 from our studios in the heart of Silicon Valley, Palo Alto, California. This is a CUBE Conversation. Hello everyone, welcome to this special CUBE Conversation here in Palo Alto, California at the CUBE studios at the CUBE headquarters. I'm John Furrier, the host of the CUBE. We're here at Dustin Kirkland, product manager at Google, friend of the CUBE in the community with Kubernetes. Been on theCUBE, CUBE alumni. Dustin, welcome to the CUBE Conversation. Thanks, John. It's a beautiful studio. I've never been in the studio. On the show floor a few times, but this is fun. Great to have you on. Great opportunity to chat about Kubernetes, of which you do a lot of some product management work at Google, but really more importantly on this conversation is about the fifth anniversary, the birthday of Kubernetes. Today we're celebrating the fifth birthday of Kubernetes. Still a toddler. Absolutely. Still growing. You think about how, you know, Linux has been around for a long time. OpenStack has been around these other big projects that have been around for, you know, going on decades in Linux's case, and Kubernetes has grown so fast, but it's only five years old, you know? You know, I remember at an OpenStack event in Seattle many, many years ago. I think it was six years ago, CUBE's on its 10th year, so we're going to have these lookbacks moments. This is one of them. I was having a beer with Lutucker, JJ. Kismatic was like one of the first companies at the time, didn't make it, but we were talking about OpenStack. I was like, this Kubernetes thing, this is really hot. This paper, this initiative, this could really be the abstraction layer to kind of bring all this, you know, cloud native. Wasn't thought of at the time, but it was like more of, you know, OpenStack, trying to move up the stack. And it turned out it ended up happening. Kubernetes then went on to change the landscape of what containers did. Docker got a lot of credit for pioneering that, got the big VC funding, became a unicorn, and then containers kind of went into a different direction because of Kubernetes. Very much so. I mean, the modernization of software infrastructure has been coming for a long time, and Kubernetes sort of brings it all, brings it all together at this point. But putting software into a container, we've been doing that in different forms for a lot of time, for a long time. But once you have a lot of containers, what do you do with that, right? And that was the problem that Kubernetes solves so eloquently and has, you know, now for a couple of years, and it just keeps getting better. You know, you mentioned modernization. Let's talk about that, because I think the modernization theme is now pretty much prevalent in every vertical. I'll be in DC next week for the Amazon Web Service Public Sector Summit where modernization of governments and nations are being discussed. Education, modernization of IT, we've seen it here. The media business that we're participating in is about not where you store the code, it's how you code, how you build. Is a mindset shift. This has been the real revelation around the DevOps movement, infrastructure as code, now called Cloud Native. Share your thoughts on this modernization mindset because it really is how you build. Yeah, I think the cross-pollination, actually, across industries, and we see that even just in the word containers, right, and all the imagery around shipping and shipping containers. We've applied these age-old concepts that have been, I don't know, perfected, but certainly optimized over decades of, actually, centuries or millennia of moving things across water in containers, right? But we apply that to software and boom, we have this step-function difference in the way that we manage and we orchestrate and administer code. That's one example of that cross-pollination. And now you're talking about optimizing governments or economies, but being able to maybe then apply other concepts that we've come a long way in computer science. DevOps is a good example. Applying DevOps principles to non-computer fields. Just think about that for a second. And it's mind-blowing. And if you think about also the step-function you mentioned, because I think this actually changed a lot of the entrepreneurial landscape as well, and also has shaped open source. And big news, this quarter is MapR is going to shut down this Hadoop, one of the big Hadoop players. Cloudera Merge with Hortonworks fired their CEO. The founder, Mike Olson, has retired. Some say forced out. I don't think, so I think this is more of his time. Amar Adal is still there. Open source is a business model. Can we be the red hat for Hadoop? The red, not really kind of the viable, but it's evolving. So open source has been impacted by this step-function. There's a business impact. Talk about the dynamics and the step-function, both on the business side and on how software is built, specifically open source. You know, you and I have been around open source for a long, long time. I think it started when I was in college in the late 90s and then through my career at IBM. And it's interesting how on the fringe open source was for so long, and so much of my IBM career and then early time I spent on site at Red Hat. It was something that was, it was different. It was weird. It was very much fringe worthy, right? But now it's the mainstream and it's everywhere and it's so mainstream that it's almost the de facto standard to just start with open source. But you know, there's some other news that's been happening lately that you didn't bring up, but it's a really touchy aspect of open source right now. And that's on some of the licenses and how those licenses get applied by software, especially databases, when offered as a service in the cloud. That's one of the big problems I think that we're working with in the open source. Summarize the news and what it means. What's happening? What's the news and what's the real business or technical impact to the licensing? What's the issue? What's the core issue? Yeah, so without taking judgment, any way, shape, or form on this, the TLDR on this is a number of open source databases. Most recently, Cockroach, DB, have adopted a different licensing model that is non-standard from an open source perspective. And from one perspective, they're adopting these different licensing models because other vendors can take that software and offer it as a service. And in some cases- Like Amazon. Like, sure, you said it. And offer it as a service and maybe contribute, maybe pay money to the smaller startup or the open source community behind it, but not necessarily. And it's, in some ways, it's quite threatening to open source communities and open source companies. In other cases, it's quite empowering. And it's going to be interesting to see how that plays out. The tension between open sourcing software and eventually making money off of it is something that we've seen for at least 25 years. And it continues to go on today. And this is, to me, a real fascinating area that I think is going to be super important to keep an eye on. Because you want to encourage contribution and openness. At the same time, you look at the scale of just the Linux Foundation's numbers. It's pretty massive in terms of now the open source contribution when you factor in even China and other nations. It's on exponential growth. So is it just open source? Is the model not necessarily a business model? So this is the big question no one knows. I think we've crossed that and open source is the model. And this is where me as a product manager has worked around open source. I've spent a lot of time thinking about how to create commercial offerings around open source. I've spent 10 years at Canonical, the first half of which as an engineer, the second half of which as a product manager around Ubuntu, building services, commercial services around Ubuntu. And I learned quite a few things that now apply absolutely to Kubernetes as well as to a number of open source startups that I've advised. And kind of given them some perspective on maybe some successful and unsuccessful ways to monetize that open source. Okay, so Dustin talk about, let's get back to Kubernetes. So I think this is the next level talk track is as Kubernetes has established itself and landed in the industry and has adoption, it's now an expansion mode. So land, adopt, and expand. We've seen adoption, now it's an expansion mode. Where does it go from here? Because if you look at the tail signs, things like service meshes, serverless. So you've got some interesting trends that are going to support this expansionary stage of Kubernetes. What is your view about this next expansionary wave? What comes next? Yeah, I think the next stage is really about democratizing Kubernetes for workloads that, you know, it's quite obvious when Kubernetes is the right answer at the scale of a Google or a Twitter or a Netflix or, you know, some of these massive services that it is obviously and clearly the best answer to orchestrating containers. Now, I think the next question is how does that same thing that works at that massive scale also work for me as a developer at a very small scale? Helps me develop my software, my small team of five or 10 people. Do I need a Kubernetes if I'm a five or 10 person startup? Well, I mean, not the original sort of Borg vision of Kubernetes. It's probably overkill. But actually the tooling has really advanced and we now have Kubernetes that makes sense on very small scales. You've got things like K3S from Rancher. You've got micro-Kates from my colleagues at Canonical. Other ways of making shrinking Kubernetes down to something that fits perhaps on devices, perhaps at the edge beyond just the traditional data center and into remote locations that need to deploy and manage applications. On the Kubernetes clustering and some of the tech side, you know, we've seen some great tech trends as mentioning Cloudera and Hortonworks and MapR. Let's take Cloudera and Hortonworks. I remember back in the old Hadoop days when it was booming, people say, oh, they were so proud to talk about their clusters. I stood up all these clusters and then I would ask them, well, what are you doing with it? Well, we're storing data, I think. So it became kind of this use case where standing up the cluster was the use case and they're like, okay, now let's put some data in it. So the question for you is, with Kubernetes it's a little bit different. We're not seeing, they were seeing real use cases. What are people standing up Kubernetes clusters for? What's specific besides saying I've done it? Yeah. What's the main use case that you're seeing that has real value? Yeah, actually, you just jogged to mine of really funny memory back in those big data days. I was a CTO of a startup, we were encrypting data and we were helping encrypt healthcare data for healthcare companies. And the number of healthcare companies that I worked with at that time who said they had a big data problem and they had all of, I don't know, three terabytes worth of data that they needed to encrypt it was kind of humorous sometimes. Is that really a big data problem? This fits on a single disk, you know? But yeah, I mean, it's interesting how that works. But the hype of the tech was preceding the reality. The needs. The needs. So it's Kubernetes, so I have a Kubernetes cluster for blank, fill in the blank. What are people saying? Yeah, it's largely about the modernization. So I need to modernize my infrastructure. I'm going to adopt a platform that's probably not the old Java WebSphere type platform or something like that. I'm investing in hardware. I'm investing in software, middleware. I'm investing in people. And I want all of those things to line up with where the industry is going from a software perspective. And that's where Kubernetes is sort of the cornerstone piece of that. Linux, of course, that's pretty well established. Right? So in this continuous delivery and integration piece of it is that the pipelining, where's the fit on the low-hanging fruit use cases of Kubernetes? Just development process or? It's the operations. It's the operations of now I've got software that I need to deploy across multiple versions, perhaps multiple sites. I need to handle that upgrade ideally without downtime in a way that you said service mesh, in a way that meshes together, it makes sense. I've got to roll out new certificates. I need to address the security vulnerability. These are all the things that Kubernetes does such a better job at than what people were doing previously, which was a whole lot of for loops, shell scripts and SSH pushing tar balls around, maybe devs or RPMs around. That is what Kubernetes actually really solves and does an elegant job of solving as just a starting point. And that's just the beginning. And without getting vendor here, Anthos is the thing that we at Google have built around Kubernetes that brings it to enterprise. The other day I did a tweet, I called it Anthem, I was typing too fast. I got a lot of crap on Twitter for that. You mentioned Anthos. Multicloud has been a big part of where Kubernetes seems to fit. You mentioned some of the licensing changes. Cloud has been a great resource for a lot of the new web scale applications from all kinds of companies. Now IT with serverless, you're seeing a lot more of these capabilities. How do you see the next shift with data state coming in? Because you've got stateless data and you've got stateful data. This has become a conversation point. Yeah, I think Kelsey Hightower has said it pretty eloquently as he usually does around the serverless movement. It lets developers focus on just their code and literally just their code. Perhaps even just their function and just their piece of code without having to be an expert on all of the turtles all the way down. That's the big difference about serverless. Having written a couple of those functions, I can really invest my time on the couple of hundred lines of code that matter and not choosing a distro, choosing a Kubernetes, choosing all the stack underneath. I simply choose the platform where I'm going to drop that function, compile it, upload it, and then riff and rev on that. Fifth anniversary of Kubernetes. We're riffing on Kubernetes. Dustin Kirkland here inside the KubeCube alumni. You were recently at the KubeCon in overseas in Europe. Barcelona. Barcelona, great city. Kube's been there many times. Stu was there covering for us. I couldn't make this trip, unfortunately. Had a couple of daughters graduating, so I didn't make the trip. Sorry, guys. What was the summary? What was the takeaway? What was the big walk away from that event? What synthesized the main stories? What were the most important stories being told? Big news, big observations. It was a huge event to start with. It was at Fira Barcelona. It didn't take over the whole space, but I've been there a number of times from Mobile World Congress. This is KubeCon in the same building that hosts all of Mobile World Congress. So I think 8,000 attendees was what we saw. It's quite celebratory. I think we were doing some pre-fifth birthday bash celebrations. Key takeaways. Hybrid cloud, multi-cloud. I think that's the world that we've evolved into. There was a lot of tension, I think, in the early days about must stay on-prem, must go to the cloud. There's going to be a winner and a loser and everything's going to go one direction or another. I think the chips have fallen and it's pretty obvious now that the world will exist in a very hybrid multi-cloud state, ultimately. There's going to be some stuff on-prem that doesn't move. There's going to be some stuff better hosted in one or more public clouds. That's the multi-cloud aspect. And there will be some stuff at the edge in remote locations, in vehicles, on oil rigs, at restaurants and stores and so forth. What's most exciting from a trend statement? What's getting you excited from what you see on the landscape out there? Yeah, so tying all of that to Kubernetes, Kubernetes is the thing that basically normalizes all of that. You write your application, put it in a container and expect to Kubernetes to be there, to scale that, to operate that, to upgrade that, to migrate that over time. From that perspective, Kubernetes has really ticked all the boxes and you've got a lot of choices now about which Kubernetes you're going to use and where. Beyond Kubernetes, a lot of variety of projects. You've got Kubeflow, you've got service meshes out there, a lot of different projects. What's a dark horse? What's something that's out there that people should be paying attention to that you see emerging that's notable that should be paying attention to? I think it's a combination of two things. One is pretty obvious and that's AIML is coming like a freight train and is sort of the next layer of excitement I think after Kubernetes becomes boring. Which hopefully if we've done our job well, that Kubernetes layer gets settled and will evolve but the sort of the hockey stick hopefully settles down and it becomes something super stable. The application of machine learning to create artificial intelligence, conclusions, trends from things, that is sort of the next big trend. And then I would say another one if you really want the dark horse, I think it's around communications and I think it's around the difference in the way that we communicate with one another across all forms of media, voice, video, chat, writing, how we interact with people, how we interact with our tools, with our software and in fact how our software interacts with us and our software interacts with other software. That communications industry is ripe for some pretty radical disruption and you know some of the organizations in there doing that, it's early, early days on those changes. Final point, you mentioned earlier in our conversation here about how DevOps is influencing and impacting non-tech and computer science related. What did you mean by that? Well, I think you brought up unexpectedly in that you were looking at the way some other industries are changing and I think that cross-pollination is actually quite powerful. When you take and apply a skill or an expertise you have outside of your industry but it adds something new and interesting to your professional environment, that's where you get these provocative operations, these really creative, innovative things that no one really saw coming. DevOps principles apply to other disciplines. Agility, taking down waterfall-based processes. That's one phenomenal example. Imagine that for governments, right? To remove some of the pain that you and I know, I've got to go and renew my license, my birthday's coming up, I've got to go and renew my driver's license. You know how much I'm dreading going to the DMV? Root canal, driver's license, all the same. Exactly, how waterfall is that experience and could we be more agile, more DevOps-y in some of our government process? The U.S. government's procurement practices are based upon 1990 standards. They still request a manual, a physical manual for every product they buy. Who does that? I know that there are organizations trying to apply some open-source principles to government but think about just democracy and how being a little bit more open and transparent in the way that we are in open-source code, the ability to accept patches. I have a side project, a passion for brewing beer and I love applying open-source practices to the industry of brewing and that's an example of where I use professional work to complement a hobby. All right, we've got to bring some cube beer. Can we private label some cube beer? If you like sour beer, I'm into sour beer. I hope that's okay. We like to get the IPAs, double IPAs for us. Final question for you, five years from now, Kubernetes is going to be 10 years old. What's the world going to look like when we wake up five years from now with the impact of Kubernetes? Yeah, I think, I don't think we're struggling with the Kubernetes layer at that point. I think that's settled science and as much as Linux is pretty settled science. Yes, there's a release and it comes out with incremental features and bug fixes. I think Kubernetes is settled science, management of those containers is pretty well settled. Five years from now, I think we end up with software, some software that's writing software and I don't quite mean that in the way that sounds scary and that we're eliminating developers, but I think we're creating more powerful, more robust software that actually creates that software and that's all built on top of the really strong, robust systems we have underneath. Automation to take the heavy lifting but the human creation still keep part of it. Humans are in the loop. We're many decades away from humans being out of the loop on creative processes. Dustin Kirkland, he a product manager at Google, Kubernetes, Guru also, Cube alumni here in the studio, talking about the Kubernetes 50 year anniversary. Of course, the Cube was president in creation during the beginning of the wave of Kubernetes. We love the trend, we love cloud, we love talking about tech. I'm John Furrier here in Palo Alto. Thanks for watching.