Travis Cox & Arlen Nipper Talk MQTT & Digital Transformation

Travis Cox and Arlen Nipper take a deep dive into the world of Digital Transformation and IIoT. They discuss the success of the MQTT protocol, the importance of education in continuing adoption of MQTT, the hurdles companies face when approaching Digital Transformation, and how a strong ecosystem can achieve a solid IIoT infrastructure. They also touch upon the importance of the Eclipse Foundation, Sparkplug Working Group and much more.

"If we're going to go on our digital transformation journey, the first place to look is 'how can I change my OT infrastructure to make it better OT? And then be ready to plug digital transformation into it.' – Arlen

Guest Bio:

Arlen Nipper has over 38 years of experience in the SCADA industry. Over his career, he has worked for Amoco and Koch Oil and then served as the President and CTO of Arcom Control Systems, and Eurotech Inc. His experience covers a broad range of technology, from the design and manufacture of embedded computer systems to complete SCADA system infrastructure implementations for many companies. Nipper is also the co-inventor of MQTT, a SCADA message transport, in conjunction with IBM, and has over 18 years of experience designing high-performance SCADA systems using MQTT and Message-Oriented Middleware. Nipper was also involved in many of the activities that have recently led to MQTT being a dominant IIoT messaging standard.

Host Bio:

Travis Cox started with Inductive Automation in 2003 and has previously served as director of training and director of support. Travis has overseen numerous successful launches of HMI and SCADA projects across various diverse industries. His exemplary work as an Ignition trainer and sales engineer has helped produce hundreds of vocal Ignition advocates who continue to drive the company's growth.

Episode Transcript:

Travis: Welcome to Inductive Conversations. I'm today's host, Travis Cox, and I'm talking with Cirrus Link's Arlen Nipper. Arlen, thank you so much for sitting down with us today.

Arlen: Hey Travis, thanks. I always enjoy this. I'm looking forward to the conversation we're gonna have today.

Travis: Now, this is your second time with Inductive Conversations and we're really excited about that. You're the first second-time person.

Arlen: Alright, I like that.

Travis: Now, last time we talked about, a lot about MQTT. The origin story, to the practical implementation of it in the industrial settings, we're looking forward to expanding on our last conversation with Don in this conversation. It seems like every year, the interest of MQTT continues to grow. As the co-inventor of MQTT, what is it like to witness its rising popularity? And why do you think there's so much interest in it right now?

Arlen: It is interesting to me, I love sitting back and watching people do all kinds of things with MQTT. From Facebook using it, Facebook Messenger, to now it comes installed as standard on Raspberry Pi for college kids to play with. But I think the reason that it's taking off is because people are finding out that it's — “Keep it simple, stupid.” It's a very simple protocol. And — I would like to say that Andy (Stanford-Clark) and I were extremely intelligent and we invented this really complex messaging transport, but the reality is, it leverages all the advantages of TCP/IP and just puts a delivery mechanism on top of that.

Travis: Right. I keep saying it's really simple and we need simplicity. There's too many protocols out there today and we gotta get somewhere that is — everybody understands that, not only the hardware and the devices, but also the applications. That they can all talk to the same language. MQTT was designed with oil and gas in mind. Have there been any applications of MQTT in other industries that have surprised you?

Arlen: Well, actually every time I look out there, again, I mentioned Facebook a while ago, and I can say, well that's one of the reasons Chris and I started Cirrus Link, 'cause we were really mad that a social media networking company actually scaled MQTT beyond what we actually had invented it for, which was industrial. But having said that, we see it in use today, all over the place. If you consider as your IoT uses MQTT, AWS IoT, IBM Watson IoT, Google Cloud Platform, SAP Leonardo, it's just becoming a very capable and consumable technology. So that when we go talk to customers, not only the IT folks heard of it, the OT folks have heard of it, and we can start having that scaling discussion on, you know, how do you transform or go into digital transformation?

Travis: MQTT, as we know, is an open standard, and has been payload-agnostic. So we can send any kind of data we want over MQTT. But for the industrial space, we know there's been a lot of development with the Eclipse Foundation and the Tahu Project and there's a Sparkplug working group, so if you can tell us more about that and why it's important?

Arlen: Well, first of all, as we said, there's a lot of applications using MQTT. Now, the beautiful thing about MQTT, is you can publish anything that you want on any topic. But the bad thing about MQTT is you can publish anything that you want on any topic. So what we did with Sparkplug, a little over three years ago, is we set down some really simple rules on, "Hey, if you're gonna use this in an industrial application, then this would probably be a really good topic namespace to use. This would be a good way to publish processed variables," so we're talking about 4-20 milliamp-loop levels, pump start-stops. So if we could just have a common way of expressing that, then multiple OEM device vendors, multiple solution providers could all use MQTT with a common namespace and some common verbs to all be able to use MQTT, but use it in a real-time mission critical industrial environment.

Travis: Right and it's all about context. You gotta have a context for that data. And Sparkplug can really provide that context.It makes it much easier.

Arlen: Well, and really, what we did if you think about it, what we did Travis, is we finally — it's like, the king has no clothes on. We finally came to the point if — last 40 years we've been using register-based protocols, i.e., I'm gonna pull for a register, I'm gonna get it back and then I've gotta give it context. I've gotta get the engineering units, I've gotta timestamp it, I've gotta to give it engineering ranges. What Sparkplug did was move that atomic representation of that process variable all the way to the edge of the network. This tenet that we talk about, single source of truth. So when I go get that tag from Ignition or any other application, even if I get it a week from now, I've got the timestamp, I've got the tag name, I've got engineering units, I have context. So, what we've created with Sparkplug is the notion of, for lack of a better name, we've defined a process variable object that we can all work with. IT, OT, cloud, time series database, SCADA. Everybody can have that same single source of truth.

Travis: Yeah, that's incredibly important as we continue going on. We can't do mappings a million times. [chuckle] It's just not gonna happen.

Arlen: No. You're gonna wear the data out before you get it.

Travis: Yeah. You look at some of these SCADA projects that are out there, how much time is put in the beginning to provide the text and the address space and the mappings have to go on there and then the context is there and then guess what happens, they upgrade equipment or things change and then we have to do all that work over again, pretty much. Right?

Arlen: Right, and it's — and you have to realize the current feeding of tags, it's not one time, it's for the rest of the life of that system. And that — we've got to eliminate that overhead. It's just not economical. I know Chevron were on the Sparkplug Working Group, a kick-off meeting that we had last week and they were saying that they found out that by using Sparkplug in MQTT they reduced the commissioning time of adding a single tank to 21% of what they used to take. And they're saying as they grow, they're not gonna get exponentially more employees to take care of their system, so to them, to Chevron, this is very important.

Travis: Yeah. And it's a way you always do this in demonstrations. How we can, in seconds auto-discover all that information that's out there with its context, and allows us to bridge our legacy, our old devices along with all the new equipment that's gonna be coming in and we're gonna plug and play with.

Arlen: Right. And because of MQTT because, you know, one of the notions that we have to think about, and I hope that people listening to this podcast still realize is that with poll response, you're always using a 100% of available bandwidth and so you only have a finite number of tags. By using MQTT if we can reduce that 80 to 95%, that's 80-95% more information that we're currently leaving stranded in the field and that's where digital transformation gets interesting.

Travis: Yeah, I know a lot of people have asked us, Inductive Automation and Cirrus Link about Sparkplug, about MQTT especially Sparkplug in terms of, "Do you own it?" But that's really the Eclipse Foundation and the Sparkplug Working Group is really, it's, the whole, the whole, that was all started because it is an open standard, it's not something that we own. Right?

Arlen: Right. We had no idea. Again, we started going to ARC Conference, with Inductive Automation. Yourself and I and Don Pearson had talked about, you know, who would adopt Sparkplug, and then it just took off. I mean it almost took off way faster than I thought it ever would. And we started getting the hard questions from very large customers, "Arlen, who owns Sparkplug?" And we said, "Oh, yeah it's on our public GitHub site." But that really wasn't the right answer because of my background and the fact that we had taken a lot of MQTT technology to open-source through the Eclipse Software Foundation, they looked like the perfect fit. Now realize, we could have gone to an Industrial Organization to these collectives or organizations, but it was, they aren't software-centric. So, ISA is not built to manage a software project. But the Eclipse Foundation are. So this is pretty disruptive, is we've taken some industrial technology but we're taking that to a software-centric organization. I like the quote that Mike Milinkovich from Eclipse Foundation said, is that, "Software is eating the world and we've got to address that." And I think that's one of the things that you asked me: “How is MQTT taking off?” Well, I think that's one of the reasons, is that software and messaging and interoperability are eating the world.

Travis: Yeah, and the idea so you would get huge adoption because that huge adoption is gonna allow a lot of products out there to work together very simply.

Arlen: Absolutely.

Travis: Obviously one big piece of digital transformation you started talking about there a little bit, is education, is the awareness about it, and of course, with the Eclipse Working Group and Sparkplug, more information there, it's gonna help get that awareness out there. Right? Get different organizations together, but in terms of Ignition I know we've been working together a lot to build a new series of videos on MQTT in particular to talk of course about what MQTT is and how that, how it works within Ignition to make that really simple. And so, if you talk a little about the importance of these videos and why we got them going.

Arlen: A lot of people, they look at MQTT and they say, "Okay, there's a topic and there's a payload and I can publish a message and then I can come over here and I can subscribe to that. And, you know, life is great." I think that's a lot of what we were seeing, is that everybody figured out using the baseline of MQTT was really easy, but then you go, "Well, Arlen, that satellite system, the network actually failed two minutes ago, did you know that?" "Well yes, we do because we have the last will and testament built into MQTT," but you've gotta explain that. So I think what the videos that you and I have done with, for MQTT, it kind of goes through the development of an MQTT infrastructure. You know, what is a topic, what is the payload, what is the last will and testament, how does the — how do clients work with MQTT servers? And we've put that in the context of how you would use that in a mission-critical SCADA system.

Travis: Yeah, I know on our side the videos that I worked on in terms of Ignition were just obviously the basic installation configuration of the MQTT modules, but also to get some of the best practices in terms of how to set it up in a secure environment and not everybody knows a lot about TLS and to create certificates, and having to manage those, but that was important. There's one thing that’s really important about MQTT is you keep on saying right, is that you didn't invent your own security problem.

Arlen: Exactly.

Travis: We should talk about that for a second.

Arlen: Yeah, Travis, that's one of those things, people come in like, "Arlen, we're trying to make a decision between OPC UA, MQTT, and we noticed that OPC UA has security built into it, but MQTT doesn't have any security," and that's on purpose 'cause remember MQTT sits right on top of TCP, it doesn't change TCP, it doesn't use UDP, it doesn't use other transports, it's a requirement. Now, 20 years ago when Andy and I were working on MQTT, that was a big decision. We could have gone UDP. We could have invented our own transport. And everybody kept saying, "Oh, no. TCP/IP, that's too heavy on the wire, that'll never work." But in hindsight, with hindsight being 20-20, it's the best decision we could have made, because now, our security model is right on top of the latest advancements for TCP/IP. And so, 20 years ago when we did this, nobody heard of TLS. Now, we've got TLS 1.2, and I'm sure something else will come out in the future and you are guaranteed that you're gonna be compatible with that.

Travis: Right. Yeah. I know the video has helped really people understand how to set it up in that way, and makes it simple when looking at it. MQTT is already really simple, and it's easy to get started with it.

Arlen: Yep.

Travis: There's some of those best practices. Some of those things that are there can really help get that scalability that people want and people need. I know you played a key role in sort of defining part of the Ignition Edge, and also with the Ignition Onboard Program. So what are your thoughts about the growth of Ignition Edge, especially with MQTT and with the Onboard program? What can that bring to the table?

Arlen: Well, again, Travis, I think about this — I was — I'm out here doing some other videos. I got to thinking that you and I got together less than four years ago when I came out here and I said, "Hey, Travis, I think we've got a good idea. It's called MQTT." And then, we said, "Well, you know what, now we've got the lightweight messaging. We can actually set out at the Edge and efficiently publish tags in. And oh, what if we took a smaller version of Ignition. We call it Ignition Edge. And while it's Java, so we can run on just about anything. It... "

Travis: Yeah.

Arlen: "Preponderance of OEM devices. It already had embedded Linux." So all of this Ignition Edge notion popped out. And now, we had everything that we needed for Ignition Edge. And to me, what that means is we've made a disruptive step in having software that can run on multiple OEM devices, out on the edge, that's all the same, that works the same way, publishes that information in. And now, an OEM can implement Sparkplug himself or they can go with Ignition Edge or they can do both. And I think what we, Ignition and Cirrus Link have — or Inductive and Cirrus Link have — brought to the table is kind of the perfect mix of being able to put, basically, a buffet out there for the customer, and let him pick best-in-class hardware, network, security, applications, and really get him all the tools he needs for digital transformation.

Travis: Yeah. We, I think we both see the Ignition Edge and being able to move, especially, move pulling out to the Edge number for legacy devices really having a distributed network, be a real key part of an architecture change that's needed for digital transformation. But in terms of digital transformation that we're talking, it's a hot topic right now ...

Arlen: Yeah.

Travis: … In the industrial space. We certainly are doing a lot of education and awareness about what we think digital transformation is and what's required for it. But you talk a lot, you do a lot of presentations about digital transformation and — but that's the kind of first frame it is and give the people listening, sort of context right now in terms of how the digital transformation is being presented and what that means to the — from an IT perspective to the OT world.

Arlen: My notion, Travis, is we talk to customers all the time, and I think we're a little bit in the forest for the trees. I.e., there's executives in a lot of big companies, there are IIoT executives now, at the C level, and they have conversations with cloud providers or — what they could do, and they wanna get started on their digital transformation. But I'm here to tell you, that's the wrong place to start because — Here, I'll give you a quote. I came up with this quote for ICC this year. And what I said was, "Ignition-based digital transformation uniquely addresses the fundamental issue that current OT infrastructures simply are not conducive to the requirements of digital transformation." What I mean is that you can put lipstick on a pig, but it's still a pig. And the problem is we have these legacy infrastructures that are poll-response, that have been there, 20, 30, or 40 years. So the first — if we're gonna go on our digital transformation journey — the first place to look is, “How can I change my OT infrastructure to make it better OT?” Full stop. And then be ready to just plug digital transformation into it.

Travis: So you're saying, we shouldn't, from an IT perspective, and just say, "Oh, yeah. No problem. We can add that data. It'll be really easy. You know what we're gonna do? We're gonna write some custom code, some APIs we've put down, locally. We're gonna, hopefully, tack it to the SCADA system that's there, and not harm anything, of course. And hope that all the data that we want is actually there." So you're saying that's not the right approach?

Arlen: The first project will work perfectly, and then, somebody will see that and they'll go, "Hey, you changed that for this consumer in the enterprise. I want you to do this." You change it a little bit more. And then somebody else goes, "Oh, will this enable mobile?" and you change this. And I've seen it time and time again, is you're gonna take a perfectly good SCADA system and infrastructure. And it'll become so customized, it'll become brittle, and then, nobody will wanna touch it, and innovation will stop.

Travis: Absolutely. Yeah. We have to transform that OT layer. OT has to think of digital transformation in the same light that the IT professionals are.

Arlen: Absolutely.

Travis: There's a lot of money for digital transformations at the C level. They wanna do some amazing things with the data: machine learning, and they want to do analytics, and obviously get a lot of wisdom out of the data that's there, but they're neglecting to realize that the majority of the data they need is, where? It's at the OT layer. And we fundamentally have to get OT systems to kinda think like IT in terms of being open and standard and being able to be interoperable, but of course have it work for the OT world as well. I know you talked about that a lot. We can't — we have to get digital transformation but we have to keep our SCADA systems running and have superior systems.

Arlen: Well, I've been on numerous projects now, for probably the last two and a half years where customers have started down the path and they've taken the IT approach in putting, like you were saying a while ago, "Well, let's put an edge of network device there, and we'll write some JSON or we'll write some Python code to convert that to JSON to publish it to one of the cloud services." Again, in a one-off experiment, that's gonna work fine. But we've gotta think more in terms of tools on platforms, not coding on operating systems. And I have seen, I've been doing this for 42 years now, and I've seen a lot of coding on operating systems and it never works. I guarantee it will never work. So when I go to someone like Amazon on some of their projects wanting to get data into Amazon, into S3, or into Kinesis, or into DynamoDB, or into their SageMaker. If they look at it from writing lambdas with Python code, they really understand that's not scalable. But if I've got a tool like Ignition, so I've got my Ignition designer, I've got my tag management, I can do UDTs. I can put context to those tags. Now, I have a tool that I can give to any engineer on the plant floor and he knows how to deal with it.

Travis: So in terms of digital transformation, now, let's talk about what is needed in order to really achieve it. What are the major steps, if you will, that companies need to be looking at an in terms of digital transformation?

Arlen: Don has said this before, (citing the Eric Eric Schaeffer book) Industry X.0. There's numerous Gartner studies that have said you've got to have — there's no one company that's going to be able to do digital transformation. In 2019 going into 2020, no one company is gonna be able to do this. This is where we're gonna require an ecosystem from — again, we're talking about the Industrial Internet of Things, that those things are devices that OEMs are manufacturing. Those things need to talk into infrastructure to be decoupled. That infrastructure needs to be plugged into by solutions that understand the data that they're getting in full context. Not editing a Modbus 40012 14 times in 14 different applications. So really, to me, Inductive Automation provides the tools, technology, and ecosystem through the Onboard program required to modernize existing OT infrastructures. And that results not only in superior OT or SCADA solutions, but an infrastructure that's now ready to address your digital transformation requirements both today and wherever you wanna go in the future.

Travis: Yeah. The decoupling you mentioned, we know it's so important to be able to decouple the applications into devices. And of course we do live in the brownfield world. There's a lot of legacy devices out there. At the same time, you're working very closely with a lot of manufacturers and new sensors and equipment, getting MQTT ready to go on that, on those devices so they can plug and play in the infrastructure. We've got these two worlds, we've got the old and the new, and there's a lot of focus on it. Of course we could bring new sensors in and all that, but we really have to make sure we can — we're not just putting a band-aid on the old stuff. You gotta fundamentally bring the old, the legacy, into the same infrastructure.

Arlen: Believe me, you and I both, we would love to — we would love for the brownfield world to never have existed. Give us a nice clean shiny slate and we'll design the best digital transformation system in the world. That's not gonna happen. There may be a few cases, but 99%, we're gonna have to address the existing applications, existing equipment, existing protocols and do that in a way that gives us a migration strategy to where we wanna be.

Travis: Right. And I think the Ignition Onboard program is gonna be a real — one key way to do that. The Onboard program is combining the hardware, the IPCs and panel PCs and networking equipment and anything that can run all computing and it'll basically have Ignition pre-installed, pre-licensed, ready to go from the manufacturer, so that with the Ignition Edge MQTT product on it, so that we can easily take data from those PLCs into this infrastructure from the legacy world.

Arlen: Absolutely, and not only can we do that from device to cloud. Again, I'm seeing more and more customers looking at edge-of-network IT. So, the IT edge versus the OT edge. But the IT edge, again, is not set up to be able to handle all of the aspects of — let's go back and call — let's be technical here for a little bit, what's called a process variable from a sensor in the field. Let's say it's a 4-20 milliamp loop, if zero to 4095, we gotta scale it. We've gotta get an engineering unit. All of those things that APIs just don't give us, but tools give us. And with Ignition we can show that, my brown field to cloud in a 30-minute demo.

Travis: Absolutely, and where a lot of Onboard partners that are being in the program, companies like Advantech and Moxa and ORing and many others, Opto 22 and many others, that is gonna make this really approachable because you could just simply go to the manufacturer, get an equipment that already has everything ready to go, you don't have to go and get the hardware, get the software separately, install it, go and configure it, tune it. All of those things. It's just ready to go, if you have... If you want strategic transformation, doing the first couple of pilots is no big deal. But it's that rollout, it's that getting everything to be brought into the infrastructure as you go along. That can be challenging. If we can reduce any of those steps, we're gonna save people a lot of time.

Arlen: Absolutely, and I think with organizations like Eclipse Foundation and Tahu, I think with ecosystems like the Onboard program, I think with solution providers understanding the value of Sparkplug. So I'll give you an example: Canary Labs is a time series database. And typically, in a time series database is just like in SCADA systems, you had to go configure every tag that you wanted to show up in your time series database. Well now, we can have an infrastructure with Ignition Edge, publishing tags, and then take an Ignition Gateway in a Canary Labs time series database, plug them in, in a few seconds, they know all the tags and all the engineering units and all the parameters.

Travis: Yeah, plus we can then have IT who wants to play around, get access to some of that data as well. All the applications they wanna be brought in or if we wanna experiment and play with some new systems that are out there, they'd plug it in we can start looking at them.

Arlen: Absolutely. It's the serendipitous nature of data. And think about, there's no way with existing legacy OT infrastructures, that you, Travis, could come to me and go, "Hey, Arlen, I just had a really good idea. If I could take these tank levels and these temperatures and then bring in some market information, I could do this great application." And I will go, "Well, Travis you're gonna have to go to OT and ask for an additional poll to be put into the poll table to get that piece of information and you gotta go through and then we'll have to re-do a point check and — well, Travis, maybe six months from now we can give you that piece of information." By that time, you're going, "I'm not interested."

Travis: It's gotta be simpler, right?

Arlen: Exactly.

Travis: So, that brings us to end-users and who have to — obviously, those are the ones that have to go through and adopt these and put these in place. But what have you seen in the past year as a progress towards digital transformation towards this? What has it been like?

Arlen: Well, from Cirrus Link’s perspective, just to even look at the adoption of Sparkplug in the sales that we've had in 2019, it's incredible. I mean, the adoption is — I could say, almost exponential now. So I think people are really starting to understand this, they're starting to leverage the capabilities. And it's two-phase. One of the things we were talking about is this serendipitous nature of data and being able to plug things in and unplug them. But one of the things I want to bring up is that even from the standpoint of “forget cloud, forget predictive maintenance, machine learning, think of a system where you never had to do a cutover.

Travis: Yeah.

Arlen: Right? So now you've got your production system plugged in, you've got tests and dev plugged in, but they're both looking at the same tag that's coming in from MQTT Sparkplug. So starting from there, and then being able to expand out. Now, I've got OEM devices, I've got much more exponentially more efficient network infrastructure. Now I can bring more information in reliably and now I can bring that into multiple applications, satisfy all of my requirements for mission-critical, for real-time, and then be able to take that information and literally plug it in the cloud applications. So I think for me and for Cirrus Link, the adoption has just been incredible in 2019.

Travis: That's great, 'cause you know — how many years we've been working together now? Three, five years?

Arlen: Well, no.

Travis: Four or five years?

Arlen: It'll be four years.

Travis: Four years. So we know we did a lot of education, a lot of work...

Arlen: Exactly.

Travis: To get people to understand the possibilities. I mean, when you came to the office that day and talked about MQTT and you drew up the architecture, I was sold in two seconds, I knew exactly what was going on. Coming from the IT world, kind of you can come to the OT space be like, "Why the heck do we do it that way?" And those are really funny — I think a slide you saw, they had a reference to a movie. It talks about if you could give her the idea of poll-response protocols. Why don't you talk about that?

Arlen: Okay, so actually this was taught at Chevron and he was talking about why Chevron are looking at adopting MQTT Sparkplug. And the first slide was Men in Black, right? It has Will Smith holding up the nebulizer, the minder eraser. Imagine you clicked that and you didn't know a darn thing about SCADA or protocols, you didn't know about Modbus or HART or Allen Bradley DF1 or DMP3.0 and you — but you were a very knowledgeable IT (person) and somebody had this piece of equipment, had all of these analogs and statuses and information flowing in, and you plugged it in and you said, "Oh you've got to pull it and ask it for information." And you go, "What? You have to ask it? This is 2019."

Travis: Yeah, that's really funny 'cause we saw that potential way back and over the years, we have been able to see the growth of that MQTT and see the adoption of MQTT and now, I think, it's being adopted, being used versus being an idea.

Arlen: Right. And that was very important. You have to realize, and I've joked with you about this before, right? Is that I think Andy and I did a very good job because we did keep it simple, but the problem was with IBM. So for the first 10-12 years of MQTT, it was sequestered to whatever you could buy from IBM. There wasn't a lot of MQTT servers out there. You know, you couldn't just go run it on a Raspberry Pi. And then the Ignition platform came along. And again I had been to all of the other Ignition competitors out there and nobody got it, nobody got it. But when we got a platform called Ignition and we could take this notion of what we thought MQTT could do and start writing the modules to plug-in MQTT Engine, MQTT Transmission and Distributor, then all of a sudden it was a perfect marriage of a platform that had the capabilities and a technology waiting to solve a problem.

Travis: And look at where we are. And I think there's a lot of really bright future in terms of MQTT because, again, we're going back to what you said earlier, the “keep it simple, stupid” model, right?

Arlen: Absolutely.

Travis: It's easy for people to adopt and be heck, right? Like you were saying, there are students in dorm rooms that are developing on this and they're the next generation to come in. They need to understand this data or we're gonna have some problems.

Arlen: Oh yeah, absolutely. I think it's the — that is one of the issues, is the grain of the engineering that know how all the existent systems work and the guys that are 55-60, like me, 61-62, that's where the knowledge is right now, but all of these students coming out of college they are sharp, they are bright, and they wanna apply their knowledge. And I think with this sort of — with Sparkplug, with MQTT, with the really leveraging TC — now, come on, the future is going to be TCP/IP. It won no argument. We're done. So now let's figure out how to leverage TCP/IP get it into an industrial environment, get that infrastructure done correctly. So now it's IT-ready going forward. Let's give OT the tools, the technologies to be able to execute what they're really good at and everything else will just fall out.

Travis: Yeah, I'm really excited about the potential. And once we do that, once we can solve the OT layer, what’s sort of amazing is we do with that data in terms of all of the great AI machine learning and the analytics and where predictive analytics or tuning a process automatically through getting that streaming, that live data whatever it might be. There's a lot of opportunity, and I think customers will really start getting benefits out of that.

Arlen: Well, to that, and I think one of the interesting applications. Look at Magnetrol. Magnetrol manufactures smart level transmitters and they've got a blue tech box that converts their heart protocol to Sparkplug B. And you know, from a customer standpoint, they just really are worrying about really expensive product and tanks, they're going up and down, but by using MQTT, and opening up the other 600 process variables that are in our smart transmitter. Magnetrol can go in there and help customers tune their level transmitters and make them work better. And there are many, many examples in there. Forget about MQTT and pub-sub and all of that. Customers showing one, 10, 12 million ROIs on better level measurement. So what I've seen, although a lot of our customers have grand plans for what they wanna do with cloud. What I've seen the biggest eye opener is for customers that are starting to use our EFM protocols and they're getting instead of getting hourly data, they're getting minute data, their operators are blown away, they're going, "Oh my God, I didn't know that was happening.” So I would say, I would agree with you, we're gonna do some fantastic things, but getting visibility of more data quicker, I think that's almost the biggest thing right there.

Travis: Yeah, like you say, what, almost 90% of data's out there stranded. And I think you're right, the first thing you gotta do, of course is get out that data is to able to — there are probably some amazing things you can do with it, just by having it and at faster rates.

Arlen: Exactly.

Travis: Absolutely.

Arlen: Again, not to beat on a dead horse here, but making it humanly accessible. So I'll give you a good example because of Ignition, I can do an expression tag, and you know this. In a flow computer, there's probably 50 or more enumerations. In a PLC, there's probably hundreds of enumerations. So oh, well, Travis, I just got a value of seven. And that's an enumeration that could be stainless steel or mono, or carbon, you know, so, you're on the phone, go. Well Travis, what brand of PLC is this? And what's the enumeration for stainless steel? Oh, that's a zero for this brand, a flow computer, but it's a three for this one. My point is, is that because we have Ignition Edge, we're able to take all those enumerations and make them humanly readable. So by the time the IT guys get them in the application it says "Oh this is your reference material, this is your atmosphere. Calculations are using AGA3, or whatever. So now you've got context, you're not having to call 15 different people to get context to a piece of data.

Travis: Absolutely, well said. Well, Arlen thanks so much for being on this conversation. Really appreciate it. Always love hearing insights you have. And of course I've seen where we can all go with MQTT.

Arlen: It's been great.

Travis: Thanks.