Red Hat CTO - Chris Wright

(Joel Beasley at 00:00:00) Hello, my friends. Today we are talking to Chris Wright, the CTO at Red Hat, and we discussed the evolution of the open source movement, an amazing explanation of cloud computing, and how the organization is planning an all-virtual experience for Red Hat Summit 2020 on April 28th and 29th. All of this right here, right now on the Modern CTO Podcast. Here we go.

(Joel Beasley at 00:00:27) This is the Modern CTO Podcast. Are you getting your exercise while we're all locked in?

(Chris Wright at 00:00:41) Yeah, I am, actually. So two things. One is, I'm a cyclist. I'm kind of a cyclist nerd, so I have a contraption that if you're not a cyclist, you might not have seen it before. They're called rollers, and rollers are a set of tumblers that you can literally ride your bike on. So there's a belt drive between the front and the rear tumbler, and when you're pedaling, it's rolling those tumblers and you're just riding your bike inside, and it's a much more real feel than sitting on a trainer where you're kind of clamped in and maybe you took a wheel off to make it all work. So I do that. I got my own little weights, you know, kind of the full home gym. And then I've still been going solo, not with groups, but solo doing some outside cycling and certainly just walks through the neighborhood. So trying to stay active and keep—it's just part of my mental sanity, so it's kind of a critical part of life.

(Joel Beasley at 00:02:01) Me too. The moment that I switched and the gym was closed, I was like, I have to figure this out. So I went and dusted off a yoga mat, right? Put it in my garage, found I had one seven-pound weight and two fifteen-pound weights. This will do. I can make a workout around this. And yeah, you just do it. Right? That's how we operate as people.

(Chris Wright at 00:02:29) Yeah.

(Joel Beasley at 00:02:30) So, you know, I like the conversation that I've been hearing a lot about the different technologists, but how we're pretty used to working remote and from home. Was that the case with you and your teams?

(Chris Wright at 00:02:44) 100%. So with a key caveat, personally, I've been remote for over a decade, almost the last fifteen years. About a year and a half ago, I actually moved, and relocation was, ironically, to spend more time in an office. And so I'm out here in the Boston area. I'm from Portland, Oregon. And going back to working from home for me has been really, really easy. I have such a long history of being a work-from-home person that things like I'm at home, so that means, you know, the door to my office is closed, and my kids just intuit what that means. It's not a new experience for anybody in the house. What is different is we're all home, and we're in this physical isolation. How do you stay social at all kind of time experience that is different? In the past, it's either school and everybody's off during their day, or it's summertime and it's relaxed and sleep in and go play with your friends and do whatever. So that experience is new. And then if you take my experience and just magnify that across Red Hat, we have a full spectrum of people who are, like me, have spent a lot of time working from home or people who have exclusively gone into the office. Everybody has—there's some new twist to this current experience. Red Hat in general is a very distributed company, and the way I like to say it is, we really mirror the external communities that we're a part of. So if you think about it, a community is this distributed thing, right? It's not one place, one location, one company. It's a whole global experience. Red Hat internally is quite a bit like that. And so I think how you communicate digitally, the kinds of practices, best practices that you build over time to really foster that distributed workforce is more natural to Red Hat, but it doesn't mean that it's somehow magically easy for everybody. You know, this is a different time. So there's a lot of challenges, and you've got the toddlers at home that can't understand why mom or dad aren't available right this moment because they need their parents' attention. And, you know, that really is different. So it's been a pretty interesting transition.

(Joel Beasley at 00:05:24) One of my favorite parts is seeing all the executives that I usually interact with in person on the cameras without their suits on. I'm like, yes. I was like, that's what you look like as a human. I love it.

(Chris Wright at 00:05:36) Yeah.

(Joel Beasley at 00:05:38) I usually don't wear a suit when I'm programming, but I definitely like when I get to see people in their raw self, like who they are. And then I like the people who you can tell when you meet them that they are exactly who they are. I like those people a lot too. I get that vibe with you a lot. When I see you and hear you talk, and when we were talking before a couple days ago or last week, when I heard the passion in your voice about what you're doing, I'm just like, this person is an authentic individual, and I like talking to those types of people.

(Chris Wright at 00:06:12) Thank you. I strive for authenticity. It's one of my goals, so I appreciate that.

(Joel Beasley at 00:06:17) Knocking it out of the park. So what are you working on these days?

(Chris Wright at 00:06:22) Well, my history is very much in software engineering. As CTO for Red Hat, my job is—it's simultaneously exciting and daunting, right? It's this big task. We have a very large product portfolio. We have a ton of technology that we incorporate into our product portfolio and, as a result, pay attention to those particular areas. So my role is to keep our technology strategy fresh and moving forward in the context of changing, emerging technology trends as well as changing, emerging market ways that the market consumes technology. So in one sense, my job is static in that I'm always looking at emerging technologies and emerging market spaces, but because of the fundamental changes associated with that, things are always changing, things are always new, and it gives me a lot of opportunity for learning. Right now, ton of focus on—besides how do we work well remotely from home—the transition I think the industry is going through, which is very much around data as a central actor to our customer base, the enterprise. So what does data mean to the enterprise? Who's creating data? How do you process data? Machine learning is almost intrinsic in that conversation. It's not always about machine learning, but it's really, really a common part of the conversation. And certainly open source, the technologies that are evolving in open source, containers being a really clear area where projects like Linux and Kubernetes combined have really changed what people think about when they think of computing platforms and application platforms. You know, those are the areas where I'm spending my time thinking of how you connect it all and what are the impacts on Red Hat's long-term view in terms of our products and our portfolio and what our customers need to be successful. So kind of a lot of stuff. Fun but daunting at times.

(Joel Beasley at 00:08:41) Yeah. That's why we need that exercise. Keep our mind straight.

(Chris Wright at 00:08:45) That's right. It's a little meditative space where you can collect your thoughts, get creative, maybe just check out.

(Joel Beasley at 00:08:53) And I find—you know what? I've had this internal dialogue recently about how when I'm working out and it becomes difficult, you know, right before you start to really feel good, it's like my mind's focused on one singular thing. It's like that workout. It's pretty clear. It's like, you know, you do the reps, you lift the weight, and it's just this unique, almost meditative state. And I feel like that, like relaxing your brain like that gives you more energy for later when you have to solve tough problems. Because I've kind of been imagining it like it's like a processor. But if it's overclocked, if you're always thinking, thinking, thinking, thinking, thinking, but when you create these separations, when you create these areas of time where you're not thinking about that type of stuff and you're not thinking at that level of depth, and then you go into this concentrated—I do work blocks. I don't know how you do it, but I do these work blocks throughout the day, and I structure my day like that. But then I noticed my mind's stronger. It's like a muscle. If you're always working it out, it's never going to heal.

(Chris Wright at 00:10:02) Right. I agree. I was thinking about this recently. Where do ideas crystallize for me in general? And it's partially the physical location, partially the activity. It's certainly not always in the act of working, and sometimes it is, and that's because of a focus work block like you described. You've really got an intention, and I'm working on or working with an idea. The two places not work-related that ideas really surface for me or clear thought crystallizing, how I think about a problem—one is on the bike, so in the saddle, the other's in the shower. I was thinking, okay, what does that mean? What's going on in those two spots? And one, I don't know if this is important or not, but in both of those cases, I'm totally disconnected from devices, right? There's no outside stimulus from emails and texts and, you know, the desire to go scroll down a Twitter feed or something like that. And associated with that is maybe just letting your mind wander. And specifically on the bike, when you're really pushing yourself hard and the only thing you're doing is focusing on how much your muscles are burning and breathing, there's just nothing else happening. And I kind of have the sense like you were describing. There's something in the back of your brain that's either relaxing or is completely not needed, so it can work at these other problems or, you know, it just creates a new space. And I think it's really critical. I think it's fundamental to even maybe the creation of ideas and how you create some separation so you don't just burn yourself out, overclock and overheat.

(Joel Beasley at 00:11:57) Yeah. I think that the science will catch up eventually, right? One day, they'll be able to measure it somehow, and then it'll be like breaking news. You need time off from work to do better. Yeah. The new productivity study. So I'm excited to be talking with you a lot. I like the future. I'm a huge fan of the future. I'm not a fan of the Boston Dynamics taking over the world future. I am a fan of Boston Dynamics building really cool things future. But I want to know because you're in there, man. You're in there with data all day. You're in there with cloud. You're in there with what customers are doing on the edge. I want to know what's going on in the future.

(Chris Wright at 00:12:39) Well, that's an open-ended question that I get asked a variant of a lot. And one thing that I like to say that's not as exciting is there is a key element to the future, which is, I'll say, less disruptive and more an evolution of a series of ideas, many of which we already know. And the reason it's not as exciting is it's not like, well, you know, robots and spaceships and, you know, imagine this shiny new world that we can't imagine today. But if you think about how new ideas are formed, so much of it is leveraging our current experience. And so to me, in the context of open source software building infrastructure and platforms for enterprises to build their businesses upon, I think the future is a combination of a lot of devices. You've heard a lot about the Internet of Things or the Internet of Everything or all these kind of buzzwords that give you the sense that device proliferation is a key part of the future. Devices create and consume data. In order to create and consume data in any way that's meaningful for the rest of the world, they need to be connected, so there's a network piece that's really important there. The piece that's processing all of that data historically has been compute platforms in the data center. And I think a big shift right now is moving those same compute platforms closer to the data, and we're calling that edge computing, although it's had other names at various points in time. And most broadly, you just call it distributed computing. And the data processing being really what enterprises are looking to to build value for themselves. So data by itself, not necessarily valuable. Understanding patterns and trends in data can be particularly powerful and valuable for businesses. So I think the combination of a smartphone and a 4G network and a cloud created about a decade ago opportunities for businesses to do things that we take for granted today. Ten years ago, we couldn't really imagine. I think that the combination of device proliferation, not just smartphones, but other devices, 5G connectivity, which changes some of the network dynamics, and edge computing give rise to that same—what is it going to look like in ten years? A little hard to imagine, but I know that these are fundamental building blocks to creating a new wave of, you know, fundamentally changed businesses. And you'll hear a lot about augmented reality and virtual reality and autonomous cars and drones and things that are—they're fun to talk about, but they're—I think it's just much more pervasive and a really practical application that's topical today would be telemedicine. How do you connect doctors and patients remotely in the context of a pandemic? This is a really timely discussion to have. And so I think that we're building refinements on what we've already known, and those create building blocks for what you might call disruptive businesses. But each technology change itself isn't sort of fundamental and, you know, amazing. It's more incremental and an evolution of ideas, which I think is super powerful.

(Joel Beasley at 00:16:38) You've got my mind going now, right? So let's boil it down. Let's talk about one specific thing. I guess what was going on in my mind was this concept of the edge. Like, what's the edge going to look like in the future? So let's first talk about—I first want to know, help me build a visual mental image of what the edge is today.

(Chris Wright at 00:17:07) Okay. The edge is a good one because it's fuzzy. It means different things to different people. So how do we bring that into focus? I describe the edge—if you imagine an onion and each layer in the onion has an edge. If you go across the industry and talk to different people, their perspective or their context or the layer in the onion in which they live will help them define what they call the edge. So you could define the edge from within a large data center being—a portion of your edge being what's connecting you out to, say, a public cloud. People might call that the data center edge or the cloud edge depending on whether you're coming from the cloud inwards or from the data center outwards. You have, when you get into a single business distributing their compute across a set of remote end offices and branch offices, sometimes called ROBO, each of those compute stacks sitting in those stores could be called edge compute platforms. If you go into a manufacturing automation scenario, the machines that are sitting on the auto manufacturing floor helping automate the process of manufacturing goods could be considered edge.

(Chris Wright at 00:18:35) They're certainly not in the data center of the manufacturer. The telecommunications industry has a network. The network has an edge. That edge, often called the provider edge, is providing services to their customers. Provider edge is a place that could go all the way out to some people will describe it as a lamp post or base station where you've got antenna, come one hop in where you might have a central office, which is gonna have enough space where you could put a small rack.

(Chris Wright at 00:19:13) So, you know, we have edge computing devices that sit on oil rigs because the rig itself is littered with sensors which don't have enough compute capacity. They're filtering data into an edge device. So there's many different ways you can view the edge. And one of the things that we look at from a Red Hat perspective is, okay, there's gonna be some consistency. One, you're gonna have to run applications.

(Chris Wright at 00:19:41) Two, a likely application delivery vehicle is a container. Three, there's probably some data processing that's critical to the workload. Four, data processing is probably going to benefit from hardware acceleration. When you're sitting in an environment where you're power and space constrained, you need to super optimize your hardware platform for the application workload that's running on top of it. And maybe five, lots of those devices across the edge become fundamentally distributed, so distributed management is critical.

(Chris Wright at 00:20:18) So it kind of plays out in all these different dimensions. You could call any one of those earlier examples edge. They don't all have the same footprint. Some could be a full rack of gear. Some maybe just a ruggedized single server device that's sitting outside, even exposed to the public.

(Chris Wright at 00:20:36) So security concerns look different. So pretty broad spectrum.

(Joel Beasley at 00:20:41) That has to be the single best explanation I've ever heard of edge, and it explains why I was so confused.

(Chris Wright at 00:20:48) Hey, that's good.

(Joel Beasley at 00:20:50) Because I get to talk to people from all different industries and all different markets, and I had an idea. Like, I had, you know, my idea, and it was close to that. But when you can articulate something with a visual like an onion, that was brilliant. That, like, the onion one, come on, man.

(Joel Beasley at 00:21:07) That's amazing.

(Chris Wright at 00:21:08) Alright, I'll take that.

(Joel Beasley at 00:21:10) Yeah, that is real. We're gonna, like, clip that and post that on, like, LinkedIn and everything. That is such a good visual because you instantly understand how a collection of people nearby can all have an edge, and it all be similar but different.

(Chris Wright at 00:21:23) Yeah, exactly.

(Joel Beasley at 00:21:25) So data is now gonna be moving, I guess, to the edge. Right? And there's a lot of good use cases when we're talking about, like, processing data more locally, all the way down to, like, I've heard Tesla and some other companies are, like, manufacturing their own silicon to be, you know, more specific to the task that they're doing. Right? That, I guess you could say that's kind of like the edge.

(Joel Beasley at 00:21:49) Like, hardware specific stuff for the edge. But, you know, I watch those graphs, you know, of the data exponentially increasing.

(Chris Wright at 00:21:59) Right.

(Joel Beasley at 00:21:59) And then I look at our processing power, and I'm a geek as far as, like, always being interested in when are we gonna have an artificial intelligence that can socially replicate humans? For some reason, I'm interested in that.

(Chris Wright at 00:22:13) Mm-hmm.

(Joel Beasley at 00:22:13) Doesn't need to be powerful, just like social replication.

(Chris Wright at 00:22:17) Yeah.

(Joel Beasley at 00:22:17) And I started to look into that and look at how you would store memories and all of that and data. And I realized that we're gonna definitely be limited by our computing power. Like, we are—I think we're limited today by our computing power. Like, if I wanna do some really cool task, I ran some of the numbers. It would cost me ridiculous amounts of money, and it would take a long time to process some of the things I was previously, like, two years ago that I was looking at processing.

(Joel Beasley at 00:22:48) And I was like, whoa. You know, we're so advanced. I can run huge queries really fast—or what I call huge queries in my context of my previous experience. Right?

(Chris Wright at 00:22:57) Right.

(Joel Beasley at 00:22:58) And because, like, the new kids coming into the game, they're just like, we need more. We need more. Yeah. But, yeah, looking back, I'm like, we're definitely limited by our ability to process data.

(Chris Wright at 00:23:09) I completely agree. Which, I'll sidetrack for one second. My background, I spent a lot of time doing development in the operating system. I'm a kernel developer in a big chunk of what I've contributed to open source communities. Couldn't be a cooler place to be sitting as a kernel developer when you're like, there's all this new hardware, because what does an operating system do?

(Chris Wright at 00:23:39) Well, one of its key tasks is enable hardware so that user space applications can get access to that hardware. And the cloud spent the first half of the last decade propagating a total lie, or at least a myth. And that lie is that all compute cycles are created equally. And so we got that from, well, they're small, medium, large compute, and that's about all you need, and it's good for any application. If you look at a cloud today, a public cloud, and you look at the number of machine types or instance types, there are well over a hundred, a hundred and twenty-five different variants.

(Chris Wright at 00:24:26) And if you look carefully at what those variants are, it's varying ways of using hardware to accelerate certain types of application workloads. It's big memory. It's high I/O for data. It's high I/O for networking. It's, you know, more powerful compute cycles, more cores.

(Chris Wright at 00:24:46) There's GPUs, TPUs, tensor cores. There's all of these different ways that you can accelerate workloads. And so I think when you look at where compute's going, number one, data is a critical driver. So how do you move it effectively? How do you process it effectively?

(Chris Wright at 00:25:09) And the outcome looks like a heterogeneous computing environment where computers are scheduling tasks that are going to be better executed on certain types of silicon. And so there's specialization because we're kind of reaching some practical limitations of traditional classical digital computers in terms of what they call it, Moore's Law or Dennard scaling. How many transistors can you fit on a CPU without it melting down? We're starting to hit some limits of physics. And so you haven't seen CPU speeds increase much in the last few years.

(Chris Wright at 00:25:47) You see core counts go up, but an individual thread is just running roughly at the same speed. And then you're seeing systems-level views of improving performance, connections to memory and I/O, ways to look at memory like it's storage, ways to improve the overall throughput of a system without necessarily changing the clock speed of a CPU. So from an OS point of view, it's pretty exciting. From a data processing point of view, I think it's fundamental, and it's just strictly required to have this heterogeneous computing environment. And we've already seen GPUs take on a lot of the tasks of deep learning, machine learning type of workloads.

(Chris Wright at 00:26:33) There's also a question of what you were talking about—scale. Massive amounts of data and training data typically happening in a large, more centralized location like a cloud or a data center. The model that's associated with that could typically be put to an edge device used for inference, which has a different set of computing requirements. But we also need to connect these two, so you can't just train the algorithm one time, deploy the model. You're gonna have some feedback. Distributed learning, I think, is a really important question that we need to be able to answer for ourselves. How can we give feedback to models without shipping all of the data all of the time? Because it's impractical. So I think there's a ton of really interesting challenges that end up being grounded in some hardware questions. Like, how is hardware evolving?

(Chris Wright at 00:27:25) And heterogeneous compute, I think, is a logical conclusion and one that we're already seeing today, not just in edge, but in general across the compute landscape.

(Joel Beasley at 00:27:37) So we reached limit in physics of clock speeds and what we could do, and then we start into some specialization areas. But is there anyone working on the, like, first principles? Like, let's go back and see if there's a different way to do this from the ground up. Is there anyone in that market?

(Chris Wright at 00:27:56) There definitely is. There's a whole world, and this is maybe not suited for every computing problem, but there's a whole world of research that's happening in the quantum computing space. So quantum computing changes the notion of a binary bit to a quantum bit, or a qubit. And that qubit has the association of a probabilistic likelihood that you're in a zero or one state, which is a very different expression from you're either a zero or one. And the applications for quantum computing get really interesting in certain areas like life sciences, where we need—how do you model the universe in a meaningful way that doesn't take, you know, hundreds or thousands of years of classical compute cycles to understand the ways two molecules interact? Or, you know, put it into a context like today.

(Chris Wright at 00:29:14) How do you do matching of genetics and potential drugs for humans? You know, there's—those are really complicated, sort of maybe math-centric modeling problems that are not as—not gonna be as efficient on the classical digital computer. Quantum computers are working in a direction that are trying to help us answer those questions that might have taken a hundred years in, you know, a matter of hours so we could actually get some meaningful insights that are fundamentally new and different from where we are today.

(Joel Beasley at 00:29:47) Do you know Bob Sutor?

(Chris Wright at 00:29:49) Yeah. Yeah. Red Hat was recently acquired by IBM. Bob Sutor is an IBM employee who's also a lead person, always out speaking about their quantum computing efforts. So actually, I listened to a couple of podcasts he's done, which have been really informative.

(Joel Beasley at 00:30:09) Yeah. I talked with him a few months ago, and he brought up the example of, like, how when we model a drug, like, the digital representation is this abstraction that's not close to the level of detail of the actual molecule or whatever the organic structure he used. And that, you know, these qubits happen to have some specialty areas. And I actually see our—I don't know. I don't have, like, a data backup.

(Joel Beasley at 00:30:41) I just have a gut feeling on this. I think that we're gonna see a bigger biological market emerge. I think we're gonna get smarter with our drugs and our health and human enhancements. Right? I've been watching, like, the different artificial limbs you can get if you lose an arm or something.

(Joel Beasley at 00:31:01) Those things have just, like—soon it's gonna be better just to have a bionic arm. Right? Like, people at the wood factory start drinking on the job. They're like, "Maybe."

(Chris Wright at 00:31:15) We'll see. No, I—

(Joel Beasley at 00:31:17) Probably not. But it's—

(Chris Wright at 00:31:18) Probably not. It's true that there's a lot of amazing advances that are happening in the medical world. And one of the things that I think is interesting is some of those are just pure technology. So this example of how do you apply a quantum computer to modeling of how could you eliminate a whole set of time associated to testing drugs with humans by doing really sophisticated modeling that aren't based on these kind of compression techniques, you know, an analog signal to a digital signal? There's compression and it can be lossy. So how do you just really model the core, say, elemental interactions? I think that's super exciting and holds a lot of promise. But also in the medical fields, actually, even with artificial limbs, where are communities evolving that help develop the ideas and concepts and techniques behind the creation of—you know, we've seen interesting open source efforts around the manufacturing of an artificial limb driven by people who are really understanding how it impacts their lives because it's a limb that they're going to use, and they're directly involved in the ability to create some of the ideas and concepts behind it. Or there's an open source project that helps people with, I believe, with diabetes manage their blood sugar levels and the ways that you manage your dosing so that you don't have to worry about setting a timer and, you know, checking your blood sugar levels—and maybe you missed the timer—and it could be a really life-threatening situation.

(Chris Wright at 00:33:18) How can you create an artificial organ essentially that's really helping keep your body highly functional, driven from open source projects from people who are passionate, typically because they're directly affected, about the topic? That's a great example of the power of human collaboration and creating ideas that are bigger than any one person, which is really core to what open source is all about. I mean, I just love that. I love that we're sharing ideas and growing our collective knowledge and our capabilities through the sharing of ideas.

(Joel Beasley at 00:33:55) So you guys have a couple platforms.

(Chris Wright at 00:33:56) You have—

(Joel Beasley at 00:33:57) You have a couple platforms. You believe that they should be open. You guys have open source is at your core. I wanna talk a little bit about some of the platforms you guys have, but I also wanna talk about this concept of, like, open source becoming mainstream. And, like, how do you keep the culture? How do you keep it cool when it goes mainstream? It's like music. Right?

(Chris Wright at 00:34:19) That's right. From a platform point of view, what we're focused on is an area that we would call open hybrid cloud. And the hybrid cloud is taking the cloud concept, which is how do you create an environment where operationally, it's a managed service? So you have this experience of total ease of operations. And then from a developer point of view, it is highly efficient because it's got everything you need in terms of your core dependencies for your application available as APIs and just services that you consume, maybe even in many cases externally from your application or your core, quote-unquote, business logic. So we're building a cloud that enables developers and operations teams to span both your internal data center as well as out to a public cloud, and the platform that we use is Linux and Kubernetes-based. And this is also saying that we stretch out to what I would call the edge, independent of exactly which part of the edge you're talking about.

(Chris Wright at 00:35:29) And so that is all being built. The technology components behind that are all being built in open source communities. And if you look at any large-scale public cloud, the cloud itself is built off of core open source technologies. Many cloud providers have something like a Kubernetes service. So even if you get to a specific project like Kubernetes, they're available in clouds.

(Chris Wright at 00:35:56) And so the idea that open source has gone into a very mainstream part of what is today's business is a hundred percent real. I mean, imagine your day and the search engine queries that you do, the social media interactions that you have, the online ordering that you might be doing. Every single one of those experiences are supported in large part by open source software. Just like, not all stuff that Red Hat products, I'm talking much broader, the whole open source ecosystem. So what I think that's interesting in there is, okay, open source is now this mainstream way to do development. You can't have the talent pool in a single company today the way you could have a talent pool spread literally around the globe in an open source community.

(Chris Wright at 00:36:55) Creates a very different potential for what happens inside a development project. The challenge is, with all of this success built around open source technologies and projects, how do we keep the core of the culture of who we are as open source developers alive? How do you keep that kind of grassroots way of engaging with one another? How do you ensure that all of the financial motivations that are associated with business success don't sort of taint the true developer experience where we're trying to build the best technologies and solve problems the most effective and creative ways? I think that's a real challenge, and I think the risk is, you know, we've gone mainstream.

(Chris Wright at 00:37:50) We get distorted. You lose some of your soul, and as a result, lose some of the magic which makes open source projects so incredibly effective and valuable for—it's not just the industry. It's the planet at large for people and how we do compute. So I'm pretty interested in that topic. I've got some ideas.

(Chris Wright at 00:38:13) I think it's something that we'll have to percolate on for a little while as we try to understand what it means to create these communities and build technology together in this modern, we've gone mainstream world.

(Joel Beasley at 00:38:29) I saw, like, the GitHub—I don't know if it released, but I know that they were talking about it—where you could do, like, some payments to help maintain open source projects. What are your thoughts on that type of stuff?

(Chris Wright at 00:38:39) Well, I think it's—I mean, anything we can do to inspire more open source developers is good, in my mind. One of the successes of open source has been going mainstream means a lot of developers are professionally employed to write code in open source communities, so that's a good thing. There's pluses and minuses of how you direct that funding, and there's foundations that have looked at fellowships. There's different ways to do the, spread the commercial success across community developers where you're not relying on the creation of a product. And so I think being creative is important, trying all of these ideas.

(Chris Wright at 00:39:29) I'm not sure what the best model is, you know, the GitHub model, other models. Many companies like Red Hat are employing developers to be directly involved in the open source communities, and that's a critical part of the community's success. Yeah, I don't know. There's a lot of ways we could tackle that one.

(Joel Beasley at 00:39:54) What do you think—I guess, like, what would you say, because you have a lot of experience in open source and community and things like that—but on an open source project, what do you think some of the most important things are for the success of that project, about building community around it?

(Chris Wright at 00:40:11) I think one is that clarity of mission. People are rallying around a common cause. The collaboration aspect of an open source community is we're collaborating together to solve a problem that we think is important. And I think that's typically a little easier because software tends to have features and functionality that are focused on problem domain. But even with that, we've seen some projects take on such a broad scope that it gets hard to know where to prioritize, and then the project itself can just stagnate because it's internally thrashing with a lack of clarity of what are the most important things to work on.

(Chris Wright at 00:40:50) There's another piece which I think is super important. If you think about who is involved in an open source community project—well, it's people. It's human beings. And what happens in an open source project is a collection of trust relationships are created. And that's probably the most fundamental thing that builds the connectivity or health inside a community.

(Chris Wright at 00:41:26) Because if you don't have trust across the different members of your community, you'll just be dysfunctional in the end. And the reason I think that's so important is that's a very human element, and we tend to think in terms of code and features and number of commits and, you know, all these metrics that don't necessarily incorporate how are you building trust relationships between the humans that are involved in your community. There's another piece that I think is really critical, and that is diversity. And I would put diversity into two separate categories. One is diversity of the people involved.

(Chris Wright at 00:42:08) The more diverse you have—ideas and ways of thinking or contexts that you're bringing to a project—I think the healthier, more vibrant, and more resilient the project will be. Similarly, for certain types of projects that are very platform-oriented, onboarding a diverse set of workloads onto the platform, being capable of supporting a wide variety of use cases, creates a type of value for users. It creates—I would say it requires a kind of architectural discipline and even cleanliness because you're going to need to adapt the software to take on new use cases. And it means the focus of long-term maintainability becomes a critical part of the community's development process. And a great example of that would be Linux, which, as an operating system, has had to adapt to running on the smallest devices to the world's largest supercomputers, all sorts of different application workloads over the course of a couple of decades.

(Chris Wright at 00:43:25) And it's that diversity in workloads and, in this case, compute environments that it's supporting, or hardware environments that it's supporting, that's created something that's got longevity. And the internal code refactoring needed to support that has created a sustainable architecture. So different ways to look at diversity, but, you know, I think those are some key aspects. Some clarity of what you're trying to do, building a collaborative environment which requires trust, having a diverse set of users and developers involved in the project, bringing a disparate set of—a diverse set of use cases and perspectives to the development process—I think really build sustainable, maintainable, resilient open source projects.

(Joel Beasley at 00:44:16) I can tell you thought about this a lot. This is awesome. This is good. It's like you have a lifetime of experience in this area or something.

(Chris Wright at 00:44:26) It's not every day you get to talk about the things that you love, so it's pretty fun.

(Joel Beasley at 00:44:30) Dude, that's why we come here, right?

(Chris Wright at 00:44:33) Yeah. I love that.

(Joel Beasley at 00:44:34) That's what I love about it. It's like, let's just hang out, talk about what you guys want to talk about, and just have fun because, you know, that's the most important thing. If we can talk about what people want to talk about, then that's an entirely different conversation. That's why I like to leave it open and flexible when we hang out, and really is one of the reasons why I kept going with the show because I would look at some other clips and they would be, like, very rigid. And I was like, I want to get to know this human. I want to know, like, what they think about in long form versus, like, a, you know, thirty-second clip.

(Chris Wright at 00:45:09) Well, it's true. Anybody can say the words. How do you contextualize it so that it's clear you're relating it to your own experience, which helps others relate to the experience? I think it's important.

(Joel Beasley at 00:45:24) Yeah. 100%. Shared experience bonds us. Now, I was pretty excited about Summit. I heard—is Summit now going to be virtual?

(Chris Wright at 00:45:35) Yes. So like many, we had to decide what to do about Red Hat Summit, which is our annual big conference, which is very much a face-to-face event. You know, we bring together thousands of our customers and a lot of our developers to give presentations about what we're doing, the state of the art, the places we're going. It's a super fun event. And given the context that we're living in in this world with COVID-19, we had to make the call to say this isn't going to work as a physical face-to-face event.

(Chris Wright at 00:46:16) We made the call to move it to a virtual event. It's at the end of April, and we'll certainly have to change course a little bit, but eight-hour-a-day multi-track format just isn't quite the same format that makes sense in a virtual environment. But we're being super creative. Our creative team is amazing. They've got a lot of awesome ideas of how we can get content out there in an interactive way so we don't lose the ability to engage with our audience, who's largely our customer base.

(Chris Wright at 00:46:50) So we'll be doing a virtual event at the end of the month and talking about all the things we would talk about at Red Hat Summit, which is our products, our portfolio, the things we're doing, the ways that our customers are using what we're doing to solve really amazing problems for themselves, and even where that elevates the entire industry segment that those customers sit in. So really exciting way to look at technology and how it impacts people, and it's just going to move from a physical event to a virtual event. And it's kind of an experiment on a certain level. And many other companies are having to go through similar experiences. And my personal hope is that it's successful in a way that we can think about what events mean to all of us involved.

(Chris Wright at 00:47:44) I travel a lot, and there's the carbon footprint associated with that. There's the effectiveness of sharing information face-to-face versus virtually that I think we can explore. And so I think it's an opportunity to really see what happens. Not that our event's the only one that's going to fundamentally change the world, but we're just in this process of collectively learning. And Red Hat Summit is one of those events that will be a virtual event, and I look forward to how we learn from that and, well, really how we can grow from the traditional event model, which has been around for quite a while.

(Chris Wright at 00:48:28) So maybe it's time for some innovation in that space.

(Joel Beasley at 00:48:31) You guys will do well. I mean, you have a really good creative team, really smart people all around. I saw the—they gave me this link before our last call, and I, like, freaked out when I saw it. So it was a link for the Edge website. Was that just, like, a single blog post that it was?

(Chris Wright at 00:48:50) Yeah. Yeah. It almost doesn't work to call it that because when you scroll through, it's a little different, but it's a really compelling space to discuss edge computing. And, you know, it's just an example of the amazing work that our creative team does.

(Joel Beasley at 00:49:06) We should send them a note after this, and we'll clip the onion explanation and see if they can turn that into a visualization.

(Chris Wright at 00:49:14) I'm sure they would be able to.

(Joel Beasley at 00:49:16) Because, like, that—and I'm going to post links in the show notes so everyone that's listening can go—but, you know, people that listen to the show, they know that, like, I'm a huge fan of design and animation, like, really good quality brand and messaging. Like, I'm just a big fan of it. And when I saw this page, I was like, this is unbelievable. It had, like, this subtle yet brilliant animation. And, I mean, there's good content too. But it's about Edge. It's about edge computing. It had some history. They, like, animated up some, like, older computers and told—and there was storytelling in there and everything. And it was just like—I bookmarked it as a favorite because it's something that I'm like, in the future, I'm going to need to tell someone I want to build a page like this. And this is the example I'm going to give them.

(Chris Wright at 00:50:08) That's fantastic. We'll have to make sure they get that feedback because it is—it's fun to work with creative people. I spend most of my time in more of an engineering context, so the creativity is focused on code and features. But I love working with the creative team and understanding how to harness that creativity either visually, so images just pop out and you understand what somebody's trying to portray just with an image, or even in that storytelling arc.

(Chris Wright at 00:50:41) Like, how do you bring somebody along the journey where it's not just the bits and bytes, but it's the whole infusion of an integrated experience, or making a journey out of it. And it's—I've learned a lot from the creative team.

(Joel Beasley at 00:50:57) Oh, yeah. I just noticed that, like, great people tend to collect together, right? So I got to talk to you, and I was like, oh, this guy's brilliant. And then I got to talk to Will and some other people on your PR team. They gave us, like, amazing storylines and stuff and notes, and I was like, oh, this is so good. And, like, each—you know, I've interacted with your brand as much as most people do with the fact that you guys basically power the Internet, right? But the services that you guys have and the platforms and how you're pushing forward everything is just—I just have a lot of respect for you guys. Like, did you think—fifth, or, like, twenty years ago when you knew what Linux was—did you think that you'd be the CTO there?

(Chris Wright at 00:51:39) I can honestly say no. And I could probably expand it out a little bit and say, twenty years ago, working in open source software, hacking away at Linux was a passion. It was a way to learn. For me, luckily, even that long ago, it was also a job, so I got to spend some time professionally working on open source software.

(Chris Wright at 00:52:11) I never imagined twenty years later the impact that open source would have just in general across the industry. You know, sounds crazy to say it, but to humanity, because it felt like a fringe radical movement. And we spent a lot of time either just working on our own projects or maybe, if we're lucky, trying to convince somebody that we weren't crazy and they should run this software. And today, it's, you know, the mainstream discussion we had earlier. Today, it's very much a part of the mainstream. And then, personally, I didn't anticipate where I would be in twenty years.

(Chris Wright at 00:52:54) You know, that kind of interview question of what do you want in five years—I'd barely be able to answer that, let alone where do you see yourself in twenty years. And so my twenty-year-ago self couldn't imagine being in a position where I'm CTO of Red Hat, and Red Hat is a $4 billion company now acquired by IBM in the world's largest software acquisition in history. I mean, just—it's amazing that the trajectory, personally, that I've been on, but also more broadly where open source has come. It's just—it's staggering. It's awesome. It's phenomenal.

(Joel Beasley at 00:53:35) I get a lot of people that'll ask me, oh, like, you know, what's the secret? Just, like, hard work over a long period of time, right? But you have been on this journey and you have put in your twenty years. But, I guess, rather than asking what's your secret, right—what sort of patterns and behavior in yourself have you noticed benefited you?

(Chris Wright at 00:54:02) The first word that comes to mind is curiosity. So, being curious and learning and wanting to know more, understand more has been a really important, just, I don't know, characteristic, attribute, tool for me. In the beginning, that was open source. The curiosity was, with all the source, you can dig in and see how things work. Today, I might be asking different questions of the people around me. And maybe related to that curiosity, and this is a harder one, you know, for me and I think others as well, not being afraid to ask questions, even when you might be convinced that everybody else already knows the answer and you're the one person who's still clueless. You know, that kind of curiosity—I don't know if that's just humility. But those two things I think are really, really valuable. And then having the ability to communicate has been super important for me. And I'm an engineer.

(Chris Wright at 00:55:16) I'm a nerd. I love the details, but you also have to work with people and communicate your ideas to people and influence people to do things, and it's not necessarily binary or black and white or just the best idea that's the most influential. Sometimes ideas are hard to understand, and sometimes people's motivations just start going a different direction because they're thinking about solving things in a different way. And so how do you communicate and build some bridges just, I'd say, more at a human level? It's been a really important way that I've interacted and I find it an important part of the way—put it this way, when I talk to our engineers in our company about what does it mean to be a technical leader, it's not just being the smartest person in the room.

(Chris Wright at 00:56:12) Being capable and bright is super important. You bring a lot with just your intellect. But being able to work with people and influence people to move in a common direction, that's a critical part of technical leadership. And I'm not sure that we always emphasize that enough, even in open source communities where it's so focused on, you know, the best ideas win. You don't want to marginalize people and give people the sense that their contributions aren't valuable even if they're not merged into the code base.

(Chris Wright at 00:56:52) So how do you balance that discussion or the dialogue about a feature without making it a personal attack? "That's a dumb idea." To be blunt, that's kind of an asshole move, and the way I describe it is we don't need assholes. Right? We need really smart, engaging, thoughtful people who are also capable of being influential by asking some of the right questions and not being close-minded.

(Joel Beasley at 00:57:22) Yeah. I've thought a lot about that because we have a leadership company, like techno leadership for technologists company. And, you know, you're always looking for value propositions and how the market's moving. And one of the—it keeps popping up as an idea or just keeps recurring in the back of my head.

(Joel Beasley at 00:57:41) It's hard to measure—when people think of efficiency or teams, I found that their default, they go to tools. Like, "I'm going to use this type of tracker. I'm gonna use this type of analytics system. I'm gonna use this, this, or that." And then I found—I had this thought one day. I was like, if you did reject someone's merge, right? You're like, "Nope." And with no explanation. That human interaction—how do you measure how much productivity that just killed of that individual?

(Joel Beasley at 00:58:10) You know, like, how do you measure that? And how do you then connect that back to a business case? I haven't found a great way of doing it yet, but it's definitely something that I know, I feel, I see happening in teams. And then I was just working with our sales team trying to figure out exactly how to do that. But it's definitely important. How you treat people is directly connected to their productivity, to everything.

(Chris Wright at 00:58:36) Yeah. I love that question, and I could think of ways of sort of playing with that idea. I worked with an analyst group who recently did a study that they referred to as developer velocity and its impact on a company's revenue. And I think they're introducing a notion, maybe it's not their own, of DVI, Developer Velocity Index. And companies with a higher DVI or Developer Velocity Index generate three times more revenue than businesses like themselves with a lower—you know, in a similar area of the market, but with a lower Developer Velocity.

(Chris Wright at 00:59:17) So you could kind of ask yourself maybe from that same question you're asking, how diverse are your inputs across your developer pool? And if you see it's always the same few people's code being merged and everybody else is hearing no, is there a way you can measure velocity, not just pure code commits, but also the number of people who are committing? And you might even be able to see people who are consistently not being encouraged to continue to participate really slow down and fade away. And in some cases, that might be a hard decision where it's a person who's not being very productive in the project.

(Chris Wright at 01:00:02) In other cases, it could be how you're handling the communication of, you know, did you need to say no and then go write it yourself and merge it? Or could you say no and give some constructive feedback and give the person some direct guidance of how to improve it and then merge their code? The end result is the right idea got merged. The difference is the process that you used to get there and how empowering that was for the person who either heard no with some feedback or just heard no, rejected. And then a similar looking idea just got merged by this other person, and that's totally demoralizing or demotivating.

(Joel Beasley at 01:00:44) Do you guys do that? Do you guys measure your velocity? Do you have any favorite tools for that?

(Chris Wright at 01:00:50) Well, the short answer is no, in terms of what I was describing. So the way we look externally at metrics and communities would be pretty basic and that's commits per time period. Internally, we do keep metrics, but similar, it tends to be more status around commits. And we're very careful not to use commits as an incentive because you get terrible outcomes.

(Joel Beasley at 01:01:19) You're telling me, man. Everything's a commit. Yeah.

(Chris Wright at 01:01:22) Yeah. But I'm gonna change, oh, you know, change a word. I could submit however many letters in that word, subsequent pull requests to slowly rewrite that word. That would be a terrible way to build your code base. But, you know, those are just more informative. You get a sense of overall productivity or activity is a better way to put it, just by looking at the commit rate. I don't know how much detailed analysis we do down to per developer, per project, or ways that—and I think you could do some natural language processing, looking at the language that developers use when they communicate with each other and seeing the sentiment analysis of developers and associating that with the quality and velocity of code that's going into the project. I'm sure it'd be a fascinating way to research.

(Joel Beasley at 01:02:17) You could probably build some custom dictionaries too off of repos, like, extract comments from repos that have more success to figure out what the terminology is. Right? Because the standard dictionary is not gonna do perfect.

(Chris Wright at 01:02:30) Yeah.

(Joel Beasley at 01:02:30) With all of our lead speak.

(Chris Wright at 01:02:32) LGTM is probably not a common part of any normal dictionary. And that sort of a terse response, out of context, might be thought of as rude. In context, that's two thumbs up. Looks good to me. That's as good as being committed. So context matters as well.

(Joel Beasley at 01:02:55) Is there anything else that we didn't cover that we want to get out there today?

(Chris Wright at 01:03:01) I talked about data before, and I talked about the importance of data, how it's kind of—data to me is a central actor in the enterprise. And I think the movement of data, which is the connectivity that all the pieces where things like 5G and Wi-Fi 6 and new networking technologies and edge bringing compute closer to data are really important. The compute environment around data processing, inclusive of heterogeneous compute focused on machine learning as a critical workload. There's also the storage of data, and data is—because it's the central actor—moving your mindset towards not "should I store it," but the answer is just yes. And because you learn sometimes later that it was valuable and you can't go back in time and store it. So, you know, we're seeing people be really interested in just collecting large volumes of data, and then the relationship between data in motion and data stored, and the analytics combined, I think all brought together in a single platform is sort of unique for where we are in the world today.

(Chris Wright at 01:04:26) And so, typically, in the past, you had more like a storage platform, an application platform, maybe a data lake and a Hadoop kind of data analytics platform. And these are separate. What I think is really interesting today with containers and Kubernetes becoming this really common building block for the industry is bringing these together on a common platform. It's really a powerful way for people to build their infrastructure and harness the value out of data. And we've been doing a lot of work to bring data more directly integrated into our platform, so it's container native, container accessible when you're in a containerized application environment.

(Chris Wright at 01:05:10) And I think that's important. And the one piece that I wanted to touch on was that storage component. Right? You know, we talk a lot about movement, a lot about the application or the analytics and processing of data, but also storing data and just collecting what you can because you'll find value in it later.

(Joel Beasley at 01:05:32) I love it. Dude, we're doing it, man. We make a podcast, and you weren't in the tiny little closet. Good. You feel good?

(Chris Wright at 01:05:44) Yeah. I feel good. I've got nothing but a mellow day ahead of me. I can actually get some focus time to focus on a few projects that I'm working on. And I really enjoyed talking with you. It's fun to have just a discussion, communication. I learned something new. I got a great shout out I can bring back to the creative team. So it's been a nice time.

(Joel Beasley at 01:06:09) Yeah.

(Chris Wright at 01:06:09) The onion.

(Joel Beasley at 01:06:10) Dude, the onion, I'm sure you've said that before because it was so articulate. But hearing it for the first time, it was like mic drop. I was like, I can already see the animation on the creative page, and it's just like, in thirty seconds, you completely took something that was fuzzy. And, I mean, yes, the people that you get when you go speak in front of crowds, they get to hear it. But we—I'll do everything I can to amplify that out as much as possible because it's just brilliant.

(Chris Wright at 01:06:43) Good. Thank you. Appreciate that. Yeah.

(Joel Beasley at 01:06:44) And are you outside of Boston? Are you in Downtown Boston?

(Chris Wright at 01:06:47) I'm seven miles due north.

(Joel Beasley at 01:06:52) Is that—that's not like Westchester, is it? Oh, I'm sorry. That Westchester is in New York. What town is that?

(Chris Wright at 01:07:00) Probably similar. New York to Westchester is where I am to Boston, but in Boston. Yeah. I'm in a town called Winchester. Maybe a slightly better, the better known maybe nearby towns would be Lexington, Somerville, Arlington, and if I go a little closer to Boston and Cambridge. And right now, it's about a fifteen-minute drive to the office because everything's empty.

(Joel Beasley at 01:07:30) Everything's empty.

(Chris Wright at 01:07:31) In a morning commute, it's somewhere between forty-five minutes. That's pretty normal. Anything funky happens, it quickly is an hour, hour ten. And if something catastrophic has happened, it's an hour and a half and more, and all bets are off. But I'm just working from home. So, nice. Who knows what it's like out there?

(Joel Beasley at 01:07:53) Well, towards the end of the year, I'm sure everything will be back to normal, and I've got a lot of people. I'm in Boston a couple times every year because, you know, there's so many—Vistaprints out there. They're a really cool company. I don't know if you know the guys over there. But they actually have a parent company. There's a Vistaprint for every country. Their parent company is in Boston. But they have a lot of health care companies, a lot of AI type startups. But I really like it. And so next time I'm out in that area, I'll send you a note, and maybe we can meet up, say hello.

(Chris Wright at 01:08:22) Please do.

(Joel Beasley at 01:08:22) Meet up, say hello.

(Chris Wright at 01:08:24) That'd be great. I'd love it. We could be past this social distancing stuff and have a legitimate beer and hang out or whatever.

(Joel Beasley at 01:08:33) Yes. Okay. Yes.

(Chris Wright at 01:08:35) Not an air fist bump, but a real fist bump or—

(Joel Beasley at 01:08:38) Boom.

(Chris Wright at 01:08:40) Shake hands. Boom.

(Joel Beasley at 01:08:41) Do the fist bump, and we'll get a screen grab of that.

(Chris Wright at 01:08:44) Alright.

(Joel Beasley at 01:08:45) Dude. Boom. Do it up to the camera. Boom. Alright. Digital fist bump. Social distancing fist bump. Yeah. Alright. Well, you have a fantastic day. Enjoy the rest of your afternoon, Chris.

(Chris Wright at 01:08:56) You too. It was great hanging out with you.

(Joel Beasley at 01:08:57) Great to talk to you too, bud.

(Chris Wright at 01:08:59) See you later, everyone. Cheers.