 Live from San Diego, California, it's theCUBE. Covering KubeCon and CloudNativeCon. Brought to you by Red Hat, the CloudNative Computing Foundation and its ecosystem partners. Welcome to theCUBE, here at KubeCon, CloudNativeCon 2019. Second day of three days wall-to-wall coverage. I am Stu Miniman. John Troyer is my co-host for the three days and we've had a great schedule but this one will be super dope, of course, because it is the one that has the right phrase to use. Kelsey Hightower, who is now a principal developer advocate at Google Cloud. Kelsey, thanks so much for joining us. Thanks for having me. All right, so, let's start. You did a keynote yesterday and I actually heard, you know, not only did it rain in San Diego, people were talking about allergies. They were grabbing their tissues, seemed to be tearing. You had stepped back for a little bit. When I first came into this show, we've been doing it for four years. It was, you know, kind of Kelsey Hightower and Kubernetes almost seemed to get top billing of the show. You specifically stepped back for a little bit and you're here this week. So, talk a little bit about that piece. Yes, I stepped back to do some serverless stuff, right? So, I worked on some cloud function stuff at Google, launching the Go support for cloud functions and really trying to understand the serverless space by being in it. And that means stepping back from Kubernetes quite a bit. So, the keynote, I wanted people to have emotion. So, no live demos, no slides, no speaker notes, and then just telling stories from the last six years of being a part of the Kubernetes community and making people feel something. And I think it resonated with folks and of course, people got a little teary-eyed. I gave people a cover so we just kept saying the allergies are starting to flare up in the room. And we really connected with people. Awesome. So, you came back, which means serverless not completely taking over and making and obviating what we've been doing here for years. Yeah, I think serverless is just another tool in the toolbox. And I didn't want to miss it. So, before I put it in its category, I wanted to make sure that I got super deep with it, used it myself, gave it a fair shot, and it definitely deserves a place. But I think the idea of serverless is the thing that's going to stick. This idea of eliminating as much infrastructure as possible and then putting that everywhere we can. I want to bring that idea of a tool in the toolbox to what we're talking about at this show. So, Kubernetes is one of the most hottest topic at the show, the CNCF now. I mean, there's dozens and dozens of projects here. Dan Kahn, when he kicked it off, talked about Minecraft. And it's like, there's that board there with all the tools, and oh boy, which one do I pick and how do I use it? How do you look at where Kubernetes fits in the overall landscape? Obviously, 12,000 people, it's really exciting. Why is there so much excitement around something that I think is really, it becomes another tool in the tool shed and baked into the platform? I think Kubernetes represents a problem that most people have. If you went down the Linux and then virtualization path, then you ended up with a bunch of virtual machines that you need to go together somehow. So, if you look inside of what Kubernetes has, like the scheduler, how it takes in the pain of running a workload. If you're running VMs in Linux, this is a problem you already have. So, Kubernetes just resonates with almost everyone that is using virtualization. This is why it's so popular. So, it fits. Now, every tool in the landscape may not resonate the same way because everyone doesn't have the same set of problems around the edges, but Kubernetes is a very obvious thing to anyone that's managing more than a handful of machines. Well, I think that brings up an interesting question of as companies and people assemble the stacks, right? Assemble the engines out of the components. Do you have any thoughts on, I guess we could take it from a couple different ways. But maybe as a person coming here for the first time, representing their team, getting started, maybe not involved online with upstream Kubernetes, but trying to make sense of the landscape here and all the different, the zoo of different projects. Lots of new people here. You talk to people, I think, about 50% or more of the people are brand new. People have been ignoring, rightfully so, Kubernetes for four or five years. Maybe I don't need it. I'm good where I am. But we're at a point now where you can ignore it. VMware's offering Kubernetes, every conference you go, whether it's KubeCon or not, this is the thing they're talking about. It's just like Linux was years prior, right? It's just the thing that people are doing. So now you're coming to see for yourself firsthand. You're coming to ask people, how's it going? Now that we're five years in, there's a sense of maturity. Things are slowing down. The ecosystem's getting a lot more mature around it. So you almost have no choice but to be here because now it's in your world. All right, so there's some people that I've been seeing online that are still looking at this a little bit skeptically and said, we've been down this path before. Oh, everybody's involved in Kubernetes. There's my Kubernetes versus some of the other environments. How should we think about that? Because it's going to be baked into VMware when they do project specific and they've got a couple of ways to get you to Kubernetes. Microsoft just announced an update. Is it an interoperability issue? Is this a universal backplane? Do you have a good analogy as to how we should be thinking about where we are today and where we need to go so that we don't repeat the sins of the past when it was the multi-vendor mess that really didn't solve the customer's problems? You can always have multi-vendors because there's too many customers for one vendor to satisfy. That's always going to be the case. There's no way around that. But the way I look at Kubernetes now is like, take the web, click around, web pages, link them together. And out of that, we extract it rest. People can build APIs. We build tooling on top. Cloud providers build APIs to manage infrastructure. So the rest component comes out of the larger picture of the web. And we take the larger components of Kubernetes and we extract out that Kubernetes API. You get Istio. You get these network control planes. You get people building 5G infrastructure using that Kubernetes model. You get all the cloud providers saying now, if the world's going to have this set of APIs that are based on Kubernetes, then I can actually build a global control plane because I can assume the Kubernetes API everywhere, not just for containers, also for networking, authorization, management systems. So it's only natural that people start moving up the stack. And I expect even more pains, even more fragmentation, if you will, because now it's so much easier to explore a new idea even if it's only for a smaller subset of the market. So I expect it to explode. Yeah, one of the things we've been looking at this year is really the simplicity of the offering. You had done Kubernetes the hard way a couple of years back. We've been looking at things like lightweight Kubernetes, the K3S, you know, how are we with that simplicity of the overall solution and making sure that Kubernetes can reach its potential to get to all of those use cases and edge points that you were talking about? Kubernetes' job is to manage the complexity. If you need to run in multiple regions across the globe, that is a set of complexity. Kubernetes has one way of addressing it by sitting on top of all those VMs globally and then providing a set of APIs. That Kubernetes setup and cluster is going to be way more complex than a microKS where you have a single virtual machine where you install the components on one machine. You don't deal with networking. You're not dealing with multiple nodes. That flow is super easy. I think I did a tweet for the canonical folks. They have a tool called microKS. You just run one command, you have a Kubernetes cluster and off you go. And that's great for a developer. But as the underlying infrastructure gets more complex, I think the overall cluster and the components that you need in that cluster matches the complexity. So I think Kubernetes has proven to scale up. And now you can see it's scaling down. So I think it's one of these things that's adaptive to complexity versus having to jump off of the platform because it can't meet either range. So Kelsey, we've talked a little bit about maybe both Kubernetes as this universal API, but also being embedded and being below a lot of application layer and other management layer things. I mean, as you think about talking to our fellow technologists, there are some people who are going to be, we've also used metaphor mechanics. There's some people who are going to be the mechanics, but like everybody drives. So as we get to this level of maturity here now, at KubeCon 2019, any advice on how people should pick, do I need to, and also online we hear a lot about, oh, I don't need Kubernetes. I don't know if I need Kubernetes. I don't know if my particular use case right now, boy, I don't know if I want to go there. So I mean, how should people be looking at it? Also upscaling, should every IT and technologists be, and developer be working towards Kubernetes? Absolutely not. Thank you. If you're managing a bunch of machines, you got two choices. You can build a lot of custom tooling and build something that looks like Kubernetes. Most people don't have the time to do that. So what we want to do is say, look, a lot of people are collaborating on that obvious thing that you should build to manage that. Now, if I give you 80% of your time back, you should go and fill in that gap between what Kubernetes brings to the table and what your developers want to actually do. And at the end of the day, this has always been the same thing. You check in code. It should adopt the company's best practice, and I should be able to get an endpoint and some debugging tools. That has always been the North Star. Even when there was virtualization, early days of cloud, Kubernetes is no different. The thing that Kubernetes represents though is that you don't have to build as much glue between if you're on VMware or you're a pre-early cloud, Kubernetes just builds all of that stuff way up to this line. So maybe you actually finished that CI CD part that you're supposed to do anyway. All right, so Kelsey, every year we try to figure out and distill down the theme of the event. Couple of years ago, the service mesh really extensions. We're going at it here. There's so many different pieces. It's a little tough to kind of pin down. We talked about some of the edge simplicity use cases, securities, of course, been in discussion for a couple of years. Anything that you've distilled so far, the things that you're finding most interesting and new kind of at the edges of this whole ecosystem. This whole thing is a Swiss Army knife. So it depends on who's holding it. Whatever problem they have, that's the piece of the tool that they're going to make front and center. So that's what this is. And right now, I think there's a lot of confusion on do I even need all the other components in this Swiss Army knife? Some people are just like, well, this tool looks interesting. I don't have a problem that this tool is for. And some people are actively creating a problem so they can use the other tools in a Swiss Army knife. I think the biggest thing that I've seen in the last two years is make the new thing work the all way. So you're getting the more traditional vendors showing up and adding their Kubernetes integrations and they're making the new thing more familiar to the people who have the existing tool. And when I look around, that's the thing that I see arise. Hey, that firewall you were using, we now have Kubernetes support. That security tool you were using, we now have Kubernetes support. The security tool works fundamentally the same. It's just now easier to adopt and maybe make Kubernetes things that are deployed in it, leverage those things. So you're saying that's a good thing, not a bad thing? It's a good thing, but it can also be dangerous in some cases where we may get complacent a little bit. And what we end up doing is recreating the world that we tried to run away from a little bit. We try to create a little distance and maybe rethink a few of these approaches, maybe eliminate some need for some of these things. But if we get stuck in recreating the old world on top of the new thing, does it really benefit anyone if we did that for too long? Yeah, it's interesting, because you talk to the enterprise and only 20% of applications are in the cloud. And if you talk about out of my entire portfolio, how many are really new cloud native applications? It's much smaller than that 20%. So we know it's the long pole in the tent of modernization, but you spend a lot of time talking to customers, you're traveling the world. What are some of the kind of best things that you're seeing out here that are helping people adopt those new environments and not just stay complacent, as you said? Pragmatism and leadership. If I see those two things, if there's someone that can make a decision, I see Spinnaker, I see Jenkins, I see a thousand CICD options. Leadership is pick one. They roughly do the same exact thing. You get someone that knows what they're doing, hire someone, get some help, make it work. And then the pragmatism is just be honest about your velocity. You might only bring in the VMs and then you go to containers. So this all or nothing approach never worked. You know it doesn't work. So I think when you have those two fundamental things, then you see a lot of success. And it's not about the age of the enterprise either. There are a hundred early companies are making it work because they have the leadership component and they're very skeptical. So they approach the problem with pragmatism. So they actually get to production. Sometimes faster than the startups that are trying 7,000 things in more of a reckless fashion, the whole thing catches fire. So those are the positive outcomes There's so many tools now. You have your traditional vendors now with skin and the gang giving you documentation. I think right now if you got those two components you're on your path to success. You know, I guess the last thing I want to get your thoughts just on this community these days, a couple of the keynote speakers today really talked about, you know, project over company. And definitely, you know, the open source ethos is front and center at a show here. Give us your viewpoint, how the community is doing any highlights you want to share. So I have one more thing on top of that hierarchy is people over projects always. And then that means that the people should be able to say, hey, I am not wedded to this project forever. There's going to be a time where we have to jump off. There's going to be a time where we have to learn from the other communities. And if you do that then we can actually be on this great path. If we put the projects too much front and center I think we start to miss the boat. Kubernetes, Kubernetes and the rest of the world is moving on. And then we look up, we've missed it. And we actually didn't even get to contribute to the new thing. So I think the biggest part about this community is that hopefully we keep the thing going where we keep reminding people. It's people over these projects. And I think in my keynote I was trying to adjust the idea that we're just kind of pace-setters. You come in, you contribute. All contributions are welcome. Documentation, code or leadership. And then sometimes you've got to jump back out and allow someone else to come in and set the pace and let the ecosystem become the marathon and let it keep running. Well, Kelsey, thank you so much for sharing with our community. I tell you, I've had countless stories of people over the years that have talked about how they've reached out to you. You've helped them along the way. And I know everybody in this ecosystem really appreciates everything that you've helped to move this to where we are today. Awesome, thanks for having me. All right, for John Troyer, I'm Stu Miniman. Superdope coverage of KubeCon. CloudNativeCon continues. We'll be right back. Thanks for watching theCUBE.