 From Austin, Texas, it's theCUBE, covering KubeCon and CloudNativeCon 2017. Brought to you by Red Hat, the Linux Foundation, and the Kube's ecosystem partners. Well everyone, welcome back to our live exclusive coverage of the CloudNative Conference and KubeCon. Put on by the Linux Foundation, I'm John Furrier, the co-founder of SiliconANGLE, meeting with my co-host Stu Miniman. We're here breaking down all the action in the tsunami of open source developers, renaissance and software demolition. You know I've been talking about our next guest. We're excited to have Kelsey Hightower, who's the co-chair of the committee here for the program, for this awesome conference that's exploding. I was also a staff engineer at Google, known in the industry as very active participant. Kelsey, great to have you on theCUBE. Welcome to theCUBE. I'm happy to be here. I feel like I've made it now. Well, not really. You make it every day on Twitter. We follow you. I mean, you've been an active voice and it's been fun to watch this community. We've been president and creation at theCUBE and KubeCon and we've been watching the evolution. It's like Jello that forms in the refrigerator. Couple of years ago, you saw it come together. Containers, microservices. The drive or the tailwind for now Kubernetes, this orchestration opportunity has changed the game. What is the bottom line? How has Kubernetes, because everything was all about containerization, that was going to change the world, but it kind of did, but it's evolving. What's so important about Kubernetes? I think Kubernetes is really an actual thing you can use that takes all the ideas we've been working on for the last 20 years and just gives us a new starting point. So that's about changing the game but actually making the game available to everybody, right? So we always talk about containers as this revolution, but if you think about containers, it's more like let's take VMs and make them faster to use, shrink them down, and then the configuration management world of deploying those things. Kubernetes wraps all that hard work into a single thing and if you start there, it feels like you just leapfrog where you were. Kelsey, I want to ask on that. So much we get excited about the cool little tool, but it's about the patterns, it's about what I can build with it. When I look at this community, that boring infrastructure stuff is important, but it's about building the applications and what I can do with it that we seem to really see coming out of this event. Yeah, Kubernetes represents the experience of like the RedHash, the CoroS, the Googles of the world into a thing you use. So when I talk about Kubernetes, it's like when we solve a new problem, just like in Linux, it rolls back into the platform, but it covers this big problem set that almost anyone writing software has, and I think this is why the traction of Kubernetes is so big, so fast. I mean, so many successes, I just love watching the tech evolution, Uber, Lyft, Netflix, building scale software on open source, and there's a lot of success stories. So two things jumped out of me on the keynote, pluggable architectures and service meshes, two dynamics that are pretty instrumental and part of it. It sounds intoxicating and it's cool, but then if I'm just a practitioner out there, I'm like, well, all this other stuff I'm used to is hard. What about security and storage? So there's a lot of other things that are important to customers, the blocking and tackling, storage, networking, whatever, and then new things are coming to the table. So you got new vocabulary, new concepts, combined with the existing, pre-existing, old guard concepts like storage, networking. How does that, how do you connect that? So for the person who's running IT or the CIO or the person that are doing technical architecture or the large big IT department or company, they got to grok this. How do they figure it out? How do you dissect it? So the problem is they didn't change. Your app takes input, does something, produces output. About 30 years into making now, that doesn't change. Kubernetes doesn't change that. Containers doesn't change that. So I think all this stuff, if you look at what you've been building over your whole career, all the bash scripts, all the tools that you brought in, their whole goal was let you to focus on building those applications. We've taken all of those things, realized what the patterns were. So if you look at Kubernetes, we just lay an OS on top of all the storage, the compute and the networking, and just says, hey, here's a new set of primitives, and we're going to make it easy to consume those. And then the next level on top of that, security is inherently baked in for the most part. So I used to work in finance. When you look at it and say, what's running? Most people can't answer that question, not easily or with a straight face. In Kubernetes, we have a declarative object that tells you these are the things running. They were started at this time by this person. That's what you get by default, even though we don't talk about it as a security problem, it totally is. Hold on, so when declarative continues innovation and integration, why is that important? Because does that speak to the distributed nature of it? I mean, why is the declarative piece so important? So distributed, I think a lot of times people have been dealing with distributed systems for a long time without understanding how to actually deal with the patterns. So we've just been doing it badly. Once you add more than one machine to your stack, you now have a distributed system. So we've been able to deal with this with the meat cloud, throw a bunch of people at it, right? And everyone just deals with their subsection of the servers. Now we're just laying a thing on top that lets you treat it like one single machine. That's how we now start to think about this new problem. So once you start to have that kind of, those premises at your disposal, just change the way you tackle this particular problem. So I'm not sure that this is like a whole new mindset required, it's just that now you can just rebase, right? Like with the mobile phone, you're not necessarily writing apps at the very low level anymore. You're writing way up here with a bunch of new abstractions. Yeah, so you brought up security. It's one of the hot button topics. There's the low level like, wait, do I put it in a VM or do I do it at the container level? What do you see kind of the status security in this space? What do we still need to do? There's two levels of this, right? There's the security in my app. So no matter how great Kubernetes gets, no matter how great we do at the very low level of like this container shouldn't do these things, you still have this layer where your app will set requests from your users. And more than likely, that's where your problems are going to be. No one's doing brute force anymore. I'm just going to come in on the port that your security team opened, and I'm going to abuse your app because there's probably some hidden behavior that you're unaware of. So that level of security, we hope that that industry starts to have more people focused at that real value layer than the stuff down here. So Kubernetes may take care of this down here. So we talked about the declarative piece. I know that this is what's running on these machines and I can be sure of it. You can actually assert things and that's part of security. Is it working the way you intended it to work? So decouple security you're saying. Do it, keep it at a declarative level of infrastructure, let the app guys fend for themselves or is that? It's more like, let's make it easy to do the right thing. Kubernetes doesn't solve all the problems, but the problems it does solve, we make security just be a built-in primitive. That's a good argument. It should solve its own problems. Don't try to do too much. But the patterns now, we start talking about security. If you think about Istio, that goes a little bit higher up the security stack. It also takes a declarative approach. So when you say only these apps can talk to each other, you can declare that and let the system do the enforcement rather than people. Okay, I got to give you kind of the question on demographic shift in the developer community here. Obviously the growth is big. The numbers are here, better than all the other events combined. How do you break down the, you had to draw a line in the sand, kind of infrastructure developers, from configuration management provisioning, all that stuff to kind of pure app developers who's like, hey, I'm DevOps, I'm just, I just want serverless, I want a full pool of resources, all that stuff's taken care of. How would you kind of 60, 40, 30, 70, how would you, because there's a lot of new people in here. What's the numbers in your mind? Just guess. In my mind, I would probably say this movement has about 70% of people who identify themselves as like, I'm a developer. I really want a different set of premises so I can move on. If we look at the last maybe five to 10 years where you've been brought into DevOps, you now have been exposed to infrastructure. And if you're going to be exposed to infrastructure, you want this kind of infrastructure and not what you had before. And I think the ops people took a little longer. They were like, I don't know, this just looks like something that doesn't solve my problems or it's only for startups. But now we're starting to see that it will work for almost any workload if you understand what Kubernetes is trying to do. It's hard to parse through the developer definition. Well, I mean, look at this. 4,000 people here this time, right? We started with like 300 people, maybe 500. And now we're at 4,000. So you start to see everyone say, all right, Kubernetes has a spot for me. Here's how I contribute and leverage the platform. Kelsey, what do you say to people that look at this environment? It's just too complex. There's layers and layers and I learn one piece and it's changing constantly. This opportunity, threat. Everything in life is too complex. Anything you don't understand is too complex, okay? But if I go to your company and say, how long will it take me to learn all of your systems? Years, probably. Not everyone knows everything. So I think all of these things by their very nature are complex. But if you think about what Kubernetes does, it at least takes all that complexity and gives it an API. You can now reason about it. So if you take the time to learn Kubernetes, all this stuff from how do I deploy my app to how we manage the hardware at least has a defined API for the first time. It isn't going to be random from corporation to corporation. We're now aggregating the complexity and giving it a name. In your mind, how would you define a high quality, pluggable architecture to leverage is the goodness of Kubernetes? What does that look like? How should someone kind of check their, check some their code, if you will, look at it and say, okay, that's a pluggable architecture. What does it look like? So Kubernetes, if you think about it, the whole thing is extensible. So when people talk about the complexity, it's because there are a lot of moving pieces. So it was designed to leverage its own API since day one. So if you want to add a new scheduler, the thing that does, where does this application run? Our current scheduler uses the Kubernetes API to do that. You can bring in your own, and Unova is a good example from two years ago, adding their own scheduler to Kubernetes. If you want like a TLS certificate from Let's Encrypt, there's a very obvious way that you would do that in Kubernetes. So our whole platform is API driven from the outset. And the benefit that is integration, right? Integration and extensibility, like one thing that has always plagued our industry is, you buy this big software package, you want to do something custom, and now you're screwed. Now what you have is, we expect it to be extended, and your technology partner of choice, we'll be able to extend it in a way that you can actually upgrade the thing. All right, so slightly a different area. Kubernetes now, there's what? 42 certified partners out there. Will anybody make money on it? You know, I come in saying, I don't think it's directly, I think it's more like the cloud platforms, the other platforms. What's your take on the whole business aspect of this? I think it's kind of like Linux, dude, how many people make money on Linux? You know, and I think even the people that do make money on Linux, it's the support, it's the service. And I think Kubernetes sets the stage for technology partners. You can't just sell me Kubernetes and walk away. You have to give me Kubernetes and envision how my business will extend on top of it. So I want to do machine learning. Kubernetes is a great platform for doing machine learning. The value is above that, with the machine learning and all that other stuff. Yeah, what's your take on that dynamic of all the contributors here? I know, joining Google, one of the reasons that if I remember right from your reading, you know, it's just their participation in open source. You look, Microsoft big on open source. You know, Adrian Cockroft is in the keynote this morning talking about AWS's participation. You know, what's your take? Honestly, if you're a big provider, the value is not proprietary software for you. I'm in a cloud provider. We sell CPU cycles. If you want to use Mesos to spend those CPU cycles, that's great. We have them to believe in Kubernetes. So we provide that based on our experience. So to me, Kubernetes is much more product of our experience than it is something that just Raleigh trying to compete in the market. So that's why I think people find it valuable. It solves problems that you have and that share amongst your peers. What's your advice to app developers? Because the impact seems to be also the value creation is going to be on solving problems in a way, new creative way. And again, we're predicting in theCUBE that we're going to see a swing back to the craftmanship of software development. I mean, Agile's great and kind of took that craftmanship but you de-risked it because you could make it run faster, but we're seeing a renaissance around craft, artisanship, not just UI. I've done about real value, subtle change, cultural impact. That's an value opportunity. Your thoughts. When you talk about craftsmanship, the thing that we always look at in craftsmanship, we always talk about how long it takes to do something. I made this by hand. This was aged for 50 years before we drank it. And I think what we're doing now in the enterprise is we don't have time to actually focus on the craft. I need it by Friday. And I also got to figure out the infrastructure first. So when you get things like Kubernetes and then you layer on platforms like serverless and these paths that sit on top, now you can actually focus on craftsmanship. Let me get this library right. Or if there's another company that has already figured it out and they've taken 10 years to get that library perfect, I get to actually use their handcrafted piece and my handcrafted piece and then we start to get to the actual vision. So I think the key missing element today is time. These platforms get your time back. Then you can actually invest in that craftsmanship. And all that heavy lifting around redundant stuff that you shouldn't have to do. I mean, no, I'm old. I remember I used to have to do our own graphics libraries. Now I can go to light. Now it's like the artisanship is coming back. 100% agree with you. But this is an opportunity that no one's yet monetized because it had never existed before at this level of speed, reliability. They're monetizing it. You're seeing the businesses monetize it. So remember, I don't necessarily think that the vendors, the traditional IT vendors will be the one that monetizes. It's going to be the Netflixes of the world, the people that have an idea and they go to market and then within two years, they have this large control of the market. Because now they look at it and say, start with Kubernetes, grab Prometheus, grab these pieces that have been handcrafted by a large community that cares. I'm just going to focus on my business piece. That's just cashing in. The value is shifting. The value is shifting. Kelsey, you mentioned time. First of all, I want to say thank you for giving us some time in this community. I've seen so many examples. People are like, Kelsey Hightower gave me a call and talked to me for 10, 15 minutes. I'm nobody. Podcasts, writing, everything else. How do you keep on about it? How do you look at seeing this community continue to grow? Honestly, you got to be, I'm a people person. And people are like, no, no, you work at a vendor. You're super biased. I actually am a people person. But you work at a vendor? Yeah, exactly. So for me, the people are first because these people helped me get to where I am today and I'm super appreciative of it. So when I get a chance, someone DMs me on Twitter and say, hey, Kelsey, I'm trying to reinvent my career. If I'm busy, I say, call me. And I pick up the phone and say, hey, how are you doing? Here's what worked for me. I'll listen for a little while and say, hey, here's my professional opinion. And I don't actually mind when other people do well. And I think a lot of times you want to shine by ourselves so much that we don't want to give away the secret sauce too early because then I might not be able to shine. I actually find it very enjoyable. If I helped you with your talk and you go and you rock the stage and you go back to work and you get promoted and then you tell me, hey, I really appreciate that. I found the ability to say, you know what? You win, I win. You know, pay it forward in community is critical. That is a great example. More people should do it, congratulations. Paying it forward is all about selflessness. But it feels good when you do it. People don't understand. It feels good when you're around other people that also feel good. You're so selfish with your selflessness. There you go. All right, final question for you. By the way, everyone should be like that because that's what communities do. Good thriving, robust communities help each other and don't mind me a little bit cocky, but that's swagger, I like that. But helping people is key. You have some good swagger. We appreciate your work on Twitter. Final question, your talk. What are you going to be talking about? What's the keynote like? Give a preview. So the preview is that I was going through the release notes of Kubernetes. And it's actually born. 1.9, if you look at what we're shipping, it's all around stability. It's all about delivering the promises that we made years ago. They're finally becoming V1 now. That's about it. There's nothing that I'm going to change in my cluster because of 1.9. And that's the major feature. We've been talking about getting infrastructure to become boring. And when I can look at a new release of Kubernetes and that freak out that I go to change a bunch of stuff, we've finally done it. We've done the part that we were designed to do. So what I want to do is say, hey, if Kubernetes is boring, where does the excitement live? And what does it look like? So I'm going to do a lot of live demos of here's what it looks like when you're doing it correctly from my point of view, based on experience. Boring is calm, boring is reliable. The action is on top. There you go. All right. Kelsey Hightower, thanks so much for spending the time. Appreciate you coming on theCUBE and sharing your insights and commentary. You'd be a great CUBE analyst. We'd love to have you on any time. I'm John Furrier, Stu Miniman here at CloudNativeCon and CUBECon live in Austin, Texas, back with more live coverage after this short break.