The Resurgence of Networking — A Conversation with Aviatrix CEO Doug Merritt

Disclaimer: The Futurum Tech Webcast is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded, and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors, and we ask that you do not treat us as such.

Transcript:

Daniel Newman: Hey, everyone. Welcome back to another episode of the Futurum Tech podcast. I’m Daniel Newman, your host, CEO of The Futurum Group. Today we’re going to be talking about security, networking, AI, and so much more. And I’ve got a return guest, someone that hasn’t been with me in a little bit, but someone that I always enjoy talking to very much, Doug Merritt, CEO, Aviatrix. Doug, welcome back to the Futurum Tech podcast. How are you?

Doug Merritt: Good, Daniel. Thanks for having me back. It’s nice to be back on the show.

Daniel Newman: It is. It’s great to have you back. It’s been a while. For anyone that knows Doug and I a little bit, you may realize that we are in the same city, but we are not in the same room. This is becoming a bit of a problem for me, but it’s indicative of the lives we lead, the schedules we keep, and the fact that, when you’re figuring out your next big move, there’s always a bit more time for lunch than there is once you’ve figured out your next big move. And, you figured out your next big move, you’ve landed at Aviatrix as the CEO in a really hot part of the market.

Give me the quick background, tell me a little bit about the decision to land here. Tell me about why you were so excited about the company. I know you had a lot of things on your mind because we met a few times, we talked about what you were thinking about.

Doug Merritt: We did.

Daniel Newman: You picked this company, and it had to be something that really caught all the things that were important to you.

Doug Merritt: Yes. Yeah, no, I actually debated, I think like a lot of people, and tried to live the life of being retired or partially involved, investing, working with VCs, being on boards, doing advising work, all the usual. I’m intellectually curious and I can’t just sit on the sidelines, but do I want the mantle of operational responsibility every day?

And I thought that would be great. I actually truly love my wife and I really love my family, and you now get an opportunity to go on vacations and travel and make sure you’re at every single function and sports game. That part was great. But, I realized, over a year and a half, I actually physically met with VCs and PEs and investigated what it looked like to be a true investment partner, to fully switch and get a different job, what it looked like to be an operating partner, what it looked like just to do my independent thing, and it just wasn’t satisfying for me.

I get why it would be satisfying for others, but I realized, for better or worse, I am an operator. I love meeting with customers. I love hiring and mentoring people. I love going through product reviews. I like setting plans, financial plans, work. We’re about the operational cadence of a company. And, ultimately, I like being on a team, and being part of a mission, and a vision, and a purpose that is singular. It’s you’re all in and either it works or it doesn’t work. And so made the decision that as attractive as that other life was, I was going to give up some of my flexibility and go back and be a full-time operator and wind up at Aviatrix.

Daniel Newman: Well, first of all, it’s funny that we have to explain that we love our families when we also have to explain that part of loving and being good is being satisfied and happy. And I think if there was one thing I recall from some of our meetings is you could just see in your eyes there was something missing. So it looks like what was missing is look after building and doing what you did at Splunk, you’d certainly earn the right to be at a lot of tables to help a lot of VCs to sit on a lot of boards. And by the way, I do all that now. I run companies, I sit on boards, I advise, I’m an LP, I work with VCs, I do all those things. I think it’s a motor, it’s a cadence that I think actually allows you in that other part of your life that you talked about to be so much more appreciative, like when I actually go on vacation because I’ve been working…It’s like when I have too much free time, my mind works too fast. So anyways, tell me about the Aviatrix selection though. It sounds like you probably had a few choices put in front of you. You had different industries. I know I look back and laugh because we got interested in metaverse and blockchain for a while and I know you and I both saw the ups and downs and you definitely didn’t go that direction, but you ended up in security. Talk about the Aviatrix pick, why you landed there.

Doug Merritt: Yeah, so a couple of things that were super important to me. One was a board that I knew and had a trust relationship with, and Aviatrix is a very tight five member board. They’ve known each other for a while. More importantly, I’ve worked very closely with three of the five members and then over my interview cycle of the seven months, almost seven months since I’ve been at Aviatrix, obviously developed a tight relationship with the other two. But I think everything starts with do you really have the confidence and trust bi-directionally between the governing group and the operating team.

And then behind that, for me, it really had to be a topic that I was passionate about and felt was super important to the world around us. And one, I actually toyed with non-enterprise. I got pretty close to a B2C opportunity that I was very passionate about. It’s an unbelievable company. It was a really hard choice, but when I really peeled back, it’s like I’m not a B2C person and I’ve developed almost 40 years of experience in enterprise sales and all the motions that happen with that. And to be learning something new is exciting, but it’s like why wouldn’t I be leveraging the investments I’ve made over my entire career?

But what I’ve seen with cloud, and we lived it with Splunk and Spades, is it’s really, really… Cloud, as we all know, has completely changed the architecture of the way that people think about building, deploying, managing, running applications. And it’s very difficult to bring a non-native service to cloud never works well. And when you look at it, what’s really worked? What are the things that have taken off in cloud and dominated? There are companies that were built for cloud. Datadog is a great example in the observability space and the app observability space and Snowflake and the actual data and the next generation data warehousing and data storage space.

We obviously went through a huge amount of pain at Splunk to refactor and rebuild so we could be more cloud native. Very proud of the work there. Super hard to do. But it just highlighted, hey, to really make cloud work, you need a problem set within the layer of technology that’s important to organizations, then you need solution sets that were natively built for that. And part of what we experienced at Splunk is networking is something that clouds certainly had to deploy. There’s no way you could be in a cloud without networking services, but that wasn’t the lead for most of these clouds. AWS started the whole trend obviously, and they were focused on how we get the developer oriented services that really make app development quick and easy and painless. Developers are not super networking aware of folks and they just want networking to be invisible.

Like, “Hey, that should just take care of itself.” And what we experienced at Splunk is networking actually really, really matters. When you look at the old seven layer OSI stack, I know that’s really dating me, but networking is one of the foundational elements within the entire compute paradigm. And ultimately what I saw is you’re only as agile, as responsive, as cost-efficient and as resilient as the weakest link in that chain. And the foundation of that chain where it’s embedded in the ground is the networking layer. And most of the services are out there from the cloud vendors. They’re working on it, but they’re incomplete and pretty complex and you can talk more about that. And then the last generation services aren’t built for the cloud, whether it’s firewalling or core networking, of course switching, core routing. They’re virtualized monoliths that are plopped into the cloud. And that doesn’t work very well as we’ve all seen as we went through that lift and shift orientation versus the modernization or both scratch orientation.

Daniel Newman: Yeah, I think you’ve found a very interesting intersection, Doug. The challenges of companies are not stovepipe. And I think that’s one of the things that is becoming increasingly evident for CIOs and boards. I had a really interesting conversation recently. I did the keynote at a observability/AIOps conference for LogicMonitor, a different industry, different business competitor in your old world, but not anymore. And one of the things we were talking about is with the advent of AI, technology has crossed into a board level conversation at a breakneck pace. Now we like saying that about digital transformation, but I think that was more of a culture move into the boardroom. And probably a very firm technological move into the boardroom. It’s, hey, how do we gain productivity? How do we gain efficiency? How do we hire differently? How do we distribute products and build supply chains? And of course, how do we make sure that we’re keeping security and privacy and data sovereignty top of mind? And to your point, companies are laid in with technology debt. They’re laid in and stacked with technology debt.

There’s some companies that may not love when I say this, but the reason we haven’t seen more enterprise workloads move through the cloud is not necessarily because of a cloud problem. There’s some economic stuff that is complicated and companies are financially like, “Ooh, that’s more expensive than I thought.” But moreover, it’s just really hard. It’s a really hard problem without massive investment in CapEx and replacement and displacement of technology. But you think to some extent maybe with things like what Aviatrix is doing, you’re cracking the code on the hybrid, the cloud, AI, security and bringing them together, creating a simplicity layer. Talk a little bit about how you’re thinking about that, knowing what enterprises really look like, which is not built from zero or built on cloud, it’s built on a whole lot of technical debt with the desire to modernize.

Doug Merritt: Absolutely. Yeah. For better or worse, the bulk of the IT industry is fueled by 5,000, 10,000, 20,000 organizations globally that all are more than 40 years old. And because of that, they’ve got generations of technology that were all purpose-built. And it’s really difficult to take an application that was built for architecture A and shift it to architecture B and have it run efficiently, which is where I think the lift and shift realization has been for companies. And again, it’s going back to that foundation of, okay, if you understand cloud and how cloud is built now how do you craft something to be native within that cloud? But ultimately the value that I see for Aviatrix is you’re only going to be as agile, as responsive, as cost-efficient as you would like to be based on the flexibility of your networking and then networking security layers, going back to that bottom layer of the stack.

And something like generative AI and large language models, which I think complicate the landscape even more, because now not only are you in multiple hyperscaler clouds, Google, AWS, Microsoft as examples, but there’s becoming more and more purpose-built compute centers. Whatever the heck do you want to call them that are very LLM specific to add into those big clouds plus your last generation on-prem data centers that most people realize, “Oh, I can’t really get out of, or I could, it’s just my costs are going to increase because I’m going to take this Java app and throw it in a monolithic approach into AWS, and that’s more expensive than running it on my depreciated purpose-built hardware already in my data center.” And now you add in the complexity of regional pops through Equinix, or Megaport, or others and you get a really, really complex landscape that is the gate to how am I going to get this amazing breakthrough capability that’s something like a large language model could bring in my development cycles or my sales cycles or my marketing cycles.

So our aim and we’re early, a lot like I think Snowflake was early 10 years ago, people didn’t understand why you’d need a purpose-built data warehouse for cloud to get to the data cloud, and then they eventually did and then they did all at once and the company became massive. I think we’re in that same cycle of, “Oh wow, I am really gated with the flexibility, agility, cost-efficiency, responsiveness that I want by my network and services,” which is my excitement for being here. And our purpose is like how do we actually help the boardroom and the technical team deliver on what they’re trying to get done through this shift?

Daniel Newman: So I want to talk about AI a little bit because I feel like it’s a sin now in some way to not lean into AI more. And I know it’s not necessarily one of the words that you use most aggressively in your Aviatrix marketing, but I also know that every company on the planet knows they need to play in the AI space. And the other thing I would say is AI is creating a really important requirement for companies to double, triple, quadruple down on investment in security. And now our data is not showing that IT budgets are actually expanding like they should. And I mean that is, right now what we’re seeing is budget shift rather than budget expansion. So companies are saying, we want to grow and expand revenue, productivity, use technology, but we don’t want to spend more on it. We just want to cut costs and this is going to come home.

This turkey’s going to come home to roost, because all this AI is going to mean a lot of security vulnerability. It’s going to mean the black hats and the bad actors are going to be just as good if not better at using this technology against companies. So as you see the security landscape, how is AI shaping demand and shaping product development and shaping your vision of how Aviatrix expands to deal with everything from network perimeters to cloud and security? It has to be a big task.

And I don’t know, I feel like security is always a second or third or fourth thing, but should it not be right there side by side like AI and security, because we say things and I’m rambling on, just give me one second. Davos, the number one topic was trusted AI, responsible AI, governed AI, but I didn’t hear the word secure AI. And what I don’t understand, is it just because no one wants to talk about it, no one wants to spend on it, or is it that people still aren’t seeing that they’re symbiotic, this really important relationship between security and the ability to reach trusted AI?

Doug Merritt: Yeah, I think sadly, security is always something a lot, like networking in many cases, that’s the add-on.

Daniel Newman: Yeah, but you need it.

Doug Merritt: Yeah. Again, they’re lower in the stack, right? One of the things I love is serving underrepresented populations that are deeply important. We did that at Splunk with IT ops. Who loves their IT ops team? Well, without them you’re in deep trouble as far as running your tech. And we gave them a bunch of tools that made them superheroes. I view the same thing with networking. Networking has almost been forgotten. It was so important in the ’90s and 2000s as the internet was emerging and now it’s become an afterthought and it is critically important to what we’re trying to get done through these higher order capabilities. And one of the advantages that Aviatrix is bringing is as you move networking to be ephemeral in the clouds, we architected and it’s possible to architect the network with security as a first order element. It’s not a bolt-on. It’s not a set of services you bring in through third parties. It’s actually part of the core delivery.

So we’ve been talking a lot over the past year about our really cloud firewall. Firewalling was something that became an afterthought. We’re not as secure as we thought. So let’s put some firewalls around these data centers to try and protect them. Why weren’t they built into the network from the beginning just because it wasn’t conceived that it should have been built in from the beginning? Aviatrix’s changed that paradigm because it’s woven through the offerings that we have. And then that new capability, that secure networking, native intertwining of security policies and visibility and control with your networking services winds up impacting your AI, your large language model, your purpose-built domain specific language model capability because you now can more natively understand the different workloads and be more thoughtful and pervasive from the very beginning at making sure that there is effective securities, effective perimeter defense interrogation around what you’re trying to do with your initiatives around AI, whatever those happen to be.

So we are definitely thinking about how we do it, security was a first order for us as a company. We’re thinking about how to apply generative AI to our offerings, but we’ve been on the path of traditional machine learning and AI for a while, while not all networking capability is going to be well-suited to a large language model or domain specific language model. And then at the parallel we’re super aggressive about how we are going to use AI to power Aviatrix? And that was purposeful. One, I’m a huge believer on the impacts that AI can have on everyone’s daily life making BDRs and SDRs much more effective and productive sales reps, more productive, your legal team, more productive, your finance team, your HR team, that there isn’t a corner of an organization that’s not be impacted positively, for now, and we’ll see what happens long term. We’re all a little bit dubious about where this thing goes, but it’s amazing the superpowers that it gives to every single function.

That was the first order. But then the second order is if you really want to weave AI into your solutions, playing with it and seeing the impacts becomes a really positive experience for now conceptualizing, not theoretically, but because I’m working with large language models or domain specific language models every day in my life, “Oh, this is how it could wind its way into a set of runbooks we’re delivering into our customers. This is how it could work its way into our monitoring frameworks for a set of customers. This is how it could wind its way into autonomous networking recommendations or security recommendations for a set of customers.” So there’s a duality that we’re working on here that we’re starting to see some really positive network effects within our company.

Daniel Newman: Network effects with your networking solutions.

Doug Merritt: Yeah.

Daniel Newman: It’s funny that you say that how we are so into the hype cycle on things that we just completely divert all focus. And the thing is, remember when we used to say software will eat the world? And I wrote a MarketWatch piece in 2019, end of the year, had no idea that we had the 2020 coming. I don’t even say the word. I don’t want to get our video censored, but the point was I said, “No, semiconductors will eat the world.” And I said, “Because you can’t run software on air.” And my point is that layers of abstraction, we got to that front end experience layer and everyone just talked about that. But that doesn’t happen on its own, whether it’s been cloud, or virtualization, or networking, or how data moves. And by the way, there’s been a bit of a renaissance because of AI that data has to move.

We’ve seen all the focus on silicon photonics. We’ve seen the focus on optics and obviously light for moving data. NVIDIA’s massive growth has come a lot from systems development, not just GPUs, but the fact that they’ve got systems, NVLink, InfiniBand, the way they move the data between the systems, parallel computing that’s going on, data has to move in a very specific way to be efficient.

Then you actually have seen actually some of the stocks that have performed best, companies like Broadcom, which is to some people, you might call it a snoozer or maybe they’ve refocused a lot on the VMware and software side, but their core business is remarkably profitable, because they actually help move data in a data center. And so my point is I think people are kind of seeing it, but in the world, you and I aren’t being asked to go on TV and talk about the networking chips or anything where it’s all AI, AI, AI, AI, LLM, LLM, LLM, and then a lot of these important things become afterthoughts. So for the fun of my audience, the audience here, Doug, here’s a quick one. Now, how long have you been here exactly at Aviatrix or roughly?

Doug Merritt: About six and a half months.

Daniel Newman: Only six months. It feels like it’s been a year.

Doug Merritt: I know it does.

Daniel Newman: I’m going to say six months.

Doug Merritt: I’ve accomplished a lot–

Daniel Newman: You’ve gotten a lot done and that’s good. In six months from now when you’re going to an event and you say, “Hey, I’m Doug Merritt, CEO of Aviatrix.” Someone says, “what’s Aviatrix?” What’s your quick two-minute, how do you explain the company to people?

Doug Merritt: We provide agility, cost-efficiency, flexibility for your cloud and AI initiatives by freeing up the network. We are the network mesh, the network abstraction layer that provides that upper level capability that you’re trying to get done.

Daniel Newman: And by the way, I put you on the spot, you did not know that question was coming. And the reason I did that is I am just genuinely curious, I’m always curious, because I get that a lot. “What’s the Futurum Group?” All right. And it’s like you get really good at that in practice. And of course you have to have a version for the technically astute and then a second version of that for the person that doesn’t really know anything about tech, but maybe is on the investor class or it’s the board level class that you do want to be on the map.

Doug Merritt: And where we’ve rotated really effectively I think is at that technical level. The typical network engineer that gets some exposure to Aviatrix so far, the 20,000 plus that are ACE certified, Aviatrix certified engineers, and the customers I met with, they are huge fans. What you guys do is really, really cool. That upper level is what’s missing right now. And that’s usually what happens is you make sure that the people that are going to use your technology deeply understand it, but then they’re gated by the lack of understanding of the economic decision makers. And a lot of what I saw was something like Terraform, which we’re super reliant on where infrastructure is code that you need for this world, this agility world we’re trying to drive. LLMs everywhere, applications that are self-generating and self-healing everywhere. It needs underlying infrastructure that is adaptive and agile and flexible and secure and cost-effective.

And Terraform became an infrastructure code abstraction layer that can run in any cloud or your own data centers. And I view Aviatrix very similarly, which is that we are that networking abstraction and semantical layer. You look inside of AWS, there are thousands of primitives that deal with different networking interfaces across our 22 networking services. That’s a very rich landscape, brutally complex. So if you’re just an AWS customer trying to run networking with the agility and flexibility and visibility and security and cost containment, you want really darn difficult. Now look at Azure, whole different set of primitives, whole different set of networking constructs, different API levels. You’ve got to replicate that. AWS team in Azure. GCP, same thing. Totally different from those other two. OCI, same thing, different from those other three. Add in, I’ve got a bunch of pops with Equinix, very different architecture. Add in my last generation data center, different architecture.

So how do you create that abstraction layer that immediately solves literal spend? We can help reduce your AWS bills because of inefficient compute and AWS resource utilization, because the network, because it’s so difficult to create an efficient networking layer. So right away we can just walk to AWS customer and say, “We have pretty good confidence, we’re going to reduce your monthly bill by 10, 20, 40%,” which generally is a good thing for AWS because they want to help lower the cost of their customer so they can deploy more compute.

Now, what do we do as far as operational efficiency and number of people that you have? What do you do with visibility and policy control? How do we help with security, take that entire stack and pretend that you’re in Azure if you’re not there yet? Now it becomes almost an imperative within these organizations. So I think they will become aware of it, that there is this semantical abstraction layer that’s necessary on the network side. And that was part of my excitement of coming here is you’re serving a usually underfunded network operations center and network engineering team that everything sits on their shoulders. Every single time there’s a slowness in an app, every single time there’s an outage, who’s the first person blamed? It’s a network. Those network engineers and they just don’t have enough budget and visibility because everyone assumes that’s cheap and free and easy and it’s not. It’s super complex. So it’s exciting.

Daniel Newman: Yeah, some type of network FinOps for the modern enterprise IT stack. It’s like if you’re multi-cloud, you have FinOps for your cloud spend, but you probably don’t have FinOps for some of these components, network spend for instance. And this could be one of those very logical check boxes in your FinOps for your overall cloud and hybrid and multi-cloud deployment where you could bring real value without compromise, without compromise, without security risk. And it’s an “and.”

Doug Merritt: Yeah, it’s an “and” and that’s the important piece is we’re not competing with FinOps vendors, what we have is deep packet inspection. We’ve physically moved bits within these clouds within a VPC, between virtual private cloud, within virtual private clouds. So we get a perspective on utilization, on cost utilization at the network layer that unless you’ve got a data plane that’s physically routing, moving bits and traffic engineering, and you don’t get that perspective. So we can give that to the FinOps players to provide a more comprehensive picture. And we’ve got a module called Cost IQ that’s just focused on the true networking spend. But we can do the same thing with the observability players. They’re great at understanding application interrogation. They don’t have that deep pack inspection. So our layer is again, way down that stack providing unique visibility on elements that it’s literally almost impossible for these organizations to get visibility around if you’re using native cloud services, because they’re not exposed from the native cloud services layer. So you need something like an Aviatrix gateway and data plan in there to provide that visibility.

Daniel Newman: Absolutely. So Doug, final minutes here. You have a big sales kickoff, I understand coming up. What’s the big message that you want to drive home to your team and of course your team then drive home to the market?

Doug Merritt: So the two core themes are trying to raise that level of conversation description, the way that we just talked about that. How do we cross that gulf to help our core technical user be able to effectively evangelize and justify our solution at the upper layers of the management chain. So we’re doing a lot, again with AIOps and with our messaging to help with business value, understanding and justification, ROITCO. But we’ve got to cross that barrier if we really want to take off. And then two, we’re at the very, very earliest innings of the opportunity and possibility for something like an Aviatrix. And we’re in the dominant pole position. We’re the largest specialty vendor in this area. We help define this category. Well, the category actually still needs to be blessed by folks like you, but we’ve been driving and evangelizing this need and trying to drag the market with us for so many years now.

I think we’re at that pivotal moment from everything I’ve seen in my six and a half months of companies that were hemming and hawing. It’s like, “Ah, I think I can just get by with what the clouds have provided me.” I’ve seen literally almost a hundred customers do what I call the boomerang effect at Splunk. At Splunk it was, “Let me go play with open source, because I can do all this stuff for free.” Then after 18 months, like, “Oh geez, man, it’s harder than I thought. And it’s not as high value as I want. I need my developers, my best guys focusing on the high value stuff. Let me go with Splunk.” I’m seeing the exact same thing with Aviatrix of, “I think I can probably just use these low level networking service and primitives, and they’re coming back saying, “Geez, it’s so much harder than I thought within a cloud. And by the way, I woke up and I’m in three clouds or four or five and I need you guys.” But that’s the earliest earnings.

I got to Splunk at a hundred million of ARR 300 million of revenue, because we’re perpetual licensing at a time and watched as we went from a hundred million to 3.2 billion by the time I left. And you can’t manufacture that. The market has to be aware of it. The market has to form and the patterns that I was thinking I saw coming in, but it helps much more when you’re inside and talking to customers. It’s like I think that we’re going to see that same market liftoff. And that’s the other message is we’ve got to keep with our technical buyer, reach the upper levels, and then have confidence that this market is at the beginning of the liftoff and we’ve defined it. We’re the gorilla in our little space. This is ours to lose. Let’s lean in with confidence and humility and serve these customers and make sure this market happens.

Daniel Newman: Doug Merritt, CEO, Aviatrix, great to chat to you. Have a great sales kickoff and thanks so much for joining me here on the Futurum Tech podcast.

Doug Merritt: Thanks so much for having me, Daniel.

Daniel Newman: All right, everybody hit that subscribe button, join me here for all of these interview conversations. Always great to have Doug Merritt join us. Hopefully he’ll come back in a year and tell us just how big this market is getting. But for this episode, for this show, I got to say goodbye for now. Thanks for tuning in. See you all later.