Aviatrix CEO pitches cloud networking to 'underserved buyer'

Three years after his departure from Splunk, Doug Merritt is a year into the CEO role at an emerging vendor, Aviatrix, and reflecting on how his experience in the IT ops market has influenced his more recent decisions.

The Aviatrix Cloud Networking Platform consolidates features including transit routing, subnets, network management automation, visibility and security into one package it claims is a simpler and more cost-effective alternative to traditional networking tools and built-in services from cloud providers.

In March, Aviatrix updated its Distributed Cloud Firewall with Kubernetes support. This month, it launched a partnership in which network function virtualization vendor Megaport will deploy Aviatrix routing, encryption and visibility services on its private edge network. And this week, the company hired former Google senior director of Google Kubernetes Service and Anthos Anirban Sengupta as its CTO and senior vice president of engineering.

Along with competitors Alkira, Prosimo, Arrcus, IBM and, potentially, Cisco's Isovalent, Aviatrix looks to harness growing enterprise interest in managing multi-cloud networking at the application layer, according to Shamus McGillicuddy, an analyst at Enterprise Management Associates.

"The challenge [for Aviatrix] is to [break] through the status quo of using native networking features and functions offered by cloud providers," McGillicuddy said. "Network infrastructure teams understand the value, but network teams often lack control and influence in the cloud."

That's where Merritt said his experience selling to IT ops pros at Splunk comes in, during an interview with TechTarget Editorial that also touched on his view of his departure from Splunk and his plans for the future of Aviatrix. 

I covered your departure from Splunk a few months after it happened and didn't talk to you directly then -- what is your point of view on what happened there?

Doug Merritt: I've honestly never gotten a cohesive answer from the board on exactly all the details of what led to the transition. What I have pieced together and what kind of makes sense is, driving any company, much less a public company, through a complete business model and technological shift is something that almost no companies do, for a good reason. It's incredibly convoluted and chaotic and challenging.

When that transition occurred, I'd been CEO for over six years. And I think the combination of a couple of executive turnover[s] and few new hires, all done during COVID -- important new hires, senior level hires; no face-to-face board meetings for a year and a half, almost two years … because of COVID; and all the changes we were going through as a company led to a very large board. We had added a number of people, and I didn't retire people. So I think at that point we were 12 or 13, with three or four relatively new members -- one a brand new member from Silver Lake and a few that were [there] less than 12 to 14 months. I think it led to the potential dissonance -- I didn't think that there was much of a drift, but there obviously was.

The observation I heard was that most of the tech industry made money during the pandemic, but Splunk earnings didn't meet estimates -- and if you're a public company, that's the way it goes. But it sounds like it was more complicated, in your view.

Merritt: When you are inverting cash flow, revenue and operating margin, and you've got dual businesses -- a recognize-[revenue]-up-front, on-prem business and recognize-[revenue]-over-time cloud business, that combination makes it extremely difficult to both predict the business and understand the business. We tried to use key metrics like annual recurring revenue [ARR], as cut-through metrics to give better visibility on what was really happening, and that growth was ridiculously insane -- on a scale close to 60% the first few years of the transition [to cloud]. When the transition occurred, it was still north of 35% and that was a $3 billion ARR mark, which was well above the growth rates of anyone before us -- ServiceNow, Workday, Salesforce, anybody.

As cloud got to be over 50% of bookings, and was nosing up to be 50% of revenue, it made an impact on revenue margin and cash flow because you're replacing $10 million recognize-up-front deals with recognizing a $10 million deal over three years on a day-to-day basis. [Former Splunk CFO] Jason Child and I, and [former Splunk CFO] Dave Conte before that, as we started this journey, said there's going to be a lot of pain. It's been pretty well written what happened at Adobe and a few others that tried this.

Shortly after my departure, [Splunk's next CEO] Gary [Steele] got very aggressive with some costs and spend[ing] but that just was kind of icing on the cake. For anyone that looks at the models and understands what's happening underneath, Gary's a great leader, but what got set up over the preceding three or four years was really what happened the year and a half [after] my departure. It's not much different than a Hollywood studio head that comes in when the last person gets sacked, and within the first year all these hits are published. It's like, 'Well, those movies actually started five or six or seven years ago and were obviously in the editing bin when you got here, so congratulations on a good release, but that really wasn't your movie.' But I loved my tenure there. It's a great company. And I'm excited for the shareholders that there was a high-value outcome with Cisco, and we'll see what happens.

Now that you've been in your new role for a year, how are you going about it, based on that experience?

Merritt: One of the key criteria [for joining Aviatrix] was [having] a known and trusted relationship with the board -- a small board, and I thought going back to a private company would be a great spot for me. The number of constituencies you deal with at a private company is very different. If you've got a small, tight board, that's just the investors and owners and you know them well, that makes the team aspect really come to life. In this case, three of the five board members I've worked with in the past as early investors at Splunk.

What I loved about Splunk was the core buyer that the company was founded around: IT ops teams, and their extraordinarily difficult job that very few people appreciate. … Splunk gave them a tool that elevated their careers. Aviatrix is at a very similar place with networking. … From my experience transitioning to cloud with Splunk and how difficult a microservices architecture is [to manage], how convoluted these cloud [networks] are, and how mission-critical networking is, I got an appreciation of [the fact that] you don't get a lot of insight and control on these cloud networking services. Aviatrix was a vendor to Splunk, so I got a tiny piece of that.

[Now] we serve this network engineer. They suffer from the same problem: cloud networking services are not organized and built the way that you would think they would be in a software-defined networking world. They didn't evolve that way -- it's [complex] and a lot of how it operates is hidden because it's all a service. So these networking teams are really under the gun. … We're back to, everyone immediately blames that network and networking team when something goes down. … And there were these characteristics around Splunk serving this under-served buyer ... that are really relevant with Aviatrix.

What's the pitch to that buyer? Why should companies switch out core networking functions from a big vendor or cloud service provider to Aviatrix? What's the advantage for them?

Merritt: The core initial pitch was ... even if you're in a single cloud, wouldn't you like to get much greater visibility on network traffic flows, the health of those flows, latency, a bunch of detailed information that's difficult to get right now? Wouldn't you like to get much more control? I know [cloud service providers] take care of everything for you, but you might want to optimize traffic in certain ways and filter in certain ways. And wouldn't you like more operational efficacy? You've got all these virtual boxes with each cloud, and the teams are pretty loaded trying to stitch them all together. We can do it with less people, higher visibility, higher control, giving you a lot of the core attributes that you knew on prem, but taking advantage of cloud. Then when you go multi-cloud, just square that equation, because each cloud [provider] has defined each of those [services] very differently.

The way that the typical application is architected, and the way cloud networking services work, you're exposed to a whole host of data processing and egress fees that we can avoid, because we're an integrated [package]. So we're not forcing packets into and out of all these different services, and we don't charge for data processing or egress. ... Depending on how you use cloud networking services, we can certainly at least be competitive with them. But in many cases, we're going to wind up being 20% to 70% cheaper.

What about cloud networking security?

Merritt: We launched our distributed cloud firewall in 2023 [to address this]. And unlike, really, everybody, including the CSPs [cloud service providers], things like network address translation and core firewalling and some of the filtering and egress that goes with that is all baked in. You don't have to use [Aviatrix] for routing to get that, but it's baked into the networking platform.

When I add all this up, the way that we're trying to frame it is the network is the cloud. We're not just going to deal with five major clouds -- [companies] will need to feather in the right LLM [large language model] architecture, which means data centers are going to be cool again. … Don't you want a consistent data plane and control, management and data analytics plane across that, given how mission-critical these workloads are?

Beth Pariseau, senior news writer for TechTarget Editorial, is an award-winning veteran of IT journalism covering DevOps. Have a tip? Email her or reach out @PariseauTT.