thing at a location or even, of course, at a remote asset or at the edge. All of that is considered edge, right? So edge is a critical consideration really for any deployment. And I think most of the conversations that we have is really around this, but how else are you really going to look at getting access to all the data?

17:02
Travis: Well, this is the way that it works. So in this case, there are a lot of key benefits of working, of having the edge out there. From having a direct interface with the onsite equipment, where you can reduce latency by collecting the data at that source. It allows you to have more distributed systems, and they're not tied to specific locations. They have more scalable solutions overall. We can have redundancy and failover. That's another big benefit of this. And that can prevent a lot of data loss with these critical systems. And for you guys, I know that it's also about working in harsh environments. And when you have those environments to have that critical system, that failover that's there. So when you look at some of the benefits of edge, what do you guys see?

17:55
Cole: Yeah, so it's a lot of what you see here, but at the end of the day it always comes down to the fact that there's no way to get the data. I love the term digital twin, because I think a lot of the time people are like, "Well, that means there's an onsite twin, right?" It's like, yes. And that means real hardware. So a lot of what you just said, it's how do you scale your organization from a process standpoint and a value standpoint? Because the big companies are all doing this today, the medium guys are doing it too. And so really everybody needs to be looking at this because these... Not only these values, but the applications you can deploy, the enhanced benefit that those provide from an organizational standpoint are business-changing, dramatically just business training.

18:46
Mike: Yeah, you think about how difficult this whole thing can be at times, right? You're talking about a PC acquiring data out on top of a wellhead somewhere deep in the Texas desert, right? Pumping oil. Keeping that thing connected, right? You're not gonna run Ethernet to it. You're certainly not gonna get Wi-Fi there. So you're using 4G. If that goes down, you better be able to store that data so you can analyze it later, right? There's value there. And on top of that, you better put something pretty reliable out there because you're certainly not gonna roll a truck to go repair that. Not to mention the downtime and everything else. So you've really got to look at that complete life cycle, all the reliability. Can I stay connected? Can I get my data? Because if you can't and it becomes a burden, that's the first time that you're going to lose that kind of corporate support to do this kind of an initiative.

19:38
Travis: Yeah, and you guys mentioned the applications, and I think that's really crucial, right? What are the applications we're trying to do? Hardware is a necessity, and we have to pick the right one. But the applications are a lot of times what drive some of these solutions. And so that could be an HMI, as we said. A standalone HMI out there. A lot of customers do this where they just need to connect those local equipment and provide the interface. But when we do that, we look at that as an opportunity because we're now, we're talking about edge, but it doesn't have to just be stopping at the HMI. That is critical data that is collecting and working with. And so when we look at a broader solution and bringing it up to a cloud or centralized application where we have long-term storage and all that, now we can do that store and forward, as you were saying there, Mike, right?

20:25
Travis: We can get that data and collect it at the source and forward it up. We don't lose that data, and we have that local interface. And we've always been talking about these architectures at Inductive in terms of like a hub that we call hub-and-spoke. And now we're just defining it more specifically as edge. This being a very critical application. But data collection in particular, we're leveraging edge and looking at technologies like MQTT to publish that data up to a centralized system efficiently. When data changes, perfect for remote locations where we have cellular or 4G, where we want to be able to be efficient so we can send more data and work with more of that out there, but get it up to a centralized system.

21:11
Travis: The technologies are really, in the landscape today, are available, are ready to go. And along with the compute, you have everything you need to make this happen. You put it all together, right? The data collection, the local interfaces, and then later looking at ML or analytics. And we'll talk more about the future. But these are some really cool applications. And I know, Cole, that you've been working with some customers, some actual practical applications, and real-world use cases of that. You want to share a couple of those with us?

21:40
Cole: Yeah, absolutely. And the one thing I'll just say before we jump into this is the cost to get a POC going for this type of application is very low. One of the things I love about working with Inductive is that you can do the trial with Ignition. So you get a piece of hardware from us, connect it up, and you can immediately start seeing value. So let's talk about how you'd actually potentially do that in real-world applications. So a lot of the time, as we've said, you're going to be dealing with brownfield deployments. A lot of the things we're going to be talking about here can apply to greenfields as well. But looking at brownfield deployments that we encounter on a regular basis, you have to deal with the legacy equipment. And the first thing that you can really do in a lot of these cases is display, visualize the data locally, and do some reporting. So to give an example, this was something that I'd worked on in the food and beverage space for chip manufacturing. And so back to Mike's earlier point, there are still a lot of people in the industry that are walking around with a clipboard, checking off valves.

22:42
Cole: And so interacting with the legacy protocols that were available, whether we could collect that data over serial, take advantage of the drivers that Ignition has built-in, we can start to pull that data from the machinery into Ignition without having to replace it, replace the machinery. And so you pair that with an industrial PC because these are not the cleanest of environments. So you're talking about salt, dust, debris, chip powder. Really, you look at a CNC shop, metal shavings. None of this stuff is conducive to a regular computer. So that's why you're looking at a specialized piece of hardware in the industrial space. But you can use it just like you would any computer. It's not like you're having to use some crazy wild OS you've never heard of. It's an x86, generally based computer, to deploy this. To start implementing Ignition, for example, now that we have that data we can automate that collection. We can visualize it locally for plant operators, right in the field, cleanly show real-time feedback, generate reports that are starting to provide true value for that customer. And now you've created a situation where you're showing immediate value, you're getting additional value out of your machinery, and you have a baseline.

24:01
Cole: So now you can look at how you scale. So we talked about a couple applications, but you see in this picture here, there are cameras online. You could deploy additional edge compute resources, a higher spec device at the edge, right on the line. Do quality control on that line. Look at those. Do machine vision. You could do predictive maintenance. Look at part failures over time. There's some great modules with Ignition to look at that and say, when do I need to replace equipment? How do I reduce downtime? There's very, very easy ways when you're starting out with HMI and recording to just show immediate value and start proving this out in your organization. And excuse if there's any noise in the background. The F-35s are overhead right now, so they're a little loud. But perhaps analyzing the data, you really can prove out what you want to do long-term. So I'll throw out one suggestion, and we'll look at another use case too. But as you scale up you can look at the cloud. You can look at, “Okay, what are these big data tools? What are these dashboards?” You can look at what tools you can bring in that really work well with Ignition.

25:14
Cole: We won't cover those here. There's some great, as Travis said, there's some great webinars that they put on, specifically running MQTT. But as you look at indoor use case, this gives you a baseline, improving that efficiency. What applications do you want to do long-term? Looking at how we can solve challenges at more of a distributed network. And really, I'll talk a little bit about wastewater, but this could be anything from wellheads, utility grids, larger wastewater networks. There's additional challenges we need to address, so these are real challenges that the customer suppliers are facing all the time. Looking at the environment, when we're talking indoor, we're generally talking: people need to be in those spaces on a regular basis. So it's fairly temperate. But when you get out in the field, you maybe not have AC and heating in that NEMA enclosure. So you're going to want to look at a device with a wider operating temperature. There are other factors we'll talk about a little bit that you generally consider with Rugged. But I'll hold off and let Mike talk to those. And the other big thing is you're going to need to look at the connectivity challenges.

26:19
Cole: So as we talked a little bit about 4G and 5G, you need a way when you don't have fiber or copper to connect to these sites. So really, the additional challenge that then causes is transmission costs. So putting the compute right at the edge to process that data to send up only what you need to the cloud or an offsite server is really important because it is expensive to transmit data over standard cellular networks. But again, edge computing will help minimize that cost. The other thing you got to consider too is if you're, there are many applications where the PC is a critical part of the infrastructure. You're not always just going to have a data controller that can, “Okay, if it fails, what happens?” A lot of the time, it's critical to plan operations. So having something that can operate outside of the network is also key. So as reliable as networks are, you always want to plan to make sure that you have as few risks for downtime as possible. So something like Ignition Edge where you can have the PLCs report into the computer itself, run the local HMI. And if for whatever reason that connection to the, let's say, remote server offsite of the cloud is lost, the PLCs can still transmit that data, store and forward that.

27:39
Cole: The local HMI still works for the operator. And then when that connection is re-established you can send those tags up. But at the end of the day, that doesn't impact plan operations. So really, in terms of scaling the application, whether it's predictive maintenance, whether that's organization-wide reporting, computer vision, the long-term goals and applications can work whether it's brownfield, whether it's greenfield, whether it's a manufacturing site, whatever the industry. But you need to really consider what additional challenges from a software and hardware perspective do I need to plan on based on those deployments. And it's something us and Inductive can help you do because we work with customers like you all the time to solve these challenges.

28:28
Travis: Perfect. So yeah, let's then take a look at what reliability means when edge is a critical part of the infrastructure. What is the cost to the customer in hardware, labor, and lost productivity? It's very important to measure what downtime really means to a customer, right? And how much does one hour of downtime actually cost? I know, Cole, that you have worked with some customers and actually looked at some of this cost analysis. You want to share some of the data there?

29:00
Cole: Yeah, absolutely. And this is one of the most important questions you need to ask yourself when doing an edge deployment is, something I've seen time and time again is that folks don't necessarily consider what the total cost of ownership means for the hardware in the overall solution stack. The bet... It can sometimes be easier to say if we implement this kind of application from the software side, we will save this much money from reduced downtime or improved operating efficiency, what have you. That can sometimes be easier. It can sometimes be complicated. But sometimes I feel that people, and what I've seen is that that piece of the puzzle, the total cost of ownership of the hardware sometimes gets ignored. And it's the reason why we specialize in this type of hardware because there are these new challenges you have to present. And if you deploy the wrong hardware, to give a quick example, and then I'll really tell a good story for you guys. If you're deploying a desktop tower, you bought it, let's say at Best Buy. You went to Best Buy. You bought whatever you saw the sales guy directed you.

30:01
Cole: If you deploy it in a factory, you're going to have issues. And it's just a matter of how much of those issues is going to cost you. A desktop computer isn't designed to be on 24/7 in terms of the components it's put into it. It doesn't have the environmental protections. And really, when you look at an hour of downtime, the cost can vary. I've heard anywhere from $50,000 on the low end per hour to well over a million dollars. You look at like, I think it was 2019, Samsung had I think a three-hour outage at one of their plants in Korea, and it cost them $3.6 billion. So for a large organization, it's even worse. But to give an example of how this applies in real-world, I was working with a company out in, and I'm not going to name names here, but I was working out in the southwest United States.

30:49
Cole: They were deploying a cheap PC at the time that really wasn't meant for the environment that they needed it for. And actually, they weren't using Ignition. So we were able to bring in Ignition too, and it simplified their software stack, so they saved you more money. But there was one really big problem for their cheap PC. They would fail all the time. They did not pick a system that was designed to go out at the edge. So every time this happened, they'd ship a system out to a motel. The tech would drive, I think it was from Dallas, he would drive out to the middle of nowhere. He would pick up the system, and he'd either sleep there or keep on driving. He'd go fix the problem, ideally. Sometimes he couldn't even get more parts, so then he's out there for a bit longer. And then he drives back. So they come in, they say, "Well, we want a system that's at this price point." I say, "Okay, let's just do the quick math here. You, by going with a system that wasn't designed for that environment to save a little bit of cost, let's say the price, I can't remember the exact price sale, so let's just say $400.

31:53
Cole: They saved $400 deploying these over what would be an equivalent system of ours. You've now cost not only two of those systems, but two days of the technicians full-time, all their expenses, the downtime at that remote facility. So a lot of times when I have this conversation of, well, we're deploying something today and it's really not working, the cost is actually hundreds of thousands of dollars of lost productivity. There's a great case study on our website from a company called American Woodmark. They were paying someone's full-time job to go out and vacuum desktop towers. They immediately were able to reprioritize that technician to switching us. So they were saving immediately. They turned a very repeating menial task. That was now a full-time technician that they basically had all free time of theirs. So it's really, really valuable. And I think I'd like to give it over to Mike and he'll talk a little bit more about those considerations and what's important.

33:00
Travis: Yeah, and one quick thing to kind of interject here is like we look at when talking to customers as well is making sure that we have the right hardware so that we can scale Ignition, that we could do more with Ignition. But it reminds me a lot of what my parents would tell me when I was younger about buying the right pair of shoes. Yeah, you can go get a cheap pair of shoes and you're probably gonna replace it because you're super hard on them. When I was in my playing sports every single day, it's a similar kind of thing. It becomes critical as part of the overall infrastructure. It becomes an important piece to it. And long-term ROI, you're gonna get, you're gonna gain by thinking of the applications and the value of the data that we're dealing with, not so much that initial investment. Yes, there's investment, but the point is, is to make that investment because it will protect that future. So I know I didn't want to steal any thunder there, Mike, but I want to mention that before you go in those considerations. The considerations are really important here.

33:56
Mike: Yeah, absolutely, Travis. Thank you for that. I'm gonna take a few minutes just to walk you through some things that you should think about, things that we've hit upon that can help you make the right hardware selection and avoid some of those issues that Cole talks about. First, environment. Environment's a huge thing. Dirt, dust, I'll extend the example that Cole gave. My father-in-law worked in Florence, South Carolina as the head of manufacturing. They had to do some data input next to a metal lathe that shot out tons and tons of metal shavings. IT brought down the desktop fan computer and put it next to that, and you can imagine the sparks that came out after about three minutes. So it was exciting. My father-in-law was not pleased, and they learned very quickly about fanless computers as a result. We have people that have got fiberglass shavings and everything else going on around. Temperature variability, you may think, okay, I'm in an indoor environment and I'm good to go, but I worked in a non-temperature controlled warehouse when I was a teenager. We used to joke that it was air-conditioned in the winter and heated in the summer, and I mean, that was extreme even being indoors.

35:06
Mike: But then you take that out to the top of a wellhead in a Texas desert sitting on an oil rig, and you've got a completely different perspective. So take your time. Really think about not what's gonna be the common condition, but what are the extremes that you're gonna face. Power input, AC power is not everywhere. AC power is great. You can plug a power adapter in, a conditioned brick, give you good clean power to your system. We offer that. But if you're in a vehicle, if you're on a forklift, if you are in auto, all those kinds of things, you're probably running DC power. It may not be the cleanest. So make sure that you're thinking about power protections. How clean is my power? Is the system that I'm gonna select, will it be able to deal with that kind of stuff?

35:57
Cole: Yeah, and looking at one of the most common use cases there with running it in an enclosure, you have a 24-volt supply you're using to power everything. If you've got terminals on that, you want a system that's got terminals on the other side so you can wire it up very simply. And that's something we've really gone a long way in the last couple of years of really standardizing on it.

36:18
Mike: Yeah, absolutely. Impact forces and vibration, rail, auto, forklift again. Cole and I were just joking about, we just got one of our K800 units that came back off of a forklift that somebody ran into a whole bunch of stock. It doesn't look so great. Works real well, but doesn't look so great. So think about, your folks don't always treat things super nice. If you're going to be putting this onto a Mack truck that's working in a quarry kind of a thing, doing 5G data transmission, again, what's your worst case? Chemical is one that we really don't think about a lot, but if you're in a, we've got a client that's a car wash, we've got people in manufacturing, we've got people that at the end of the day, they've got to spray down their equipment with caustic chemicals. Again, what is your typical screen going to look like after being sprayed by caustic chemicals? So these are all important things. And finally, embedded life cycle. A lot of folks are going to want to standardize on a PC that they can image, support, train people on, and it's going to be out in the field for a long time.

37:28
Mike: So really think about, in the commercial consumer space, you're looking at Dell, HP, those kinds of folks introducing a new model every six to 18 months. You're looking at more of a five to 10-year life cycle in a lot of cases. So just things to think about. I/O, I can touch on this super fast. It would be a wonderful world if everything was connected with high-speed Ethernet, not the case. When I walked into that manufacturing plant, there were still PLCs and things that were 20-plus years old. Serial connections to being able to interface with those using the Inductive software with drivers that are built-in. All that's really critical. Think about how many I/O connections you need. Think about what variety. Look for a cost-efficient way that your system provider can configure that way for you. That's one thing that we really pride ourselves on. As Cole said, we build order. We have a lot of options on the front end. If it gets a little overwhelming, just give us a call. Our sales folks will be glad to walk you through that. But really think about that modularity. Make sure that you can meet your needs at a good price point.

38:46
Mike: Regulatory compliance. I am learning more about this every day. I'm a former software designer and we didn't have to worry about these things. UL, safety, certain industries, medical, very, very stringent requirements. This is the one thing that I've seen derail more deployments than anything else, is they get deep in, and then somebody goes, “Oh, hey, you need to worry about this.” And a lot of times we can kind of scramble and reconfigure and do things like that. But I mean, Cole, I know you've run into this so many times as a sales engineer.

39:28
Cole: Yeah. Absolutely, I think one of the... It's not something that I think is often discussed. One of the things I'll hear, well, I never got that other question from your competitors. I'm like, well, it might not be something that derails your project now, but I've seen like Mike said, it's like 11th hour, you're going to have issues. If you talk to us upfront we've got a number of systems that can easily be, UL industrial IT equipment-rated right out of the box. We just gave you a custom label and a custom configuration. You order that, you'll always have a UL-rated computer. But stuff like getting into specific countries, we covered North America, Europe, Australia, New Zealand easily, and then Japan for most cases. It's not something a lot of people think about, of what does it take or where am I deploying? What does it take to get my PC hardware in there? You're trying to deploy to South Korea without the appropriate markings, it's going to get rejected at customs. So it's something to consider upfront and something that we have an entire team dedicated to helping customers solve the challenge.

40:29
Mike: Yeah, and trust me, I appreciate having that team around here because it is very, very complex. Okay, mounting, just a couple of quick words, some environments you'll be very safe to put your PC right on the back of your monitor using a VESA mount, maybe a wall mount, something like that. Industrial situations, you're probably going to have something more that looks like the DIN rail that's pictured here. Think about enclosures. Think about what's going to happen as far as heat inside that enclosure. Think about your spacing. Think about attention to form factor, dimensions, those kinds of things. You want to be able to close the doors, obviously. Things like HMIs are another thing to think about. Are you going to have your panel PC flush with the face of your enclosure? What does that mean? What does it mean in terms of access?

41:17
Mike: All those things are things that you need to think about ahead of time to make sure that you've got something that's not only usable, but also maintainable and supportable. And then finally, like Cole said, we do have some servers, edge servers that have got more processing power. You'll find those in a rack mount option, but the vast majority are going to be in your more traditional mounting options. And I think this is my last slide, is connectivity, and connectivity is huge. It's one thing to have the data, but by goodness, you better be able to get access to it. Industrial atmospheres, noise, electrical interference, all those kinds of things. Hardwired Ethernet will always give you the best results. But there's plenty of places that it's just not convenient to get wiring to.

42:04
Mike: Wi-Fi is going to be the key thing for you to be able to get to that stuff. But when you're far afield, looking at these towers and top of the well hat, things like that, cellular network is going to be the way to go. 4G today, possibly 5G tomorrow, maybe lower WAN. So we're experienced with those kinds of deployments. We can walk you through the options, talk to you about backups and failovers and things like that. Generally, it's best to not have just one way of getting to your data. Try to get that stuff in a timely, secure manner.

42:37
Cole: Yeah, and another case just to throw out there for people that are maybe not doing remote sites, one of the things that's really cool about cellular is that means that you're not on the customer's network. So if they're not going to let you on with whatever reason, cellular can be an easy workaround for that problem to solve.

42:55
Mike: Yeah, yeah, absolutely, absolutely. I jumped the next slide a little bit quick. I got to show off my two new babies. I'm not a salesperson, I don't get commission. This is just me really enjoying the fact that we can hit a lower price point as an industry and specifically as OnLogic. Our processors, we're using an industrialized Raspberry Pi, compute module and a brand new Factor 201. Really good little gateway, get access to your stuff that's perhaps not as intelligent, a little bit older, in a very cost effective way. We have an Ignition special unit for this. It's already pre-installed with edge. And then we've got something that looks a little bit similar, a little bit bigger of a form factor, which is going to give you some digital input output, some analog in, even some temperature sensor stuff.

43:47
Mike: Starts kind of getting into the area where some of the PLCs play. It gives you a lot of different options in terms of how you can interface with your equipment and your environment. And again, this is coming out soon, and it will be preloaded with Ignition Edge, ready to go out of the box.

44:04
Cole: Yeah, and while we still got a whole range of x86, like you said, Mike, and like we said earlier in the presentation, the ability to start deploying and seeing the value of data right away for a really effective price point, but still industrial is really, really cool. Talking to people at ICC, I know a number of folks are running Perspective on these and then just to connect it to a monitor if they have a mouse and keyboard, just to get started, you've got the additional functionality, you've got everything you need there. So it's an interesting way to try to get as... A lot of systems out there for a lower price point, but still reliability. Back when it was only Intel processors, that was all great, but we couldn't lower the price point beyond a certain threshold and still keep the reliability, so that's a little different.

44:52
Mike: Yeah, absolutely. Let's talk a little bit later about how we can get people started. 'Cause I agree with you, getting that proof-of-concept out there, starting to find the value of Ignition software, of hardware, those survives at the edge. That's all really interesting. So I love proof-of-concepts, but I need to turn this back over to Travis to talk about software.

45:15
Travis: Yeah, absolutely. And as there's hardware considerations, of course there's also software considerations, and that we have to look at. Really, software needs to be able to scale, and you can always scale by adding more, but when it's built-in, when you kind of think about it at the beginning, it lays that groundwork, right? For low-cost scaling, it's just about adding more to the system, and that's Ignition. It's fundamentally you want to get out of your way. So we have the compute that's there. We can get out of the way with both areas and get more data, do more functions and have more available, and you kind of got to think about the future, which we'll say here in a minute, but you all seem to figure out how and where you're going to get your data and where it's going to need to go. You mentioned the connectivity in terms of like cellular 4G and wireless and all that, that's obviously very important to the northbound, getting it out. But we got to look at the southbound, right? What are we talking to? Legacy PLCs, what protocols we're dealing with? Is it OPC?

46:04
Travis: Are we going to be sending data through MQTT? Are we going to the cloud, going through cloud connectors? What does that all look like specifically having kind of a roadmap of detail of what we're going to go? It makes for a much smoother system. And then last of course, do you have to have redundancy? How critical is the system? Do you need that redundancy or that failover? So those questions change what you look at in terms of hardware. If you're looking at it from the application side, it all goes together. I think looking at these all at once is an important concept. A lot of people come to us, look at some considerations, and then we don't, we stop there, right? And then they may go to OnLogic, have some considerations, we can all talk together about what's really needed with this. Now, but going forward, we want to set ourselves up for future success. It's a really important piece. And we talked about this quite a bit. What can the edge be going forward? The investment that I'm making today lays a foundation for the next five years or longer. And there's changing environments, there's different problems, there are different solutions that we need.

47:10
Travis: We don't want to replace systems, right? Like you guys talking about this. Cole, you showed example of having replaced things. It's expensive out there. So I know that we've looked, we've kind of all discussed about the future, but in terms of... We mentioned data collection and HMIs, but we know that there's Industry 4.0 initiatives, right? So what kind of things do you see that people are starting to think about now in terms of compute that they need at the edge?

47:42
Cole: Yeah, and to that point, and Mike, I'd love to hear your thoughts on this too, but to the point around the investment, it's you can start small, but don't shoot yourself in the foot by saying, "Well, let's just start here, deploy a thousand, and then we'll just call it a day.” The true Digital Transformation, you can do those really enterprise-wide, or enterprise-level effect that you can get out of these tools is something to consider. So if your ultimate goal is to go ahead and deploy a more complex application, consider that upfront and say, “What does that need look like in terms of hardware and software in the short-term and the long-term?” I always say start small with the PLC, but start small with the PLC. Don't do like a thousand of them and then be like, all right, that's the PLC, because you're gonna have problems then.

48:35
Mike: No doubt, no doubt. When we work with the customer on a PLC, typically we'll start with a higher spec unit, and then we'll slowly bring it down because of affordability and everything else, right? But then we don't typically end up all the way down with the customer because we want them to think about, all right, what else will I be doing at the edge in one year, three years, five years? So leave yourself some room there, because the one thing that we know is edge is not gonna get simpler. Edge is absolutely gonna get more complex. So pick your hardware carefully, and make sure that you're thinking about the future and not just today.

49:13
Travis: Yeah, especially running very complex machine learning models at the edge, we're seeing more and more need for that. So just as a consideration, and there's a lot, yeah, as we look at our infrastructures, these are the kind of questions that we get and the kind of discussions that we're having and want y'all to hear what those considerations should be. Now we gotta get to the questions here, guys, but a couple of quick calls to action. In terms of Ignition, it's never too early to try it out for yourself. You can go to the website, you can download it, and you get a two-hour trial version completely for free. Takes about three minutes to download, to install. It's very simple and easy, and you can evaluate an entire system, really see what Ignition can do for you without having to purchase anything. So it's really, really easy to get started on that front. We have a great university too to learn about how to do things in Ignition. There's lots of videos on there, very comprehensive, along with the documentation system. And you can also contact either a distributor or our account executives here at IA to learn more about Ignition. I know with OnLogic, you guys have similar ways of allowing people to get started with your hardware. Wanna share that real quick?

50:27
Cole: Yeah, absolutely. So we do have a program to help you pick out what's the right hardware for your program. So we can work with you, figure out what your project needs, help you choose that hardware, get it out in the field for you. And if you don't like it, you have a built to order system, return it. We can try something else and figure out what would work instead of that. So we really wanna work with you and find, what we never do is we never say, “This doesn't fit for our hardware, but we're still gonna try to work with you on it.” We will tell you this is within scope of what we can do. This is not within scope. If you're sending a computer to the space station tomorrow, I can't help you. But 99% of the time I can. And we really, we have experts here who can help work with you. And especially when it comes to deploying Ignition, the amount of times we deployed it is, I couldn't even count, it's hundreds. So we can help you even further tailor that hardware based on use case for sure. So the first step really for people is to, you can always subscribe to our regular communication where we'll talk about new hardware, cool stuff we're doing. But if you have something you have questions about, about hardware right now, feel free to contact myself or Mike, we can get you started.

51:46
Travis: Absolutely, and all of us are available here if you wanna speak with us to learn more about what we talked about here today. Please don't hesitate to reach out on that. Let's get to some questions here guys. There's some good ones that did come in. The first one here is, as we talk about edge, we normally make a link to UNS, so the unified namespace. Using Ignition solution, UNS environmental is already done. That was the question. And actually yes, if we have Ignition on OnLogic, edge computer, where we get access to that data, that brownfield data especially, those PLCs, we bring that in, we add context, we provide that data modeling. We ultimately make it... We standardize that and clean it, and then that we can send up. And especially leveraging MQTT Sparkplug, that is what can provide that unified namespace across the infrastructure. And we have a lot of webinars that talk about MQTT. I encourage you guys to take a look at that to learn more. In fact, we just did one a couple of days ago. So definitely take a look at that, but it is definitely unified namespace concept is starting to tack on, and edge is critical to make that happen.

52:56
Travis: Now, there's a question for you guys here. This is just kind of a simple one is that they know how system integration works for us internationally for Ignition, but what about OnLogic? How can SIs leverage OnLogic internationally?

53:14
Cole: Yeah, so if you're an SI, feel free to reach out and we actually have a member of the team here who's fully dedicated full-time to work with you. But internationally, like I was kind of saying, if you're in the United States, Canada, anywhere really international outside of Europe and out of the EU and UK, you can work with our US sales team and our solution architects or sales engineers. Those are the folks you're gonna work with to find out what's the right hardware fit. When it comes to the EU and the UK, then you're gonna wanna talk to our EU branch about that. So we can really help you out there. And we work with SIs all the time, but we don't operate on distributors, come right to us, we can totally help you out.

53:56
Mike: Absolutely, yeah.

53:58
Travis: Perfect, so another question for you guys said about TCO and reliability. Intel with vPro and AMD Pro platform offer advanced remote management interface. Do you see something comparable going to ARM or the PI platform?

54:16
Cole: So this is actually a great question. So with the built-in Intel with vPro and then using the mesh commander suite, it does offer that advanced functionality. The challenge really becomes that is only available when you go i5 or above. So when you're talking about deploying, for example, a large number of Perspective or Vision HMIs, you really only need the seller on Pentium. So you're gonna incur more cost there, or if you're deploying a really small data collector. When we're talking about managing a large number of devices, and actually, this is something that really it's a big consideration feature is, you need to look at tools that are going to scale across devices and really provide that functionality. Maybe not to the same level of out-of-band management as vPro, if you need that, there's not tools that exist today in the ARM space.

55:10
Cole: But yeah, Intel has that. I would say looking at tools that can manage containers, manage, and I'm not trying to give anything away. Maybe there's a future webinar about that at some point. But there are tools today that exist that we can talk about that we have, feel free to contact me. We've solved this challenge for customers to some degree already using partners that have been at ICC like Red Hat and OpenShift, we can solve some of those challenges today. Managing devices at scales is its own challenge and something really you've got to look at when you're talking about deploying thousands and thousands of computers.

55:51
Travis: It is a really important question to look at that fleet management or that full stack management. And it is a question that a lot of people are looking at today. And the great thing is we're in kind of a, nothing short of a technology revolution when it comes to how we can deploy these systems with container orchestrations and the various managed tools that are out there on the stack as well as Ignition has its own administration tools where you can manage all that centrally. It just makes it a lot easier. So while there's nothing specific for ARM right yet, there's other tools, right? That we could take advantage of and leverage. So definitely a good question there. So we are just about out of time here, but we definitely encourage everybody. We didn't get to all the questions. If you want to contact us, more than happy to work with you guys and stay tuned... Like OnLogic is talking about their communications, subscribe to your communications, very similar for us. We have lots of updates on social media, as well as we have a weekly newsfeed that we provide out there. And we have future webinars that are happening.

57:00
Travis: I know we've got another one from Inductive Automation coming in 2023 in January on about HMI design. So keep an eye out for that. Cole, you mentioned a potential webinar. Any shout-outs you want to have there?

57:10
Cole: Not yet. I'm going to tease that out, and we'll talk, but I think scaling is something that's not... It's so well understood in the IT space already. It's brand new in the OT space, and we can talk about future tools, but yeah, I'll keep that to lock and key because you guys are going to have to wait to find out about that.

57:30
Travis: Yeah, awesome.

57:31
Mike: Absolutely.

57:32
Travis: So keep a look out.

57:34
Mike: But the key message here is get signed up for our stuff. Follow us. We'll get those notifications out. We've got a big library of really great YouTube stuff. So a lot of good information there.

57:45
Travis: Awesome. All right, well thank you guys so much for being a part of the webinar here today, and thanks everybody else for joining us today. Hopefully, it was a very useful session. I'm going to give you guys one last chance to say what is your... In terms of edge, what is your one tip that you want to give to the audience here today? Mike, we'll start with you.

58:01
Mike: Wow. Think ahead.

58:07
Travis: Love it. How about you, Cole?

58:11
Cole: Mike stole mine. No, I'm kidding. Yeah, it's, think about it as much as you do your software stack because generally that's made upfront. Think about the hardware. Think about what it looks like and really think about what we were talking about when we say the complex challenges deploying at the edge. Make sure you're doing those generations. And we can help you do that. But if you don't consider it, you'll consider it when you have systems down in the field and nobody likes that. So that's your takeaway.

58:40
Travis: Perfect. And I'm going to say, Guys, do it. Just go do it. There's a lot of advantages. A lot of awesome applications you could build with it. Take advantage of it. Get a PLC going. Talk to all of us. And we hope to see some of the amazing things that you guys are all doing. Share with us. So anyways, thanks a lot for joining us today and bye for now. And we'll see you next time. Have a great rest of your day.

59:02
Cole: Thanks for having us, Travis.

59:03
Mike: Take care everyone.

59:03
Cole: Bye now.

59:05
Travis: Bye bye.

Posted on November 16, 2022