DevOps Chat: Enterprise Multi-Cloud Backbone With Aviatrix

Our podcast guest believes the world changed “on a Tuesday” six months ago when the enterprise move to the cloud became imperative, not just an aspirational goal. That eureka moment was enough to pull Steve Mullaney off the sidelines and back into the game as CEO of Aviatrix after successful executive gigs at Palo Alto Networks and Nicira.

“Go build” might be a wonderful approach for cloud-native apps, but enterprises want the framework laid out ahead of their move to the cloud. Think enterprise cloud IT reference architecture, much like Cisco provided for enterprise IP networks, DEC mainstreamed for client-server and IBM established for mainframe computing. You have to bring the cloud to them. That enterprise cloud reference architecture is what Aviatrix targets, in the same buy-with-a-credit-card model that fueled customers’ unfettered move to AWS.

On our DevOps Chat podcast episode, Steve Mullaney, CEO of Aviatrix, talks about what enterprises want in a true enterprise multi-cloud backbone to support their needs.

As usual, the streaming audio is immediately below, followed by the transcript of our conversation.

Transcript

Mitch Ashley: Hi, everyone, this is Mitch Ashley with DevOps.com, and you’re listening to another DevOps Chat podcast. Today, I’m joined by Steve Mullaney, CEO of Aviatrix. Our topic today is enterprise multi-cloud backbone—meaty, good stuff.

Steve, welcome to DevOps Chat.

Steve Mullaney: Thanks for having me, Mitch.

Ashley: Great, good to have you here. I’d like to start by having you introduce yourself. I know a lot of us know you from your background, you’re a fairly well known guy, but tell us about yourself.

Mullaney: So, yeah, my name is Steve Mullaney, I’ve been in networking and security infrastructure for, really, the past 35 years. I actually started my career in the mid-80s as we were kind of evolving from, transitioning from the mainframe computing model into client server. I was actually an engineer on 10Base-T from the company called SynOptics. So, those of you who’ve been around long enough to remember when Ethernet was starting and—

Ashley: I do recall.

Mullaney: 10Base-T started, yeah. So, I spent most of my career in networking and security, I was a very early guy, actually, also at Palo Alto Networks, I was the first VP of Marketing there, employee number 25. And then was CEO of a company called Nicira, which was really kinda the beginning of a lot of this SDN craze and network virtualization. We were acquired by VMWare, and then I stayed at VMWare for a couple years and grew that business, which is now almost a $2 billion business and a ________ business.

And then, for the last five years, I just, I kind of said no more operational roles and was on boards and I was content to just do that the rest of my life and very happy living the dream. And I saw the world change about six months ago, where enterprises have been talking about the movement to a cloud computing model—honestly, talking about it for the last 15 years, but I noticed really about 6 months ago is when mainstream enterprise really said, “Okay, we are doing this.” And I just looked at that opportunity and said, “My God.”

And I was on the board of Aviatrix at the time, so I could see this happening up front and just said, “This is just gonna be too much fun not to be involved.”

Ashley: So, what was it? What was it six months ago or so? When was the nexus point when you saw a change that pulled you out of retirement?

Mullaney: Yeah. What was interesting was, I love to say, six months ago on a Tuesday. And it felt almost like that much—like, literally on a Tuesday. And that’s how enterprises move, right? They talk about it, they talk about it, they talk about it, they talk about it. It’s like going on a diet. You talk about it and, “I’m gonna go on a diet.” “Well, how about today?” You’ll say, “Well, you know—not a good day.” “Well, how about tomorrow?” “Well, you gotta start on a Monday,” right? And then, “How about next Monday?” “Well, that doesn’t really work, because I’m going on vacation.”

I mean, there’s all these excuses, and enterprises, about five years ago, really said, “Alright, we are gonna move to the cloud.” And honestly, that’s when every enterprise vendor went, “[Gasps] We’re dead!” Right? “We’re dead. Everything’s gonna go into Amazon and we’re dead.”

Ashley: Mm-hmm.

Mullaney: And what happened was nothing, because the enterprises were just talking. They didn’t really mean it, right? So, everyone went along their merry way with quarter after quarter of record highs. Until about six months ago, I’m on the board of Aviatrix, a bunch of other infrastructure companies and I notice the logos are changing. This isn’t Netflix any more, this isn’t small to medium businesses, these aren’t early adopters. These are Midwest manufacturing, conservative companies. What the heck is goin’ on? They shouldn’t be doing this!

And people think just because AWS is at a $30 billion run rate that we must have crossed the chasm by now. We haven’t. And I’ve noticed, across all of my infrastructure companies that I was on the board of, they’re seeing all the same thing. No one’s done anything different, it’s just, the market just changed. And the way enterprises go, they are a herd mentality. When the majority, you know, the early and late majority of customers all decide to do something, they all decide on the same Tuesday morning, and I saw that happen.

We saw it with client server. I was at SynOptics at the time. Mainframe computing was the way you did enterprise computing, and client server—PC client server was treated as a toy, just fun and games. So, print sharing, right? PCs—that’s not real computing, right?

Ashley: Mm-hmm, mm-hmm.

Mullaney: And then what happened? All of a sudden, in the early ‘90s, on a Tuesday morning, everybody decides the Internet Protocol is the only protocol that matters. All the other protocols went away overnight, and PC client server is the way I’m gonna build up my architecture—and it happened overnight. I saw the same thing happen with cloud.

Ashley: Mm-hmm. So, it’s really kinda going mainstream, you’re seeing a real—

Mullaney: It’s gone mainstream. And I’ve been at the company now for three months as the CEO of Aviatrix and every day, I just see more and more evidence of that, that companies are burning the boat—meaning, the corporate data center, the corporate backbone that is dead. It is now an expense. I view it as an expense, and I view anything I do in the public cloud, leveraging the public cloud, as an investment.

Ashley: Mm-hmm.

Mullaney: And that means all my people are focused on that as well as all my expenses. And it shifts overnight, and it’s a herd. Every single enterprise around the world is now making that same decision on the same Tuesday morning.

Ashley: Mm-hmm. I love your Tuesday morning analogy.

Mullaney: Isn’t it funny? It’s really true, though—it’s really true!

Ashley: [Laughter] It just seems like it suddenly happened.

Mullaney: Yeah.

Ashley: So, you’re on the board of multiple companies. What was it about Aviatrix that you said, “This is the one I’m gonna ride. This is the one I’m jumping on board, I’m gonna be CEO, take this baby where it’s gonna go”?

Mullaney: Yeah, I think it goes—because it goes in terms of what we say our mission, our mission is to build the enterprise multi-cloud backbone.

So, when you look at AWS and Azure and GCP and even AWS’ mantra to enterprise is, “go build.” Well, that’s great when you’re an early adopter and they hand you the power tools, and you—

Ashley: Mm-hmm, the greenfield.

Mullaney: Yeah and you are that innovative, you wanna go build. That does not go very well when you’re a Midwest manufacturing, conservative company. I’m gonna cut my arm off with power tools. I don’t want power tools. I want a house, I want furniture, I wanna walk in with it done. I don’t wanna build anything.

Ashley: I want a landing strip, I don’t wanna build the runway.

Mullaney: Absolutely. I do not wanna build. And so, I think this, honestly, has caught AWS off guard, too, right? Because they’re thinking—all of our customers love that, a “go build” mantra. Not the mainstream, right? Not when it flips. Not when you’ve crossed the chasm and now you’re in the tornado and everybody just, they’ve gone and said, “This is it.”

And so, I looked at that, and I said, “My God, all of these enterprises are begging—begging—someone to define the architecture for them.” This is just like in the client server where they needed Cisco to come out and define the networking and security architecture for them. Give me the canonical architecture that everyone’s gonna do, and then I will go implement that.

Ashley: And it’s proven, we’re not inventing.

Mullaney: And that’s what people want, and they want a validated design. They wanna be able to go, “This is what everyone’s doing, we’ve all agreed, this is what you do. And oh, by the way, I can’t get that from Amazon, because Amazon’s trying to lock me into Amazon. I can’t do it—I can’t get it from Azure or Google. I can’t get it from the old world, I can’t get it from Cisco, because they don’t even understand this,” right?

Being able to—it’s just like, you wouldn’t go to IBM or Deck for the client server architecture. This is a whole different computing model. You can’t—by definition, the leaders of the old are never the leaders of the new. So, who do I go to?

Ashley: Yeah, and to the point of multi-cloud, you’re probably not gonna be in just one, yes? It might be appealing to be in Azure for some Microsoft apps, but you also have compelling reasons to be elsewhere.

Mullaney: I have not met an enterprise that says they’re only gonna be in one cloud. They’re in at least three. Because then you’ve got, if you’re in China, you’re in Alibaba. Then you’ve got, you know, data sovereignty and GDPR and all these kinda other issues that I may have to have multiple providers in Europe, country by country, right? And not even AWS.

And so—sure, absolutely, it is going to be, and it should be. Because not that you’re gonna move workloads between one to the other. Like, that’s not gonna happen. But what is gonna happen is, maybe the marketing team likes GCP for AI reasons or, you know, I started with Office 365 in Azure and so I’m moving off on that. Or I’m a big enterprise, manufacturing enterprise and Microsoft really understands me. Or I’m a retail and I will not have anything in AWS, right?

Ashley: Yeah.

Mullaney: Or I see, even, I’m a customer, I’m a company—a software company. I’m selling to other enterprises. I cannot have my infrastructure if I’m selling to retail. I cannot even have my own internal infrastructure on AWS because they don’t want that to happen.

So, absolutely, it has to be across all this. And then the other thing is the “go build” mantra—so, even if you could say, “I can figure out how to go build on AWS,” the constructs in Azure and GCP are different.

Ashley: Right, you’d have to go build everywhere.

Mullaney: So, now I need to learn those constructs, and now I have to go build there, but with different tools that work differently—like, one’s in metric and one’s in an American system. It’s like—oh, my God, you know? And the tools don’t work. Like—okay, this is ridiculous.

Ashley: So, you painted a picture of multiple compelling reasons somebody needs to lay the groundwork for me. I’ve gotta be multi-cloud, so I need that reference architecture that’s gonna work across multiple providers. I don’t wanna stick build any more, I wanna be able to walk in and plug my application into a framework, a reference architecture that I can use—that’s what Aviatrix is about?

Mullaney: Yes. And then layering on top, leveraging the constructs, right? Leveraging the wonderful underlay, right? Of all—think of all the fiber, and the pops, and the data centers, and the regions, and all the wonderful stuff underneath that AWS Azure and Google provide, and growing.

Ashley: All that infrastructure you don’t have to do.

Mullaney: That is unbelievable, right? And guess what? I’m an over the top networking and security software player. So, when you think of guys like WhatsApp, you know, AT&T used to always get really mad because WhatsApp gets bought by Facebook for like 30 billion or whatever it was, and AT&T with all their fiber and all their different pops and all their right of ways and everything else like that—what’s their market cap? I don’t know. Probably not even much more than that. And you go, “Why is that? Like, I do all this stuff and I get no credit.”

Ashley: Spend a lot more money getting there.

Mullaney: And then WhatsApp with, you know, six engineers gets bought for $30 billion, where’s that? Well, yeah, because the over the top guys, that’s where all the value is. They’re leveraging it.

So, in a sense, we are an over the top software player, but what we’re over the top of is not the internet on top of AT&T, it’s the new internet. It’s the hyperscalers. It’s AWS, Azure, Google, et cetera, all with massive bandwidth and the best latencies you can get in peering points to each other such that when we overlay these decoupled security services on top, the combination is an amazing architecture and network.

Ashley: So, talk about those services. Talk about security operations, extensions, transit—all of those things that you need to take care of, orchestration—how does that work with ABHS?

Mullaney: Yep. Yeah, so in this what we call the enterprise multi-cloud backbone, there are, you know, today probably seven or eight different services that a customer can start with. And the beautiful thing about this, again, you can use Aviatrix. We have over 300 customers right now and growing rapidly all over the world—customers all over the world. We don’t even know that, because this is low friction land and expand. A real true cloud model where they can actually start. It might be $500 a month. It might be $50 a month. It also could be $3,000 a month. They just start using it, and they might stop with one use case. “Let me try it out.” They get their credit card, it integrates in through AWS Marketplace or the other marketplaces, and all of a sudden, they just start charging. And guess what—next month, they add. Next month, they add. Next month, they add. And then they say, “Oh, I now need additional things,” right?

So, they may start with a service we call user VPN. So, basically, we take OpenVPN and we add a lot of additional functionality to it, integrating it with SAML and other authentication means, add in some more security controls, and say, “Hey, I’ve got users, and I want them to VPN into the cloud, because I have resources in the cloud. I want them to be able to securely connect them to the cloud.”

That might be the first use case you start with, and maybe you’re paying $1,000 a month, but from there, it’ll start growing where then you say, “Well, now I need transit networking,” right? Or, “I need egress filtering. So, guess what, I start going into the cloud.” Guess what, most of your VPCs and your resources in the cloud have to access things on the internet for build packages or et cetera, right? Because this is a very distributed, hybrid world. So, it’s not just all inside the cloud. Well, guess what? I wanna do filtering on that outbound, and I wanna do what’s called FQDN, fully qualified domain name, which, I don’t wanna do filtering based on an IP address or let them go anywhere they want. I wanna say, “You can go anywhere, yahoo.com or wherever you need to go for that build package.”

Ashley: Mm-hmm.

Mullaney: And that’s it. So you wanna be able to whitelist it, and you wanna be able to define it per user and per VPN. So, you wanna be able to have security policies of who can go where. So, that’s another use.

Other use case is encryption. You know, right now, people say, “Well, you run”—you know, six months ago, when the cloud was kind of fun and games, “Hey, I’m putting workloads up there, it doesn’t really matter, it’s not my enterprise.” You know, maybe you didn’t worry about encryption. But now, you’re a retail or you’re a bank, you know, and all of a sudden, you care. So, you’re gonna wanna—

Ashley: Yeah, if you’re gonna put in an enterprise application, you’re gonna miss encryption, security, it’s high [Cross talk].

Mullaney: Yep, yep. You’re gonna wanna encrypt in flight, right? So, not only do we allow encryption, but we also allow high performance encryption, right, to 10 Gig, 20 Gig, 30 Gig and more, where—you know, and that’s important as people are putting PII data and other things. And also, from a security perspective, people know that—look, I’m just gonna encrypt everything. The data as well as in transit.

And then, you know, as you talked, being able to do this across multiple clouds, right? So, I maybe can connect, I wanna be able to connect up multiple regions within one of the clouds, but then I also wanna be able to go to the other clouds as well and create one kinda common framework upon what we’re doing.

And then I’d say another service is what we call firewall network service, and that’s another one, I can talk more specifically about that. But everybody that’s coming to the cloud—again, this isn’t fun and games any more, this is serious business. The first thing they say is what? Security and policy, I need to bring my next generation firewall policies into the cloud with me. And I need—so, if I’m a Palo Alto Networks customer, the first thing I wanna do is, I wanna take my VM series into the cloud. And when I try to do that, the construct of how AWS says you can connect your VM series firewall into AWS forces compromises in performance, scalability and visibility that are just untenable. And, with Aviatrix, we eliminate those trade-offs.

Ashley: It’s certainly a fundamental component of where you’re gonna start to build out your infrastructure. If you can bring as much of your own network security into the cloud, all the better—that’s your strategy, yeah.

Mullaney: Yeah. Well, you know, I think what’s important is, and I think guys like Amazon are realizing this, is when you have to deal with, when you’re dealing with the enterprise—and I mean, the enterprise. The Midwest, conservative, classic mainstream enterprise, right? You have to bring the cloud to them. You can’t make them go to the cloud, right? They’re looking at this and they’re saying, “Okay, I’m all in. I’ve burned the boat, I’m gonna get rid of my data centers, I’m all in.” But then they look and they go, “The grass ain’t as green as I thought it really was.”

Ashley: Mm-hmm.

Mullaney: There’s a lot of burned out spots, here. This “go build” mantra? Whoa, whoa, whoa, whoa—wait a minute. I thought I was gonna hit a couple clicks and be done. And oh, by the way, CEOs and CIOs of companies, they’re all bought in and said, “Hey, you know, we used to have hundreds of people managing this infrastructure when it was on prem, but I understand cloud. I only need tens now,” right?

Ashley: Mm-hmm.

Mullaney: So, the problem is, [Laughter] they go to the cloud, and again, I’m seeing this everywhere, every enterprise said, “We don’t really have enough people.” Like, we got kinda sold a bill of goods, here. Everyone kinda told us this was like click, click, click—you’re done. It ain’t that easy, right? And then you throw across, “Now, I’m gonna do it across multiple clouds—I don’t know how we’re gonna do this.”

Ashley: Well, let me ask you, then, we’re coming up on the end of our time here, together—if you had one parting recommendation thought for those enterprises that are in that first six months window and really looking at moving to the cloud or are in that process, they wanna check out Aviatrix, they wanna maybe do a prototype or look at the reference, how they build a reference architecture on the platform—how do they get started? What’s the best way to do that?

Mullaney: I mean, the first thing I always tell people, the best way to get started is actually just get started. The good thing about this modern world is, we don’t have to do a POC. We don’t have to—you can actually get started over the weekend, and that’s how many of our customers start. They just start playing around with our software, like on AWS Marketplace, it’s a great place to start. They just spin up a controller, they spin up a couple gateways, they—and we have quick start guides to show you how to do it in a very easy way. And people start playing with it and they go, “Oh, that’s pretty interesting.”

And so, they start with kinda one use case, right, and a very small little bit and they effectively do their own little POC. And they say, “Okay, that’s pretty interesting.” Now, we can also do it for them, but that’s a pretty good way to start.

And then I’d say the other thing is, what you really are gonna look for—I mean, there’s only one reason why people aren’t using Aviatrix right now is because they’ve never heard of us. As soon as they—as soon as we talk to people, they’re sold. Because they look and they go, “This is exactly”—people say, “That’s exactly what I’m looking for. I need someone to help me build out my multi-cloud, my enterprise—my, my, not someone else’s, my enterprise’s multi-cloud backbone, and all these networking and security services that go along with it. That’s what I need, and you guys are doing it for me? That’s exactly what I need.”

Ashley: Well, great. I thank you so much for being with us. We’ve reached the end of another DevOps Chat podcast. It seems like there’s never enough time. I’d like to thank you, Steve—Steve Mullaney, CEO of Aviatrix for joining us.

Mullaney: Great.

Ashley: And thank you, you, to our listeners for joining us also, of course. This is Mitch Ashley with DevOps.com. You’ve listened to another DevOps Chat. Thanks for joining us.