Founders on Fire: William Morgan, CEO and Co-Founder, Buoyant

Today we’re catching up with Buoyant, our winner of the 2021 Containers Trailblazers Award. We have the pleasure of chatting with William Morgan, CEO and Co-Founder of the firm. Chief Trailblazer Rose Ross quizzes him to find out more about how Buoyant came about, based on experience originally gained working for Twitter and their move to Kubernetes.

William shares his recommendations for staying focussed and what a software product should actually offer its users. He also describes how their pioneering service mesh, Linkerd, ended up as an Open Source project and how Open Source has evolved into much more than just being a way to avoid paying Microsoft for their products.

He also suggests how learning Go is useful in dealing with how to live with uncertainty, even if you suck at it. Listen to the full podcast here:

YouTube:

Also available on:

• Spotify
• Anchor

Interview transcript

Rose Ross: Hello everybody, and welcome to the Founders on Fire. My name is Rose Ross, and I’m the Founder and Chief Trailblazer at the Tech Trailblazers Awards. I’m delighted to be joined today by William Morgan, who is Founder and CEO of Buoyant, our Containers Trailblazer. Hello William, how are you today?

William Morgan: Hi, Rose. Very well, thank you. Thank you for having me.

Rose Ross: You’re very welcome. So, obviously, containers, the cloud native world is where you are now as an entrepreneur, could you give us a bit of a ‘how you got there’, how did you found Buoyant, what was your journey, how did you become an entrepreneur in the tech startup world?

William Morgan: I started as a software engineer, and I spent most of my life at various startups or big companies over the years, and I think probably most relevant to that is that I worked at Twitter as a software engineer, and then as an engineering manager there, in the early days, 2010 to 2014, something like that. That was at a time where Twitter was going through a very big infrastructure transformation, which is where I was, and the stuff that was happening there was so profound and transformative to Twitter itself, in its ability to actually stay up and not have the Fail Whale, that when I left Twitter I was fortunate enough to grab one of my colleagues, and the two of us said ‘basically, hey, what we just did there was so important, we think the rest of the world is going to need to do that too’. And that was the genesis of Buoyant, and everything we’re doing here is this particular set of experiences we had at Twitter, coupled with the realisation that okay, Twitter, at that point was a little bit ahead of the rest of the world. But, it wasn’t 20 years ahead, it was two or three years ahead. So that was the opportunity.

Rose Ross: I have to say, I’m a big fan of the Fail Whale, we in fact use a Fail Whale somewhat like the Twitter one, and we’re talking to people so we don’t miss out on things. But it’s a long time since I’ve actually seen one on Twitter to be fair.

William Morgan: Yeah, I mean it was funny because it became a cultural phenomenon, which was entertaining, but also probably shouldn’t be the case.

Rose Ross: No, a bit frustrating but at least it had a nice whale on it. So, you brought about the demise of the Fail Whale.

William Morgan: Well, me and about 2,000 other people, but yeah!

Rose Ross: You were part of that solution, and that inspired you and your co-founder to spin that out and go, ‘Hey, this is the stuff that other people need, let’s do that.’ That was the beginning of it, so you’ve not had any experience of founding an organisation before, but you’d obviously been integral in a lot of them, previously.

William Morgan: Yeah, that’s right. I had been early in some startups before and got a sense for not really the management or the company building aspect, but I got a sense for the chaos. I kind of half knew what I was stepping into, or at least I thought I did.

Rose Ross: Ah, well, baptism of fire, how did you find it with your colleague in the end?

William Morgan: I found it on the whole, amazing, and great. It’s like turning everything up to level 11 on the dial. And I think coming into that as an engineer, especially, you have to build some muscles around emotional EQ, I guess, managing your own psychology. I had to build up some of those muscles, but so far it’s been great.

Rose Ross: So, let’s take us on to Buoyant. You’ve talked a little bit about the problem that you were trying to solve. But do you want to iterate that a bit more clearly, and how you are now addressing that, and why you believe that that approach is obviously making you shine in our Containers Trailblazers alumni from last year.

William Morgan: I’ll try and keep it high level, although what we provide is a fairly technical solution in a fairly technical space. But obviously, we’re in the container world and we’re specifically in the world of Kubernetes, which is the next step after adopting containers, is well now I’ve got a lot of containers, how do I manage them in this aggregate fashion? Kubernetes, for better or for worse, is the way that that most people are starting to really tackle that. And it’s funny because Twitter was not using Kubernetes, it wasn’t even really using containers, but it was using some analogues to those things. And so a lot of what we learned there was kind of a match from the design pattern perspective, but wasn’t actually a technology match. Kubernetes was just getting started when we started Buoyant, but we attached to that very heavily.

So the idea is, a lot of companies are basically re-platforming on Kubernetes, or if they’re early, or companies are adopting Kubernetes, are greenfield, and it’s this particular way of deploying your applications, and managing your applications, where you abstract away the fact that there are a group of machines there. So, you guess there are machines that are running your applications, but you’re telling Kubernetes, ‘Here’s 20 VMs, you manage it, and I want you to run five copies of this, and 12 copies of this, and seven copies of that, and you figure it out’. So, that’s what Kubernetes gives you.

Our realisation at Buoyant was, well, that’s not actually enough to really run a full-fledged application, there’s a bunch of other stuff you need to do, especially around network connectivity between those applications. Because when people are adopting Kubernetes are typically not taking this giant monolithic piece of software and running it there, they’re breaking it down into microservices. They’re breaking it down into these components, those components talk to each other over the network, and in a coordinated fashion that is your application that you’re running. And those connections between those components are things that need to be managed and observed, they need to be secure, they need to have reliability semantics layered on top. And so all of that led us to coining this idea, which now is very popular, but at the time was a meaningless buzzword called the service mesh.

What Buoyant does today, what we’re most known for, is we build a service mesh, it’s called Linkerd, it’s Open Source, and it’s part of the cloud computing, Cloud Native Computing Foundation, the same foundation that hosts Kubernetes and Prometheus, and a bunch of other foundational technologies. And so what Linkerd allows you to do is, basically for a Kubernetes application, get observability, get security, get reliability, in a way that Kubernetes itself doesn’t provide. That’s not a failing on Kubernetes part, what Kubernetes got right was they had a very well defined boundary, they said ‘this is what we’re going to do, and everything else is going to be handled elsewhere’. And so Linkerd sits on top of Kubernetes and gives you some of those critical security and reliability features you need.

Rose Ross: Fantastic. And I did notice on LinkedIn, you describe yourself as a service mesh connoisseur.

William Morgan: That’s right. Well, Linkerd was the first service mesh, and now there’s all these other ones, there’s competition and all stuff.

Rose Ross: But you’re the champagne of the service mesh world.

William Morgan: Absolutely. Well, in that world what Linkerd is known for is its simplicity and lightweightedness. It’s very easy and here we get into some of my theories about infrastructure software, but it’s very easy to make infrastructure software that is the ultimate solution to all problems, and has every feature in the world, but when you actually try and run it, when you actually try and operate it, it’s a nightmare. Things break and you don’t know why, things fall apart. When you’re building this software, as an engineer I think especially, you want to focus on the power, you focus on, hey, we’re going to do everything for all people, and it’s going to be the ultimate solution! And you’re not really spending a lot of time thinking about, well what would it take for a human being to actually operate this, to actually have to wake up at three in the morning if this thing falls down.

It takes some discipline, and it takes some focus to put yourself in the mindset of, okay some poor human, their life is going to be determined by how this thing behaves, so what do we do to make this simple, what do we do to make this reliable, what do we do to make it something that you can understand, and build a mental model of, and operate as a person? And so that’s the philosophy we brought into Linkerd, I think it’s pretty unique in the infrastructure world, we’re just a little bit behind considering the humans in the loop. And that’s been a big part of Linkerd’s success.

Rose Ross: And obviously, having been around at this sort-of-beginning of the service mesh revolution, if you want to call it that, or the comfort blanket that you’re wrapping around Kubernetes, because it does sound…

William Morgan: Yeah, I like that.

Rose Ross: … or the people running Kubernetes, has obviously had you being very involved with the CNCF, the Cloud Native Computing Foundation. And the whole Open Source Initiative, we’re seeing a lot of that, it’s been around for a long time let’s face it, Open Source and such like. But it does seem now to be more and more embraced in so many different parts of the IT infrastructure around the way that enterprise computing works, whether it be from a security standards perspective, or for infrastructure, or having a cloud native approach, Cloud First. What do you see as the really important part for you, of why do you think Open Source has become so popular? Because a lot of people have always said, you’re not really sure who owns it, because it’s owned by everybody, effectively crowd driven. How has that maintained and built on popularity when people can be incredibly risk averse, particularly in corporate IT buying scenarios?

William Morgan: I think Open Source has evolved over the years, and the reasons why people do Open Source, and the reasons why people want things to be Open Source have changed too. And that evolution is not always recognised when I’m an older gentleman from an earlier time, and at that point Open Source was how we were sticking it to the man, like Microsoft was a big bad guy, we were running Linux on our computers, because we didn’t want to pay Microsoft money. And this was nights and weekends we’d contribute to the Open Source, it’s like this effort because it was this moral thing we were doing, we didn’t want big Microsoft to win.

And now, fast forward many decades, painful decades later, and it’s very different. Now, the nights and weekends approach doesn’t really work for Open Source. A lot of times if a project is just maintained by volunteers, you’re like, ‘Well gosh, it’s scary to want to adopt that, because what happens if those people get bored, or if they get a job that requires more of their time or something?’ So now the standard for Open Source is, well it’s good for there to be a company behind it that’s actually contributing, and that has a commercial incentive here. Because then you have this engine that’s continually powering the project. And I think what people have realised with Open Source, and part of why it works, is that the code itself is not actually that valuable.

The code itself is just this moment in time, it’s just this artefact. And for software in the modern world, software doesn’t stay still, you need it to be continually updated, the rest of the world is updating around it and changing. If you’d have a static piece of code, even if it’s Open Source, two years later it’s like it doesn’t work, you can’t compile it anymore because all the dependencies have changed. So, that means that we can be a lot more liberal with the IP considerations of, ‘oh well, here’s the actual source code’ – the source code itself is not so valuable. The valuable is, do you have a maintenance programme or engine that’s happening, that’s keeping the software up to date, that’s making sure it’s always usable by the rest of the world. So, that’s good for Open Source I think in a lot of ways.

Why do people want things to be Open Source? I think ostensibly it’s to reduce lock-in. In the worst case, if you get really annoyed, you can just run it yourself. But in today’s world, I don’t know how practical that really is. I think at this point the requirement for things to be Open Source is a little bit of a cargo cult, it’s like we think it’s the right thing to do, but the reasons why are not really right. But regardless, for me I no longer have a moral justification for Open Source, it’s a practical consideration. And for us, Linkerd, Kubernetes, Linux, everything in that world was Open Source. And so Linkerd had to be Open Source, otherwise it would have been weird, it would have stuck out like a sore thumb. And so if it’s going to be Open Source, let’s embrace that fully, let’s not be half Open Source, let’s really be Open Source and let’s take advantage of all of the benefits of Open Source, because the one thing you get there, and I think it’s really powerful, is you can have a community of people who are really enthusiastic about the project, who really feel like they’re a part of that project.

They bring it to the rest of the world. They tell their friends, they write blog posts about it, they feel very comfortable because it’s Open Source, and that’s effectively a marketing function, right? If you want to be mercenary about it, this is like, Okay this is how we’re going to do marketing, we’re going to have the Open Source community do this, instead of having a giant marketing department. Now, you can’t be too mercenary about that viewpoint, community is something you have to invest in, and for that to work you have to have a very honest and authentic relationship with your community. It’s very easy to lose their trust, and then poof, that’s all gone. So you have to be willing to really embrace what you’re going to have to do to make that a successful Open Source project, especially if you are then trying to build a company around the success of that Open Source.

Rose Ross: So, it’s multilayers, isn’t it?

William Morgan: Yeah, it’s a little nuanced.

Rose Ross: Indeed. So, aside from what you’ve been doing from that perspective, obviously as trailblazers you would consider the fact that you were so early, effectively set the boundaries of what service mesh is providing in this type of environment. What other trailblazing stuff have you guys been up to since you won your accolade with us?

William Morgan: I think the big innovation from us, obviously Linkerd continues, and we’ve got a lot of exciting features that we’re continually adding to the project. The thing we’ve been working on recently that I’m really excited about is on the operational side, and this is where the commercial counterpart to Linkerd comes in, we have a SaaS product called Buoyant Cloud that pairs with Linkerd and basically takes on the operational burden for Linkerd. And I really, really like this model because running software is painful, even if it’s Open Source, if you are responsible for maintaining it, if you’re responsible for keeping up to date with security patches, if you’re responsible for alerting and monitoring, that’s a lot of work, that’s a full time job. And a lot of that we can actually automate for you. And so that’s what we do on the commercial side is we say, ‘Hey, you want Linkerd? Great, you can run it yourself and that’s totally fine.’ And some people really want to do that, and that’s fine. Or you can allow us to run it for you. That’s what Buoyant Cloud will do.

Either way, you’re getting the benefits, it’s Open Source, it’s running on your cluster, you get security and reliability, and observability, and all the benefits of a service mesh, and you can decide, the engineering leader in the company – am I going to pay my engineers to maintain it, or am I going to pay Buoyant to maintain it? And I think that’s a really nice choice, it’s very clear what the boundary is between what’s in the Open Source, and what’s in the in the commercial stuff. Maintaining that boundary I think it’s really important, because you don’t want to cannibalise your Open Source community.

Rose Ross: And what verticals are you seeing? Obviously, you’ve had a customer base with the Open Source. Who is embracing this sort of managed service, Linkerd as a service?

William Morgan: It’s really interesting, because Linkerd adoption itself is really across verticals, it’s really more a function of who’s adopting Kubernetes, than what industry we’re in. So we have a lot of folks in FinTech, that’s a big one for Linkerd, and healthcare is a big one, because they’re both very interested in the security aspects. It’s very sensitive data that goes across a wire that they want to protect, and Linkerd gives them a good way of doing that, a great way of doing that. But we also see ecommerce retailers, so Nordstrom is there, we see video games, Microsoft, Xbox Cloud, their cloud gaming thing is using Linkerd in some massive way, using it for security.

I gave a nice talk at the last CoopCon, in Valencia, Spain about how they deployed Linkerd across all these clusters, tens of thousands of pods. And so it’s everything from… I’m here in Texas, there’s a grocery store, very popular grocery store called HEB and HEB uses Linkerd to power their curb side and their delivery service, and they rolled that out really quickly in the midst of the pandemic, because they wanted their customers to get groceries without coming into the store, and they use Linkerd to do that. So there’s all these different stories of people using it across verticals. I wouldn’t have expected any of this early on. I wouldn’t have predicted, oh yeah, it’s going to be in grocery stores and powering video games, and people buying clothing online.

Rose Ross: You can be clothed, fed, and entertained then. All via Linkerd.

William Morgan: Yeah exactly. You got it, you’ve got the trifactor.

Rose Ross: Got it all covered, brilliant. For you as an entrepreneur, as a former software developer, part of a startup environment moving into giving up the day job, so to speak, to build a company; what do you think you would pass on as advice to people who are perhaps a little bit earlier in their journey for that? What’s been important, if you have three things that you say that you think have helped you, either as lessons or as things that you’ve embraced as a principal?

William Morgan: I think the first lesson is being comfortable with ambiguity and being comfortable with the unknown. My perspective here is, I started as an engineer and then became a founder, and now I don’t write code any more, which I miss, every day I miss it. But I can’t, I can’t let myself do that, that would be a bad sign if I was writing code at this point and would probably be terrible. So I think the first thing is getting comfortable with ambiguity and with the unknown. As an engineer your instinct is you want to understand everything – here’s a problem, here’s a thing, well let’s figure it out. Let’s sit down and let’s come up with a master plan that takes into account every possible way that this could go wrong. And that’s what you’re trained to do.

When you’re writing software, you’re always thinking, what’s the edge case here. Well, what if someone puts in a letter instead of a number, what if someone puts in a negative one, what if the network is down? You have to be able to take a step back and say, ‘Okay, I can’t actually control for every variable, so let’s leave this area ambiguous, let’s leave this thing uncomfortable and unknown, and let’s focus on what can we control, and what are the contingency plans.’

For me, I actually took some inspiration, there’s a game, if you’re a huge nerd then chess is not cool enough for you, so you play this game called ‘Go’ which came from China, and Japan and Korea. I’m a terrible Go player, I suck at it. But the one thing that I learned from that game is that there is ambiguity there, and a lot of the way you play Go, you say ‘this area is ambiguous but it’s actually not important’. It’s more important for me to focus on this other area where I think the gains are going to be bigger. And that same mindset I think it’s really important. You have to have a really clear idea of what are the priorities, what’s important for you to spend your time on.

So that’s made my second point, which is you’ve got to get really good at prioritisation. To get good at prioritisation, you have to get comfortable at least with saying no to stuff that you know it’s good to do, but it’s not the most important thing to do. So there’s going to be too much work to do, not just you, but for your company to do. There’s always going to be stuff that would be good to do, so you have to be very ruthless and saying, ‘Here’s what we’re going to do, and this is good and important. And here’s what we’re not going to do, it might be good but it’s not important right now, and we’re just going to leave that on the table, because we have to.’ And it’s hard to get yourself mentally in that place.

Third one, I alluded to this early on, but emotionally you’ve got to manage your own psychology. I think it kind of depends on where you are as an engineer. Over time, if you progress up the engineering career ladder, you get into more and more of a leadership position, even if you’re still an engineer, and those same principles apply. You’re going to have people looking to you, and trusting you, they’re putting their faith in you. If they’re joining the company a lot of that is they’re putting their faith in you as a leader, and you might be really nervous, something terrible might just have happened, you might have just lost some big deal. You might be totally freaked out on the inside, but a) you have to let that stuff go, and b) you can’t communicate that outward. People can’t be like, ‘Oh my God, this guy, he looks like something terrible happened, what’s going to happen to us?’ You still have to inspire confidence and put on the show, even though at times it feels a little hollow. And then 24 hours later you’re like, ‘Oh, I’m on top of the world, this amazing thing happened!’ So you have to realise that there’s this emotional cycle that’s going to happen. For me it’s every 24 hours, I switch between feeling awful and feeling amazing about how things are going, and you’ve got to internalise that, and not let that affect your outward communication.

Rose Ross: Well, if you’re having one of those bad days, you’re doing a very good job of covering it up today.

William Morgan: Oh yeah! I’m dead on the inside.

Rose Ross: I know, I can tell! (laughing) You had actually turned the light on behind you to help, just to give you a bit of a zizz. But no, I do understand what you mean. It can be quite draining, and as you say, you can’t bring that to the table.

William Morgan: Yeah, not when you have people relying on you and people that you need to be in good spirits, to be energetic, enthusiastic, and bring their best souls to work.

Rose Ross: Lots of really good lessons there from you William. Is there anything else that you’d like to share today, whilst we’re still chatting and the green light is still on, so to speak?

William Morgan: Well, if anyone out there is using Kubernetes and would like an incredible service mesh then come talk to me, or just search for Linkerd, there’s a lot of cool stuff that we can do for you. Yeah, that’s probably that’s probably the big one. Always, always interested in more Linkerd adopters out there.

Rose Ross: Well, we’ll be rooting for that.

Thank you very much. My name, as I said, is Rose Ross, and I’m the Chief Trailblazer and Founder of the Tech Trailblazers Awards. I’ve been delighted to be joined here today by William Morgan, who is CEO and co-founder of Buoyant, our Containers Trailblazers winners. If you’ve enjoyed what you’ve heard, do reach out to William and the team at Buoyant, and obviously if you’re interested in being part of the awards this year, please do look us up on www.techtrailblazers.com, find us on LinkedIn, or reach out to us on Twitter. And no Fail Whales please with that one, we’re trailblazers. We welcome you all, and we have a deadline of the 31st of August. So we look forward to hearing from everybody before then.

And thank you very much William for joining us today.

William Morgan: Absolutely, thank you for having me Rose.

Rose Ross: Brilliant. It’s been a pleasure.

William Morgan: Likewise.