Podcast #162 The Fog of Industry

Ken: Good day, and welcome to episode 162 of our Momenta Digital Thread podcast series. Today, I'm pleased to be joined by David C. King, CEO of FogHorn Systems, a Silicon Valley company providing edge intelligence software for industrial and commercial IoT applications. Before joining FogHorn, David co-founded AirTight Networks, a technology leader in secure-managed Wi-Fi. He served as chairman and the chief executive officer, leading the company through four successive up rounds of venture capital funding. Prior to AirTight, he served as chairman, president, and chief executive officer of Proxim, a pioneer in WLAN and the first publicly traded Wi-Fi company. David led Proxim through a successful IPO in December 1993 and drove 20 times revenue growth as a public company during his tenure as CEO. In addition to his board positions at AirTight and Proxim, he served on Netopia, Cayman Systems, and Mobilestar, all pioneering companies in the broadband access and networking industries. David holds a BA in Economics and an MBA and JD degrees, all from Harvard University. David, welcome to our Digital Thread podcast.

[00:01:51]

David: Thank you, Ken. It's great to be here.

[00:01:52]

Ken: And it's great to have you. I love the pre-conversation both of us had about our respective histories and how many times we've crossed paths. It's good to be able to get you on this program finally. I always like to start asking about one's digital thread, if you will. In other words, the one or more thematic threads that define their digital industry journey. What would you consider to be your digital thread?

[00:02:17]

David: It's great to be here with a fellow pioneer in the industrial IoT domain. My digital thread extends back over 40 years. So I'm dating myself a bit back to the late 70s, when I was actually in college still, working my first college job at IBM, of all places. Back then, in the late 70s, after 40, 50 years of success, IBM was by far the dominant player in the technology world. They had probably about half or more of the industry's revenues and more than a 100% of its profits. That isn't the only profit company, but with that leading market position, and some would say dominant position, my role was in the minor systems division. Specifically, today we call them server computers. Back then, they were the early stage mini computers. I was working on the smallest system, which was a precursor to the IBM PC. So the beginning of my digital thread goes back 40 plus years ago in a very, very different world. We're working on very disruptive technology within the IBM context where the smaller systems were starting to take over certain tasks and roles from mainframe computers, the big iron. Then really the gleam in the eye with these precursors to the IBM PC, which came 5, 10 years later- the idea of a personal computer was being hatched. After that, I got multiple degrees, as you mentioned, at Harvard. And during that time that, I worked at McKinsey as a consultant, and one of the seminal experiences I had there was consulting with Apple in the mid-late 80s. About a decade later, the PC had been launched. Apple had gone from the Lisa to the very early versions of the Mac, and that's precisely the time I was there consulting. And so, I kind of witnessed the birth of the personal computer, both the precursor generation, as well as the first true PCs that IBM and Apple had brought out at the time. From there, I went after consulting to take over a little company called Proxim. I first joined as head of marketing. And believe it or not, acting CFO- it's a small company at that time, but it happened to be the moment when wireless local area networking or wireless LANs- first took hold in the early 90s. Proxim had an exciting pole position, developing spread spectrum radio modules and eventually the first wireless LAN NIC cards. These built-in modules were plug-in modules for mobile devices. And that time was primarily industrial devices, barcode scanners, and such mobile medical equipment, but also had the first products for personal computers. These laptops were beginning to ship at the time. Again, very early stage pioneering technology, almost a decade before the term Wi-Fi was invented, the whole Wi-Fi industry took off. Again, disruptive technology at the time and the wireless networking space, where people were fundamentally doing everything over ethernet and trying to get from 10 megabits per second up to 100. We were pioneering the first one to two megabits per second wireless LAN systems, pushing forward to get to 10 megabits and eventually 11. From there, we went to AirTight networks in the early 2000s. I co-founded the company AirTight in 2004, or '05. That company was the first to do secure cloud-managed Wi-Fi, along with other pioneering companies like Rocky. And so we went from wireless LANs plugged into computers and access points that cost 1000s of dollars hang on the walls, or over in the drop ceilings, into a world where Wi-Fi was starting to be embedded in laptops and then smartphones. This company went into the cloud and provided a wireless LAN system that had built-in intrusion protection. So fundamentally kind of pioneering in disruptive technology in the Wi-Fi world for over 20 years. And then finally, I ended up in FogHorn in 2015, where we first met. Foghorn was, at that time, a very pioneering company in the edge computing space for industrial IoT, which we'll talk a lot more about because that's the subject of focus of this particular podcast. But over these 40 years, the digital thread has been- it's an overused term, but disruptive technology, right? From small systems, when the whole world was extensive systems at IBM, two PCs, when the world as a whole was just trying to figure out what a PC meant, right back in the 80s, and 90s, right into wireless networking, long before it became pervasive in the post-2000 period. And now, it's still frankly in the primary phase of bringing edge computing into the industrial IoT world and specifically bringing the capabilities of AI down to the Edge.

[00:06:36]

Ken: I like that kind of- the dual threads, disruptive and pioneering. As you said, that defines the journey you've taken and the leadership positions you've had. Excellent leadership record there, and one win after another. All of this, of course, culminated in you joining FogHorn in 2015. What attracted you to the company and, I guess, more importantly, the industrial IOT space that it represents?

[00:07:02]

David: Well, I was brought into FogHorn by some financial investors who back the incubator, the hive that created FogHorn in late 2014. And they've given it a name, hired a few engineers. And the concept that pushed forward was this idea of an edge computing stack for industrial IoT. That was kind of the target space. But frankly, there wasn't much definition around that yet. What did that mean?

As you know, companies like ThingWorx, which you create, had pioneered this idea of a mash-up layer, application development environment- an IDE, if you will, for building industrial IoT platforms principally in a cloud or centralized data center environment. And so the whole idea of FogHorn was, 'Hey, we need this extra layer or an additional layer of computing where you don't want to send all the data to a centralized computer resource or the cloud because the data needs to be acted on immediately for latency reasons. You've got security concerns about sending it somewhere, but remarkably, you want to do this in a more low-cost, more hyper-efficient fashion. So the concept was there. But just as a show, we were pioneering wireless networking, when the world was still wired LANs 5, 10 years before it became a pervasive technology and maybe even ten years plus.

Frankly, the same thing was happening today as we speak, in edge computing, industrial IoT. So targeting industrial was simply a function of the use case, the big, the most valuable in the industrial or the commercial setting. The idea is that you're going to take these captive OT systems, Operation Technology systems. PLCs, DCSs, and all these acronyms that come out of the OT world, whis very foreign to IT professionals, and into folks in the modern technology world. I live in Silicon Valley, right? The idea that you're going to take all this, kind of captivating the OT world and send it to the cloud, really didn't make sense, and so you needed to have a different paradigm for computing. When I was introduced to FogHorn, this is what we were trying to do in wireless LANs 15, 20 years earlier. When the entire world was trying to figure out how to get data over wires faster.  Now more mobile platforms, like laptops, and eventually- PDAs, finally became smartphones. As more and more mobility entered the computing world, we needed a different paradigm for networking and wireless LANs, and specifically, Wi-Fi became the way that happened. Likewise, as we see the pervasiveness of IoT, getting beyond consumer- mobile devices and connected vehicles into the industrial world, into the real physical world- whether it's in manufacturing, or its energy production, oil, and gas, whether it's in smart buildings and smart infrastructure, smart transportation. The idea that you're going to send all of the world's physical sensor data to the cloud doesn't make much sense economically from a security perspective, especially from an application development perspective. The idea of Edge computes targeting industrial IoT, to me, kind of went hand in hand. And so frankly, my experience approximate were our first customers, way before Wi-Fi had been invented, were companies in the warehousing, manufacturing, retail, and transportation space. And so, to me, there are many similarities that to make this technology genuinely pervasive, you need to pick the verticals. You need to choose the application solutions. These use cases will drive adoption. And so we've been on that kind of a journey at FogHorn, and finding what the best use cases are, how do you show enterprises, the fundamental value proposition around this technology, and from the early beginnings of POCs, and early deployments, to scale, and that's the journey we've been on for the last five years.

[00:10:41]

Ken: Certainly, a continuation of your disruptive and pioneering theme earlier, especially relative to edge computing- it's interesting, as some of the other companies you mentioned, ThingWorx as an example, there were always the two components. As you call it, the mash-up environment, the visualization, if you will, and behind it always some form of distributed- they call it to edge microserver but some small component that would sit down in a gateway and help collect the data and bring it back up. And it's almost like the IoT platforms, industrial IoT platforms kind of bifurcate around that. You've got the ones that- truly, the kind of the visualization cloud level, and the ones that have focused on the Edge, and you guys have been a leader in both. To help understand FogHorn a little bit better for the audience, can you talk a little bit about some of your key use cases and wins?

[00:11:29]

David: Yeah, so the use cases that we've focused on had been those that involve what we would call more advanced computing, rather than just kind of collecting data as you described, and getting it to a data center or getting it to the cloud to do the visualization, and to try to execute machine learning or advanced analytics. FogHorn's vision from the beginning was to put- what my co-founder, Sastry Malladi, and I after we got in and started running this company in 2015- was put what we call the intelligence at the Edge. So we'll talk more about that, I'm sure, throughout the discussion, but to us, Edge intelligence means putting a lot more compute capability that is very advanced complex event analytics, machine learning, deep learning, really AI operating on live sensor data. And for us, the key that uses cases- the value proposition is far greater if you've got high volume or velocity data, like video cameras, audio sensors, high-density vibration sensors. Or you're talking about contextualizing data coming from multiple disparate sources, sensor fusion of a lot of different data sources, and providing real-time response at the Edge that is on the live data. And so the use cases that we've done, I call our company the last few years of being on a journey of letting a hundred flowers bloom if you know that historical reference. Where literally, we've done over 100 plus successful POCs and initial stage deployments in many, many industries. Discrete manufacturing, whether it's automotive or semiconductor, industrial products, consumer packaged goods, any type of discrete manufacturing, you can imagine we've probably done use case, similarly, in process industries, especially oil and gas, where we've had a major investor, Saudi Aramco, the largest industrial company in the world, that has been on a digital journey with us to pioneer many, many use cases in the oil and gas production and midstream processing side of their business. And I would say that what we see in both manufacturing oil and gas is split between video-based applications, image processing, machine vision- where we're running our stack and a camera right next to a camera to do very advanced things - it is more than just object detection. And you're trying to do on-the-fly decision making now, right, whether it's for worker safety, whether it's for reducing emissions for floor stacks, whether it's video quality inspection, manufacturing. So looking for those anomalies or improvements in processes, where a video camera can make a difference- the camera is the new sensor. We've also done a lot of work, of course, in that high-velocity realm with vibration. Vibration sensors have been introduced over the last 5, 10 years. They're becoming quite pervasive, adding a much greater layer of clarity and granularity to predictive maintenance. Predictive maintenance when you're just looking at long term historical patterns of machines that break down looking for signals and temperature and pressure, velocity, torque, whatever your traditional digital sensor vibration gives you much better signals earlier if you can weed out all the false positives. You can deal with the kind of profusion of data being produced every second of the day. And so video vibration, hype velocity sensors, that's one part of it.

The other, I would call it, slower data sources, things like traditional digital sensors, temperature, and pressure, but bringing that compute right down where you'd have measuring temperature every second. But you've got thousands of these sensors or tens of thousands of these sensors in the building or a transportation system be able to process those without trying to transport all that data to a centralized data source like the cloud, where frankly, it's just not very efficient. It costs a lot of money to store that data much less transport it. You want to find use cases where you can drive a kind of immediate business value. I've recently done a lot of work in energy management, specifically around HVAC optimization, bringing 60% of the bill. Anybody's utility bill a month is HVAC. And that HVAC system, by the way, you can drive 20, 30% savings, by frankly, turning things off or not having things run when they don't need to be running or tuning them optimally for the right temperature setting at the right time. These kinds of things can drive a huge amount of savings to the business value and reduce carbon footprint at the same time. So the use cases are many and varied. And as I said, we've done a lot of work in manufacturing, in oil and gas, in smart buildings, infrastructure, and have done some in transportation around rail and fleet as well. But there are just so many different sectors and verticals to cover much fewer companies- I think we only scratched the surface.

[00:15:52]

Ken: Yeah, you guys have a pretty wide remit of interesting use cases that you've done. I guess in the same sense. There's a diversity of use cases, and there's also a diversity of platforms that are out there. I think IoT analytics account is up to 700 IoT platforms, given their last report. Only a few, of course, I think you and I would agree, are what I'd consider being truly enterprise great industrial IoT quality. But given even that large group, what I'm impressed with is FogHorn always seems to stand out in terms of market presence. And even most recently, in the Magic Quadrant with Gartner. How do you guys do that? How do you maintain that standing out among all of the other peers that are there?

[00:16:30]

David: Well, thank you. I appreciate the compliment coming from the founder of ThingWorx. It was one of the first to stand out as it- first, independently as part of PTC, you paved the path, if you will, in terms of pioneering efforts here. But frankly, FogHorn's focus was not so much on the platform per se. In terms of IoT, it was meant to be a specialist in building the world's best edge computer, edge intelligence platform. And so yes, there is the ability to vanish from the cloud in which we provide as part of our lightning product, our OEMs like initially, GE, and eventually Honeywell, and others that have leveraged our platform, fundamentally have found value, specifically in the very advanced analytics, deep machine learning that FogHorn does at the Edge. And you can use our complete platform, which is a true end-to-end edge to cloud platform. But really, we also OEM the platform, the edge part of it, that can be plugged into other companies. They may have their data ingestion. They may have their kind of mash-up layer cloud infrastructure, cloud architecture that we can plug into. So our biggest key to success has been our singular focus on having the world's best edge platform that can be either our own in a full end to end platform, which is really where we make some of these competitive analyses, but also as part of other companies' broader Edge to edge cloud systems. By focusing on the Edge and pioneering it every step of the way, first with Edge analytics, that Edge machine learning, and an Edge deep learning, Edge AI- we've architected the product from the beginning. I give my co-founder, Sastry Mallady, all the credit. He architected this from scratch to be not only very flexible but having a leading-edge, no pun intended- technology. Things like containerization- we were very early adopters of Docker back in 5, 10 years ago. Docker was still fairly new. And when we started back in '15, Docker was becoming a pervasive platform, but we made a good choice. And then, of course, moving to Kubernetes as a containerization strategy for Edge to deploy distribution. I think we've also made all the right calls in terms of the ingestion protocols. It's not enough just to have MTT and maybe OPC, UA, and DA, which many companies have for certain manufacturing environments. But if you're going to energy, you need Modbus, both TCP and serial, if you get into transportation, CAN bus. If you get into building technology, BMS is great. And so, we've added all the right industrial protocols. And of course, the architecture is flexible enough if you've got some proprietary protocol that's not supported, right? We have, you know, an SDK that makes it very easy. So you can extend our platform and take our full end-to-end solution, if that suits your needs, or work with any of our OEM partners that have adopted our edge stack as part of their end-to-end solution.

[00:19:09]

Ken: Perfect. And just so we don't continue to perpetuate a misnomer out there, I am not a founder ThingWorkx. I will give that honor to Rick Bullotta, whom we all know and love in the industry, and together with Russ Fadel and John Richardson, they were nice enough to invite me in very early to be part of the company. But thank you for putting me among the highly esteemed group. It is interesting, though, being part of ThingWorx and a couple of other companies we've invested in this space. Many analysts, especially those I think, are coming from the IT side, kind of bemoan how slow things have moved in the industrial IoT. Perpetual trough of disillusionment, I heard somebody say not too long ago. When you joined Foghorn in 2015, I'm sure you had an idea of the trajectory of the company's going to go as and the industry. Do you think we're on track for that? Or do you think things have moved slower? And if so, why?

[00:20:02]

David: Well, early on, frankly, one of the colleagues that recently joined Momenta, Mike Albeck, who is our series a co-lead investor from GE ventures at the time- I think said it best, which is there's built-in friction and barriers to adoption for industrial in any of the industrial sectors that we're serving, whether it's manufacturing, oil, and gas, smart infrastructure, transportation. All of those sectors, it's their massive sector, they have half the world's GDP, but they're also notoriously slow to adopt new technology. And they are focused on their OT systems, which were kind of firewalled or air-gapped off of IT systems, they didn't want to be susceptible to security breaches, and they didn't want any disruption of real-time production. And so we come to the industry knowing that there are some built-in barriers to adoption and built-in friction in the system. As I talked about it with Mike way back, that just as Mike saw the value back then in how GE could leverage the next instantiation, if you will, the value in IoT was going to be the Edge not simply in the cloud, we knew it was going to be a slow pace. So I always joke that, hey, this could take decades to play out. And so one of the questions I had for the investors that brought me in is that they know what they're signing up for, right? These are financial investors that invest in Valley companies whose typical lifecycle might be five to seven years before you monetize. But here, it could take a lot longer. So that's one of the reasons that we pursued companies like GE. Initially, Mike, and eventually a lot of the other leaders in the industry like Honeywell, Saudi Aramco, Bosch, etc., to co-invest in the company because we wanted that industrial perspective, both as a path to gain customers and get adoption in the market, and validation in the market, but also because they bring the perspective about the pace at which change happens in the industrial world. So we kind of knew that going in, and having had my practicum experience, we were building very specific use cases and applications for wireless LANs, whether it was in retail barcode devices for price markup, markdown, or inventory control, or whether it was mobile patient management systems, or mobile diagnostic equipment and medical- I knew there was going to be a path to try to find those use cases. It has been slower than probably any of us expected. But part of that I would attribute to COVID.

Frankly, we were starting to scale, I think, well, a few years back, two, three years ago with these 100 flowers. So the use cases that we were doing started to find the ones that had many scalabilities attached to them. They're a customer base. When COVID hit- and frankly, not getting the plans and working with operators slowed everybody down. Thus, this threw a curveball at the entire industry, which I believe we are only now beginning to emerge from. We're now you're the digital transformation efforts, the investments in IoT, and all these different IoT-related technologies. It's just starting to pick back up again. I'd say COVID didn't help us. And then probably hurt the cause for many who are beginning to get traction. And fortunately, we had enough staying power with a lot of our significant customers, especially Saudi Aramco and oil and gas and some of our more substantial manufacturing customers, to keep going. And we also diversified at that time into energy management, which with its real sustainability, kind of carbon footprint reduction, and economic savings drive, give us a different pillar, if you will, to build a business on.

[00:23:14]

Ken: Given your prior background in Airtight and Proxim- both of those ten years plus in terms of your tenure, perhaps the original investors that brought you into this had that in mind because you're a rarity in Silicon Valley of putting ten years into one company. So I'd say you probably have the staying power. And I'd say you're probably going to see the value out of there, hopefully, proportional to or similar, I should speak to your prior companies as well. Look, quick question. You mentioned some of these great clients, customers, and investors you have. How do you know when an organization is ready to utilize FogHorn?What are some of the best practices you've seen in organizations using these platforms?

[00:23:58]

David: Yeah, it's a good question. I think the surest sign, of course, is you've got staffing and budget. That's the kind of generic answer that I think any tech company would say once an OEM sets up a group and or puts on the roadmap embedding edge technology into their IoT platform or their IoT applications, right? We're on the technology roadmap, but until they, frankly, start to staff it right and allocate budget, right to the edge intelligence layer, if you will, you're never sure kind of how ready they are. But once they start to staff it, then you know you're on the right path. And that applies both to OEMs. But it also, of course, applies to end customers. Now, at the end customer level, the one big tell, if you will, or the surest sign that you've got a customer that's ready to go is when we started to work directly with the operation staff when you're starting to work with the plant person to help the plant manager, the people on the line, the OT professionals that work with those operations folks to get these applications to deploy in the plant on the machines in the process. Once that happens, you know you've got a customer who is not just a POC, a digital experimentation project. This is more than just proof of concept. This is something that's going to be utilized and potentially scale. That's the surest sign that you're on the right path. If you're talking just to the digital folks about putting something in a lab or doing some kind of experiment where you're working just on historical data or simulated data, you're not there, right? It's only when you're starting to work on pulling the data out of the protocol servers in the plant floor, we're attaching to PLCs, we're plugging new vibration sensors, or putting a camera in place to run the models until you're at that point of working right in the physical environment with the actual operations people, you're not there. And as I mentioned, COVID, where you can't get the plant environment, you can't send staff into the plants to deploy this. That's a natural barrier. We were just starting to see everybody now opening up all the plants and getting back to more normal business practice, which is a good sign because we'll now see the pace of innovation pick up again.

[00:26:03]

Ken: Yeah, I fully agree. And I've heard the same thing across our portfolio of companies as well. The good news is, all of us are generally in the use case of remote asset management, and that has been the critical use case when you couldn't get people out to these things. The industry generally has done pretty well during this time. But the constraint has been the ability to install these systems in many places. Look, I know you've proudly been a Silicon Valley company. And you've mentioned COVID several times. I'm curious because I've no longer live in Silicon Valley. Given this- what they call 'tech migration,' how have things changed in Silicon Valley regarding setting up a company, work-life, etc.? Is everything the press says it is or much less?

[00:26:45]

David: Well, I think we're not entirely out of it yet. We're still primarily locked down in the Valley; Santa Clara County is one of the strictest probably- at least in the United States, maybe the world in terms of really restricting the work environment. So essentially, I think most companies are still remote or largely remote. Many of the Valley's huge tech titans, some of whom have already gone to the extreme, arguing that everyone can do remote and nobody has to come in. Most say they want- especially on anything on the technology side of the house, where in-person collaboration can make innovation. They're trying to encourage people to come back. But the restrictions are not quite fully relaxed yet. So we're still waiting to see how it all turns out. I think in the end, it wouldn't be the worst thing. If traffic went down by half and traveled what went down by half, I've been saying this since the start of COVID. I think we were all way too invested in everybody going to the office or everybody being on the road in front of customers that some of it were probably unnecessary. So this could help us in the end. The hybrid environment everybody talks about, I think, would be a good endpoint. But we're not there yet. At least in the Valley companies, you just started bringing employees back. We have not yet brought everybody back. We allow folks to come in, but many restrictions around masks and other things are still in place. So it's not kind of the fully open environment that we hope to get in the next few years.

[00:28:07]

Ken: You mentioned hybrid events, generally trade shows, etc., is where I see the most significant difference. And I'm not sure what I would do except for those exceptional ones, live events, because the hybrid ones are getting good in terms of how the session materials and everything else is going, and I suspect that'll be one of the biggest changes we've seen. And you talked about a climate-friendly approach, and nobody was jumping on aluminum tubes to fly across the Atlantic or Pacific to go to these things. So in closing, a question I always like to ask is, where do you find your inspiration?

[00:28:40]

David: Well, I saw the question- Books online- there are many ways to get the information. But fundamentally, for me, it's about the people and the companies. When I think when I talk about disruptive technology, pioneering innovation, I draw my inspiration- everybody can point to folks like Steve Jobs, or Jeff Bezos, or Elon Musk. Those are the easy folks to point to truly disruptive personalities in their companies. The success of their companies demonstrates the power of their vision and their disruption. But there have been just generations of great pioneering leaders in the early days in the Valley. You had pioneered in the early generation, not as powerful probably as this generation. Still, you had the Bill Gates and the Scott McNealys and the Larry Ellisons of the world, the last generation of tech powerhouses. But to me, it's not just the pioneers, the entrepreneurs that kind of drive to the ultimate success. It's also some of the outstanding managers who have taken firms to the next level. Folks like Satya Nadella and Shantanu Narayen from Microsoft, Adobe. Folks like Nikesh Arora from Palo Alto Networks, Lisa Su at AMD. Sometimes it's a great leader-manager coming into an existing company and just making it better. You think of GE and Larry Culp coming in and turning around that battleship trying to do what he's doing. For me, it's all about the leaders, the people. That's what inspires me. Whether I'm learning about them, whether on television or online or reading about it in books, it's just the story of the people. Then the organizations that they lead and the powerhouses those companies become. I just start started jotting down names when I saw the question. I came up with two dozen, three dozen of these influential leaders, learning about their journeys and what they've done to transform their organization, and driving forward, there are so many success stories in technology. And that's what inspires me.

[00:30:24]

Ken: I put your name among those as well, given your serial entrepreneurial history, and indeed your successes and the success that FogHorn undoubtedly will be. David, thank you for sharing this time and these insights with us today.

[00:30:37]

David: Thank you. Great to speak with you.

[00:30:39]

Ken: Yes, as well. So this has been David C. King, CEO of FogHorn systems, and if I may, a surreal, disruptive, and pioneering digital leader. Thank you for listening. And please join us next week for the next episode of our Digital Thread podcast series. Thank you, and have a great day. You've been listening to the Momenta Digital Thread podcast series. We hope you've enjoyed the discussion, and as always, we welcome your comments and suggestions. Please check our website at momenta. one for archived versions of podcasts, as well as resources to help with your digital industry journey. Thank you for listening.

[The End]