Podcast #103: Industrial DataOps

Good day, and welcome to episode 103 of our Digital Leadership Podcast, produced by, for, and about digital industry leaders. Today I’m pleased to introduce Tony Paine, Co-founder and CEO of Momenta Ventures portfolio company HighByte, a company providing industrial DataOps.

For the past 20 years, Tony has immersed himself in industrial software development and strategy at Kepware, most recently serving as CEO where he led the company through a successful exit to PTC in 2016. Tony has contributed to a variety of technical working groups, helping to shape the direction of standards used within the automation industry, as a strong advocate for STEM (Science Technology Engineering & Manufacturing), he sits at the Dean’s Advisory Council for the University of Maine College of Engineering, where he also provides industry insight in evangelism related to technology education.

Welcome Tony, it’s great to have you on our podcast.

Thank you Ken, that was a great introduction, it’s a pleasure to join you today to talk about my industry background, as well as what I’m working on today as part of HighByte.

Well, thank you for thanking me for the introduction, this was obviously verbiage that had come from you, but well-written and actually a great background. So, let’s start with your professional journey, tell us a bit about your background and how it has informed your views of digital industry.

Well Ken, I’m going to go quite a way back, from my early childhood I’ve always been passionate about how technology works, specifically around the Nexus software and hardware integration. I’m young enough where I don’t go back as far as vacuum tubes and punch cards, but I do remember being quite satisfied with my first computer that had about 64k of RAM, 4 MHz processer, and just two 5 ¼ inch floppy disk drives. I remember when I later upgraded to a system with a 20 MB hard disk drive, I thought that was a game changer, and that was during my middle school years. I knew right then and there that I wanted to go into a profession that would drive technology advancements with a focus on software.

That passion led me to pursue a degree in electrical engineering where I focused my attention on the overlap of computer hardware and software design, as I recognized that one would not be able to exist without the other. When I fast-forward to my recent years, industry has continued to evolve year over year at a very accelerated pace, this year’s technology outperforms what was available last year, and at less cost. The digital industry consisting of customers and vendors is continuously improving, companies are becoming more agile and digesting a new way of doing things at faster rates than ever before. Software has been a critical component to this, because it gives you the ability to upgrade capabilities, and give hardware platforms such as legacy distributed control systems, programmable audit controllers, and today’s IoT appliances, a much longer lifecycle. I’ve actually enjoyed being part of this evolution.

That’s great. So, I may date you a little bit, but I started with punch paper tape, and 8-inch floppy drives! I can certainly relate, yeah. So, you chose to focus very early on what I would call the non-sexy part of the overall technology stack, but very critical space of connectivity software. What inspired you to do so?

Yeah, that’s interesting. So, when we started Kepware in the mid-90’s our focus was on creating essentially a windows-based human machine interface, an HMI product, but it was already at a time where it was a growing and crowded marketplace. There were many new software entrants that also saw an opportunity to leverage the off-the-shelf PC and phase out proprietary hardware and software-based HMI solutions of the past, and like any company we sought out how to differentiate ourselves from the competition. That’s where I kind of leant back on my background, along with another early day colleague’s experience a lot of stuff focused on creating a large library of device drivers, focusing on that passion around software and hardware integration.

What happened was, eventually we caught the attention of some of our HMI competitors, none who were threatened with our very basic HMI capabilities, but there was interest and excitement on how they might be able to leverage our large Kelsey based drivers within their own products. We stepped back, we evaluated the opportunity, saw that it made sense for us, and began to pivot by providing external access to these driver capabilities that we had created. We had done that through at the time, vendor specific toolkits, initially based on Microsoft’s dynamic data exchange technologies. The two that come to mind are Rockwell’s Advanced DD toolkit, and Wonderware FAST DD toolkit.

What’s interesting is, during this time another effort was brewing, there were companies getting together and working on an open standard for device connectivity, based on Microsoft’s latest technology at the time, object linking and embedded technology, also known as OLE. This was actually OLE for process control, or what’s become known as OPC. OPC really opened up the door for us to focus on our particular niche, that we were not only passionate about, but we were really great at, and so we decided to transition our company’s business model into one of more collaborating with partners, rather than just competing in this crowded HMI space in isolation.

It’s funny, this is all coming back. I was probably the guy at Wonderware at the time, I was certainly in FAST DD I attended the initial OPC kickoff meeting by Microsoft way back then, I think it was 1994, and representing Wonderware at that. So yes, FAST DD I know well, as well as OPC. It seems with the early traction there, that the industry however has moved somewhat slowly in terms of truly linking IT and OT. Why do you think that is?

Well, I would say some of those technologies that we just talked about, DDD, OPC, they provided a way to share data between applications, regardless of the vendors who actually developed those applications, or created the devices. This occurred because industry got together and agreed on a standard application programme interface contract, that could be leveraged by anyone. So, at the time these technologies were heavily tied to Microsoft, first there was the DDE stuff, and then there was the OPC stuff based on OLE. Microsoft continued to rebrand that technology, it went from OLE to the Component Object Model known as COM, and these technologies were great for sharing data between applications on the same machine, what we refer to as inter-process communications, but it didn’t take long for the market to demand that this technology work across various machines, so very much of a distributed manner.

So, on the DDE side, a technology called NET DDE was born, and it was an effort started up by your old company Wonderware, it actually was eventually included as part of the Microsoft operating system stack, for people in our industry to use for COM, Microsoft had an answer for that requirement, it was called Distributed COM, or DCOM. What I would say is, with remote capabilities obviously comes the need for a strong security model, and as we all know when security gets involved it usually becomes complex, and the complexity wasn’t just for the vendors developing the applications, but it was for the customers who leveraged the applications based on these technologies, and so when you need computers to be networked together, IT is going to get involved. Of course the security model of these technologies, Net DD and DCOM, they really weren’t on par with the security models deployed by IT higher up in the enterprise, I’d say it’s definitely probably not the only issue, but it definitely was a pain point, the setting up of DCOM, that was a notable point in history to where operations and IT staff really started, or as we like to joke, were actually forced to communicate more frequently.

I’d say that over time IT has become much more involved in operations, the control system networks, and the ongoing maintenance to ensure that the organization is secure over time. So, I’d say that I’m personally glad to see that IT is having a much bigger voice today than it has in the past, because it allows them to contribute to areas where they have much more experience that operations can benefit from.

I remember doing my first Wonderware implementation, I was at Lockheed Missiles and Space, control systems engineer, we were running on a Beta version of Windows for Workgroups which included the Net DD you mentioned, Net DD component. So, you could imagine at a military defense contractor, the conversations I was having with the IT people about why I wanted to roll out basically a peer-to-peer network there! It was not fun, but it worked phenomenally well.

And you were relying on the same technology that Microsoft used to be able to allow people to play Microsoft Hearts, the game Hearts, across the network.

Yes, true, or monetary printer settings! So, the acquisition of Kepware by PTC created a bit of a panic in the OT industry, I know because we were out there talking with a lot of the companies, especially companies like GE and Cisco, which started to look to hedge their connectivity software bets, you guys were and are ubiquitous. So, I always thought it was an interesting move by PTC to acquire you guys, in fact one of our companies at the time, PLAT.ONE, we actually sold of course to SAP that’s now Leonardo, but a large portion of the valuation in that was because of the software libraries that were underneath it. One of our companies now, Litmus Automation, also puts a lot of energy around the driver library, they have the edge if you will. For this area being so critical it just seems kind of strange, there’s still not more general industry options in this. Why do you think that is?

Yeah, you’re absolutely right, connectivity is critical, and the options are pretty sparse. If I go back in time, I really believe that many companies thought that the vendors who produced hardware, the VCSs, the PLCs, CNCs and so-on, would be the best company to actually develop a connectivity solution for their own products. In fact, I sometimes joke with the past president of the OPC Foundation that they never thought they would actually enable what is referred to as a new cottage industry, for software companies to exist and thrive around data and device connectivity.      The belief was that if industry came together and agreed on the standard contract for sharing data between applications and devices, then software application vendors would implement to the standard, and device manufacturers would create the one and only OPC server for their product, and all interoperability problems would just go away.

What I would say is, in fact many device manufacturers actually did create their own OPC solutions, but we never saw that as a barrier at Kepware. We actually believed two things, that customers would leverage devices from multiple vendors, and the second thing was that not all OPC implementations, regardless of them being a standard are created equal. And so, us and a couple of other what I would call OPC only vendors, vendor agnostic if you will, took hold of the market and became the de facto standard for obtaining device data consistently within an operations environment, regardless of the devices that were being connected to.

What I would say is, before we knew it, meeting hardware manufacturers started endorsing our solution, they wanted to resell it, white label it, and no longer progress their own third-party connectivity initiatives. So, as you mentioned earlier, device connectivity is not sexy, but it’s critical, and without it applications like HMI-historians, Line Management Systems, and even todays IoT dashboards would provide limited value to their users. Even so, the competition in the vendor agnostic connectivity business was really limited to a few players.

And again, lots of credit to PTC. At the time Ben Tao who is now our Head for Private Equity & M&A within Momenta, was Head of Corp Dev there. I told him, I said that was one of the best acquisitions you guys ever made, outside ThingWorx of course! But pretty close.

Look, all of this has culminated in your founding of HighByte, a name I love by the way, which you describe as industrial DataOps. So, what problem are you trying to solve, and why?

I appreciate that you like the name, it took us a little while to come up with something creative that we thought would resonate with what we wanted to do, and we could continue to grow into. I’d say technologies like OPC have done a great job of solving the data inoperability problem, that is, how do we generalize device connectivity, collect data, and then share the results with various applications. This has worked really well, but it makes one big assumption, and that’s that the data once available in the application actually makes sense to the user. So, for a simple example, if I pull a value of 75 from my device, share it with an application, and the user sees that value of 75 on the screen, what does it really mean? Is it a temperature? Is it a utilization rate? Is it a level in a tank? Does it have units of measure? And then is the value actually acceptable, the value of 75 within the appropriate operator range for the particular parameter that it belongs to. And so here the data really lacks context in this sort of legacy format.

The problem is further compounded when units of measure for a particular piece of data differs by application or sight. For instance, something as simple as temperatures where the US plants have temperatures in Fahrenheit, we just will never change and conform to the rest of the world I guess. Whereas, everybody else in the world uses Celsius. So, when I look across all my plants, is my data consistent, has it been correlated to a steering measure? This is just what I would call the basics around data, interoperability across multiple sites. I’d say that more advanced is being able to monitor and analyze assets, processes, systems, other complex widgets that are made up of multiple data points, maybe hundreds if not thousands of data points in some cases. So, here the end user application has to actually model one of those things and wire up all the data points.

That works okay if you’re only using one application, but if I have a mix of operational tools like HMIs and historians, and now newer IoT-type tools like analytics and machine learning, I have to model the same thing multiple times in all my applications, every vendor of all these applications have their own way of creating models for consumption. So, I say that’s where we enter DataOps, a new layer in the industrial technology stack that abstracts away modelling, the piecing together of data points into a sophisticated model that’s standard at the context that I mentioned and correlates the data across the entire enterprise. And so, from a maintainer ability perspective, a change to the DataOps layer is automatically propagated to all my applications. This allows for a faster time to value, it allows for much more streamlined maintenance as the environment involves, and these are the problems that HighByte is working to solve.

It brought back memories when you talked about the Fahrenheit to Celsius, you may remember back in – I want to say ’99, the Mars Climate Orbiter crashed into the surface because a simple misinterpretation of what the acceleration values were, they thought it was English units and it ended up being metric, so they could have used you way back then. Is this another form of digital twin that you’re talking about?

We would enable a digital twin, so we ourselves probably aren’t going to go and create a visual representation to be able to mimic a real-world device, because there are a lot of folks out there that are great at doing that visualization, and graphics intensive processing. However, we see that we can actually go and populate those models, and perhaps by working and partnering with companies who are creating digital twins, that may have their own say modelling schemers, we could automatically generate those on behalf of the user and just simplify configuration.

You chose an interesting time to create this start-up, it was founded in late 2018 as I remember. As you know, the World Economic Forum I’ve mentioned several times in these podcasts, has used the term, “The Great Reset”, referring to this long term impact of the COVID-19 pandemic, what do you see is the impact of this reset I guess, generally on the future of industrial software, maybe more specifically on HighByte?

Yeah, I know, COVID-19 has actually been really interesting. As you mention, we founded HighByte in August 2018, about a year later we launched a data program to start getting customer feedback, discussion with partners going, and then we released our first product, the first version of the product at the end of this January, and looking back when we founded the company we would never have guessed that today we would be in the midst of a global pandemic, so shortly after launching our first product into the market. We like many other companies expected to travel, hit the trade show circuit, meet with partners and customer prospects, and those plans just changed overnight.

Customer products were put on hold, the priorities have shifted to keeping visiting installations up and running with minimal staff, and like any product in the manufacturing space that needs to run 24/7 and be reliable, customers require time to evaluate our product in the lab setting, then move it to a pilot phase, and then eventually if the solution actually meets their needs, roll it out over a period of time. And so COVID has put these timelines behind schedule, its impacted our customers getting real hands-on time with our offering. That said, what I would say is, ‘The Great Reset’ is only putting more emphasis on being able to run all the company’s plants from any location, it’s putting more emphasis on being able to run smarter and make predictions before failures occur.

So, this is resulting in companies accelerating their digital transformation timelines, which involves the integration of these new applications with the old. And of course, in turn that will increase the need to address the very problems that HighByte has set out to solve with its DataOps technology. So, we see it as an opportunity to be there and help our customer base rapidly adopt, and digitally transform hopefully at faster rates than they had originally anticipate.

I know you were part of some of the CEO roundtables we put together shortly after the pandemic first hit, and the shelter and place orders were given in a lot of the geographies, and one of the themes we generally saw out of those roundtables was that – at least early-on, and it certainly has proven that during the time our general investment thesis of remote asset management has actually played out well, in the sense as you say, people are looking to remotely access, operate, maintain their equipment. Obviously in a lot of cases service personnel couldn’t get out there so, where it has generally been tough on the economy, at least a sector of it, i.e. this sector seems to be benefiting from it as crazy as it sounds! In terms of the fast-forwarding of, if you will, the new normal. What has been your largest hurdle in your entrepreneurial journey?

Well, I want to stay away from COVID on this one, obviously, that’s been interesting but…

That wasn’t a leading question by the way!

Yeah. No, actually I kind of looked back and said okay, one of the largest hurdles that actually occurred for me anyway, was at my last company. It really had to do interestingly enough, with the company outgrowing a particular senior leader on the team. This person, and I kind of mention them up above, was instrumental in helping us get the company off the ground, working with me on creating a wide range of device drivers for that early stage HMI product, and really helped us to move from start-up phase, to being able to generate enough revenue to self-sustain and grow the business.

Unfortunately, what happened is, at some point we became a certain size where more leadership was needed across the company, and I’d say both horizontally and vertically. This senior leader really struggled with that, wanted to be involved in every decision, and eventually became an impediment for others getting their work done. I had a strong relationship and friendship with that person and I really struggled with parting waves given the history with their company, but it was a necessary decision to move the company forward. Looking back at it we ultimately became better, more well-rounded, because we created depth by giving people a longer leash to actually go and innovate, and a better organization because of it.

That’s an interesting one. As investors we look a lot at the make-up of the founding team, and one of the first things we look for is the willingness to step aside at the right time, i.e. what got you to a seed funding or a Series A funding, won’t necessarily get you to the Series B funding. And even as a business ourselves at Momenta we’ve had to go through the same types of things as well, you get the builders, and then you get the – call it scalers, and then you get the maintainers, and those are generally different people, different mindsets and different abilities in that regard, so that’s quite astute. I guess in turn, what advice would you offer to aspiring entrepreneurs looking to emulate the Tony Paine school of start-up management?

Well, you gave me a great leading with your feedback on my last answer. I’d say there’s really a few things. The first is to look at problems in the market that you’re really passionate about, and are best positioned to solve, because that’s where you’re really going to go and make a difference. When you can live and breathe, developing the best solution for a particular market problem, people see you as sort of a domain expert, that’s just a great place, it’s a fun place to be.

I’d say, don’t do it yourself, surround yourself with people who are smarter than you, and spend as much time asking questions as the leader, and learning rather than just doing. So, you don’t need to solve every problem by yourself, and in many cases someone else will probably have a better idea.

Lastly I would say, and I think this is pretty common today in our fast paced agile world is, if you’re going to fail, fail quickly, figure out how to use that new knowledge that you gained, and pivot as necessary to basically put yourself in a better position for success.

Agile, I think we all like to call it.

Exactly.

What I also love about your background, and it was right in the bio, is you’re very well-rounded, and the work you’re doing in STEM I really appreciate, i.e. giving back. They say your professional life is divided into three phases, there’s the learning, earning, and returning phase. I’ve found a subset of people can figure out how to do both of those together, “Returning with return”, I like to call it, which is why I like to be in Venture Capital. But I like the fact that you’re doing that already in your life. Tell me a bit about the work you’re doing, and some of the thoughts you’ve got relative to stem education, especially in the US.

STEM is actually really important to me, and when I look back at my career, I was actually lucky and fortunate to have started tinkering around with technology at a young age, coming up with some ideas of what I really wanted to do in life, before I actually mapped out my academic career. When I look back to my time in school, I can’t tell you how many times there were students that were in engineering classes where they really didn’t know what they wanted to do after they graduated, they didn’t really understand why they were taking some of the classes they were taking.

So, when I got into the workforce and started becoming a professional, and getting to the point where we were making enough money where we could start giving back, I tried to connect the dots and said, well if we can get kids to get interested in technology, and start thinking about some of the things that they want to do, then they’ll have a much greater appreciation as they get through their academic career. Learn those difficult topics for which everybody once in a while wants to say, “Why do I need to know this?” Hopefully they’ll have more of an appreciation.

So, what we did with Kepware, we did a few things, we started sponsoring a local elementary school once a year a $10,000 grant to the class that would come up with a great way of taking technology and incorporating it into the classroom. So, we had companies that basically were in rural Maine that did not have computers, they sometimes would go up to the middle school and get 15 to 20 minutes of computer time, and many of the kids didn’t have computers at home, given the community they lived in. So, we actually ended up sponsoring and providing them with a variety of laptops that they could use to incorporate another classroom. There’s been robotics and there’s been other stories.

But for me when you go in, you basically go see those students’ months after they get access to whatever we help provide them, and they can incorporate in the class, and see how excited they are about going to school. Literally I’ve had students say, “I used to hate coming to school, and now I love it because I get access to technology, I get access to being able to play with robots and learn how things work”, its great satisfaction. And so, I’ve done that there, we’ve done it under scholarships and internships where we can get kids into the professional world before they graduate, so again they can get that real hands-on experience. I feel if we can just turn a handful of people into better engineers then we’ve done a great job for the market.

Excellent, what a great way to pay back your own inspiration, but more importantly pay forward the next generation of science and technology leaders out there. So, in closing can you provide any recommendations of books or resources that inspire you?

Yeah. I’ve done quite a bit of reading lately, but one book that I’ve actually read is, “Measure What Matters” by John Doerr. John the author does a great job of introducing what’s called an objectives and key results system, or short for OKR. What I’d say is, objectives as you know are concrete, they’re action oriented, they’re what really needs to be achieved, whilst the key results are the specific steps to meet those objectives; they’re measurable, they’re quantifiable, they’re verifiable. By leveraging OKRs, a company can clearly articulate its goals and expectations to the entire organization, if they can put them on a dashboard where people can access them on a daily basis, they know exactly what’s the most important thing, and how they fit into that big picture.

So, we’ve begun leveraging this at HighByte, and we see it as a great tool to get faster alignment as we grow and incorporate new people into the company. It’s also been great a great tool as we’ve stood up a new board with some external representation, to give them a better understanding of where we fit in the market, and the things we’re trying to achieve.

What a great recommendation. I actually had the honor of meeting John at Kleiner Perkins probably 10 years ago, and yeah phenomenal personality, certainly well-deserved in terms of the market attention he tends to get, and this book, “Measure What Matters”, is a great one in that regard. Well Tony, thank you for this insightful interview.

You’re very welcome Ken, I really appreciate this opportunity to talk about my background and share a little bit about what we’re working on today.

So, this has been Tony Paine, Co-founder and CEO of HighByte, a serial entrepreneur and digital industry leader. Thank you for listening, and please join us next week for episode 104 of our digital industry leadership podcast series, produced by, for, and about digital industry leaders. Thank you and have a great day.

[End]