(electronic music) >> Narrator: Live from Boston, Massachusetts it's The Cube covering OpenStack Summit 2017. Brought to you by the OpenStack Foundation, RedHat and additional ecosystem support. >> Welcome back, I'm Stu Miniman joined by my cohost John Troyer. We always want to give the community what they want. and I think from the early returns on day one, we brought back Mark Shuttleworth. So Mark, founder of Canonical, had you on yesterday. A lot of feedback from the communities, so welcome back. >> Thank you, great to be here and looking forward to questions from the community and you. >> Yeah, so let's start with, we love at the show you get some of these users up on stage and they get to talk about what they're doing. We were actually, John and I, were catching up with a friend of ours that talked about how a private cloud, the next revision is going to use OpenStack, so really, OpenStack's been a little under the covers in many ways. The composability of OpenStack now, we're going to see pieces of it show up a lot of places. We've heard a lot about the Telco places, maybe talk about some of the emerging areas, enterprise customers, that you find for Ubuntu and OpenStack specifically? >> Sure. Well it seems as if every industry has a different name for the same phenomenon, right. So, for some it's "digital", for other's it's essentially a transformation of some aspect of what they're doing. The Telcos call it NFV, in media you have OTT as a sort of emerging threat and the response, in every case, is really to empower developers. That's why it's such a fun time to be a software developer, because the established guys realize that if they aren't already competing with Silicon Valley, they're going to be competing with Silicon Valley. So in each industry there's a sort of challenges or labels that they give this process of kind of unleashing developers and it's fun for us, because we get to be part of that in many cases. I think the big drivers under the hood, other than the operational and economic dynamics of cloudification, I think the really big changes are going to be machine learning, which seems to be moving very quickly into every industry. Retailers are using it for predictive analytics on what to put in store or what to recommend online. It just has this huge effect on almost any business when you figure out how to use your data in that way. All of that is developer driven, all of that needs this kind of underlying infrastructure to power it and it's kind of relevant to every industry. For us media is a key prospect, you know that we've done very, very well in Telco. Media is now a sort-of critical focus. Companies like Bloomberg for example us Ubuntu as an elastic platform for agility for the developers. They're a pretty astonishing operation; media company, but very tech-centric, very tech-savvy. I don't know if you've had them on the show. In retail, Ebay, PayPal it's kind of a crossover finance. They're all using Ubuntu in that sort of way. They may now see the major financials who are looking at the intersection of machine learning and transactions systems effectively as the driver for that kind of change. >> Stu: So in our last interview we talked about are companies making money in OpenStack and your answer, resoundingly, was yes. >> Mark: For us, certainly, yeah. >> One of the things we always look at is kind of the open source model itself. I was at DockerCon a few weeks ago, it's like everybody's using Docker. How do they make money? The question I get from a number of people in the community is, everybody I talk to knows Ubuntu, uses Ubuntu, when do they transition to paying for some of the products? >> Well so one of our key tenants is that we want to put no friction in front of developers. So many of the people that you'll meet here or that you'll meet at other developer-centric summits, they're developer-oriented. They're creatives, effectively. So our products, our commercial products aren't really designed to tax developers effectively. What we want is developers to have the latest and greatest platforms, to have that absolutely free, to be able to have confidence in the fact that it can go into production. When applications get into production, a whole different set of people get involved. For example the security guys will say, does this comply with FIPS security? And that's a commercial capability that customers get from Canonical if they wanted so we're now getting a set of security certifications that enable people to take apps on Ubuntu into production inside defense industries or other high security industries. Similarly if you look at the support life cycle, our standard public free support maintenance window is five years, which is a long time, but for certain applications it turns out the app needs to be in production for 10 years and again that's a driver for a different set of people. Not the developers, but for compilers and system administration operation types to engage with Canonical commercially. Sometimes we would walk through the building and the developers love us as everything's free and then the ops guys love us because we will support them for longer than we would support the developers. >> Can we talk about Open Source as a component of business models in general maybe, and how you would like to see the ecosystem growing, and even Canonical's business model. In the course of the last decade in the industry itself, right, a lot of people sniping at each other; "Well, you know open core is the way to go, open source is not a business model" there's a lot of yelling. You've been around, you know what works. How do you a set of healthy companies that use open source develop in our ecosystem? >> So this is a really, really interesting topic and I'll start at the high end. If you think of the Googles, and the Facebooks, and the Amazons, and the Microsofts, and the Oracles, I think for them open source is now a weapon. It's a way to commoditize something that somebody else attaches value to and in the game of love and war, or Go, or chess, or however you want to think of it, between those giants open source very much has become a kind of root to market in order to establish standards for the next wave. Right now in machine learning for example we see all of these major guys pushing stuff out as open source. People wouldn't really ask "what's the business model" there 'cause they understand that this is these huge organizations essentially trying to establish standards for the next wave through open source. Okay, so that's one approach. On the startup side it's a lot more challenging and there I think we need to do two things. So right now I would say, if you're a single app startup it's very difficult with open source. If you've got a brilliant idea for a database, if you've got a brilliant idea for a messaging system, it's very, very difficult to do that with open source and I think you've seen the consequences of that over the years. That's actually not a great result for us in open source. At the end of the day, what drives brilliant folks to invest 20 hours a day for three years of their life to create something new, part of it is the sense they'll get a return on that and so, actually, we want that innovation. Not just from the Googles, and the Oracles, and the Microsofts, but we want innovation from real startups in open source. So one of the things I'd like to see is that I'd like to see the open source community being more generous of spirit to the startups who are doing that. That's not Canonical, particularly, but it is the Dockers of the world, it is the RethinkDBs, as a recent example. Those are great guys who had really good ideas and we should caution open source folks when they basically piss on the parade of the startup. It's a very short-sighted approach. The other thing that I do need to do is we need to figure out the monetization strategy. Selling software the old way is really terrible. There's a lot of friction associated with it. So one of the things that I'm passionate about is hacking Ubuntu to enable startups to innovate as open source if they want to, but then deliver their software to the enterprise market. Everywhere where you can find Ubuntu, and you know now that's everywhere right? Every Global 2000 company is running Ubuntu. Whether we can call them a customer or not is another question. But how can we enable all those innovators and startups to deliver their stuff to all of those companies and make money doing it? That's really good for those companies, and it's really good for the startups, and that's something I'm very passionate about. >> We've seen such a big transformation. I mean, the era of the shrink wrapped software is gone. An era that I want to get your long term perspective on is, when it comes to internet security. Back to your first company, we had Edward Snowden and the keynote this morning talking about security, and he bashed the public cloud guys and said "We need private cloud, and you need to control a lot more there" any comments on his stuff, the public/private era and internet security in general today? Are we safer today than we were back in '99? >> We certainly are safer in part because of Edward Snowden. Awareness is the only way to start the process of getting stuff better. I don't think it's simplistically that you can bash the public clouds. For example Google does incredible work around security and there's a huge amount of stuff in the Linux stack today around security specifically that we have Google to thank for. Amazon and others are also starting to invest in those areas. So I think the really interesting question is, how do we make security easy in the field and still make it meaningful? That's something we can have a big impact on because security when you touch it it can often feel like friction. So for example we use AppArmor. Now AppArmor is a more modern of the SC Linux ideas that is just super easy to use which means people don't even know that they're using it. Every copy of Ubuntu out there is actually effectively as secure as if you've turned on SC Linux, but administrators don't ever have to worry about that because the way AppArmor works is designed to be really, really easy to just integrate and that allows each piece of the ecosystem, the upstreams, the developers, the end users to essentially upgrade their security without really have to think about that as a budget item or a work ticket item, or something that's friction. >> Mark, any conversations on the show surprise you? Excite you? There's always such a great collection of some really smart and engaged people at this show. I'm curious what your experience has been so far. >> Sure. I think it's interesting. Open Stack moved so quickly from idea to superstar. I guess it's like a child prodigy, you know, a child TV star. The late teens can be a little rocky, right? (Mark laughs) I think it will emerge from all of that as quite a thoughtful community. There were a ton of people who came to these shows who were just stuffed, effectively, there by corporates who just wanted to do something in cloud. Now I think the conversation is much more measured. You've got folks here who really want these pieces to fit together and be useful. Our particular focus is the consumption of OpenStack in a way that is really economically impactful for enterprises. But the people who I see continuing to make meaningful contributions here are people who really want something to work. Whether that's networking, or storage, or compute, or operations as in our case but they're the folks who care about that infrastructure really working rather than the flash in the pan types and I think that's a good transition for the community to be making. >> Can you say a little more about the future of OpenStack and the direction you see the community going. I don't know. If you had a magic wand and you look forward a couple of years. We talked a lot about operability and maintainability, upgradeability, ease of use. That seems to be one of the places that you're trying to drive the ecosystem. >> One of the things that I think the community is starting to realize is that if you try to please everybody, you'll end up with something nobody can really relate to. I think if you take the mission of OpenStack as to say, look, open source is going to do lots of complicated things but if we can essentially just deliver virtualized infrastructure in a super automated way so that nobody has to think about it, the virtual machines, virtual disks, virtual networks on demand. That's an awesome contribution to the innovation stack. There are a ton of other super shiny things that could happen on any given culture and ODS but if we just get that piece right, we've made a huge contribution and I think for a while OpenStack was trying to do everything for everybody. Lots of reasons why that might be the case but now I think there's a stronger sense of "This is the mission" and it will deliver on that mission, I have great confidence. It was contrarian then to say we shouldn't be doing everything, it's contrarian now to say "actually, we're fine". We're learning what we need to be. >> The ebb and flows of this community have been really interesting. NASA helped start it. NASA went to Amazon, NASA went back to OpenStack. >> Think about the economics of cars, right. It's kind of incredible that I can sit outside the building and pull up the app, and I have a car. It's also quite nice to own a car. People do both. The economics of ownership and the economics of renting, they're pretty well understood and most institutions or most people can figure out that sometimes they'll do a bit of either. What we have to do is, at the moment we have a situation where if you want to own your infrastructure the operations are unpredictable. Whereas if you rent it it's super predictable. If we can just put predictability of price and performance into OpenStack, which is, for example what the manage services, what BootStack does. Also what JUJU and MAAS do. They allow you to say, I can do that. I can do that quickly, and I don't have to go and open a textbook to do that or hire 50 people to do it. That essentially allows people now to make the choice between owning and renting in a very natural way, and I think once people understand that that's what this is all about it'll give them a sense of confidence again. >> Curious your viewpoint on the future of jobs in tech. We talked a little bit before about autonomous vehicles. It has the opportunity to be a great boon from a technology standpoint but could hollow out this massive amount of jobs globally. Is technology an enabler of some of these things? Do we race with the machines? We interviewed Erik Brynjolfsson and Andy McAfee from the MIT Sloan School. Did you personally have some thoughts on that? In places where Canonical looks about our future workforce, do we end up with "coding becomes the new blue collar job"? >> I don't know if I can speak to a single career but I think the simple fact is there's nothing magical about the brain. The brain is a mesh network competing flows and it makes decisions, and I think we will simulate that pretty soon and we'll suddenly realize there's nothing magical about the brain but there is something magical about humans and so, what is a job? A job is kind of how we figure out what we want to do most of the day and how we want to define ourselves in some sense. That's never going to go away. I think it's highly likely that humans are obsolete as decision makers and surprisingly soon. Simply because there's nothing magic about the brain and we'll build bigger and better brains for any kind of decision you can imagine. But the art of being human? That's kind of magical, and humans will find a way to evolve into that time. I'm not too worried about it. >> Okay. Last thing I want to ask is, what's exciting you these days? We've talked about space exploration a few times. Happy to comment on it. I mean, the last 12 months has been amazing to watch for those of us. I grew up studying engineering. You always look up to the stars. What's exciting you these days? >> Well the commercialization of space, the commercial access to space is just fantastic to see, sure, really dawning and credit to the Bezoses and the Musks who are kind of shaking up the status quo in those industries. We will be amongst the stars. I have no doubt about it. It will be part of the human experience. For me personally, I expect I'll go back to space and do something interesting there. It'll get easier and easier and so I can pack my walking stick and go to the moon, maybe. But right now from a love of technology and business point of view, IoT is such rich pickings. You can't swing a cat but find something that can be improved in a very physical way. It's great to see that intersection of entrepreneurship and tinkering suddenly come alive again. You don't have to be a giant institution to go and compete with the giant institutions that are driving the giant clouds. You just have to be able to spot a business opportunity in real life around you and how the right piece of software in the right place with the right data can suddenly make things better and so it's just delicious the sort of things people are doing. Ubuntu again is a great platform for innovating around that. It's just great fun for me to see really smart people who three years ago would say, do I really want to go work at a giant organization in Silicon Valley? Or can I have fun with something for a while that's really mine and whether that's worth 12 bucks or 12 billion who knows? But it just feels fun and I'm enjoying that very much, seeing people find interesting things to do at the edge. >> Mark Shuttleworth, appreciate being able to dig into a lot more topics with you today and we'll be right back with lots more coverage here from OpenStack 2017 in Boston. You're watching the cube. (electronic music)

>> Announcer: Live from Austin, Texas. It's theCUBE, covering DockerCon 2017. Brought to you by Docker, and support from its ecosystem partners. (bright electronic music) >> Welcome back. I'm Stu Miniman, joined by Jim Kobielus for two days of theCUBE's live coverage, DockerCon 2017, here in Austin, Texas. We are the worldwide leader in live enterprise tech coverage, happy to welcome to the program, a first-time guest on theCUBE, happens to also be a local here in the Austin area, so Dustin Kirkland, the Ubuntu Product and Strategy, with Canonical, thanks so much for joining us. >> Thanks, Stu. >> All right, so Dustin, give us a good thumbnail, what's your role, and how excited are you to be at another local show. All the open source shows seem to be here in Austin. I mean, we love doing it. >> I'm super glad. >> Dustin: We love sharing Austin. Glad for people to come and visit. Just make sure you go home at the end of it. (chuckles) >> Jim: Keep Austin weird and keep it open. >> That's right, that's right. Yeah, it's great to be local, it's great to have the Docker community back in Austin. It was, a lot of these people were here for OpenStack. We'll be back for CubeCon later this year. OSCON in between. >> All right, and tell us a little bit about your role. >> Yes, so I lead Product and Strategy at Ubuntu. We make an operating system that runs in the cloud, on public clouds, private clouds, bare-metal, physical servers, down to desktops and embedded devices. >> Okay, so, I have a serious question for you. Every time we see the surveys of OpenStack, the surveys in the public cloud, Canonical's always there. I mean, everybody's using your stuff. >> Dustin: Good! >> But where are people paying money for it? What's kind of from the business standpoint, maybe you can give us the quick update on that. >> People pay money when it's mission-critical. When Ubuntu and OpenStack and soon, Kubernetes, certainly more and more, Docker, when that's part of the mission-critical infrastructure, they pay for that. They pay the support and the services, they pay for consulting, for design, for leads, for architecture. They pay for access to the product roadmap, and so we do have some really brand-name customers who pay us good money for that. >> Okay, it's our third year doing theCUBE at this show, and every year, it seems we come in with one of the same questions, which is like, all right, is this ready for production, is anybody using it? We backed you to knock down the doors of everybody here, and give us more customers to talk to, so, what do you see, what's your answer to that? >> Yeah, I mean, it strikes me as really odd when people are still asking, "Are containers ready "for production?" Containers have been part of our DNA in Ubuntu for almost 10 years now. Shipping an OS that boots into a container that's able to run LXD containers, Docker containers, and run those at tremendous scale. We'd run containers underneath as the control plane of every OpenStack cloud we've ever deployed, every Kubernetes cloud we've ever deployed, every Hadoop cloud we've ever deployed. So containers are part of our production system. >> So do you guys have a marketing term? You guys are the hipster Linux container company. You were doing it before it was cool. >> I guess so, I mean it's, I guess, it's like asking, and I wonder, you think cellphones are mainstream yet? It's like, yeah, it is now, but you're probably one of the first in your family to have a cellphone, right? It's, we're kind of at that juncture, where we've been doing for a long time, and it's good to see others finally taking advantage as well. >> In the keynote this morning, we talked, we saw a lot about the maturation of Docker. They really started out working with the developer, they've really grown, working with the business, working with the enterprise. Talk to us about your customers as it fits into the container space in general, Docker, specifically. What are you guys seeing? >> As an operating system that delivers the latest and greatest open source software across multiple architectures, public and private clouds, Docker fits into that very well, in fact. It sits alongside LXD at giving that machine container, replace your VM's experience, but also the new way of writing applications. Solomon talked about applications, and if you're going to develop an application, Docker is a great application development platform. So when applications are being developed, (mumbles) or microservices, from scratch, Docker is a fantastic approach, and we see more developers using Ubuntu desktops and Ubuntu in the cloud, as that development platform. As that matures, then we get into a situation where it becomes mission-critical, and then we have really interesting commercial discussions around how do we really help that platform succeed? >> All right, we just Microsoft on the program. >> Dustin: John, right? >> Yeah, John was on, talks about, (mumbles) Microsoft is talking about being open, Microsoft's talking about choice. They actually talked, John mentioned, your company and your operating system. When we get to cloud solutions, Canonical's supported everywhere. How do you guys differentiate? How do you make sure that they're choosing your product as opposed to something else? >> So Ubuntu itself, always latest and greatest. It's fresh, you're never more than six months away from the next latest and greatest everything across the board. You're never more than two years away from an LTS, a long-term support release. That's really the key differentiator for Ubuntu is its freshment, its velocity, and that maps very well to the container world, where things are revving very, very quickly. >> All right, security was a big focus this morning also. What's your viewpoint as to where security lives, how that works with all of your environment, and what you guys do for that-- >> I've been a security nerd for most of my career. In fact, it's one of those jobs you leave but you always kind of get sucked back into because you care about it, honestly. Ubuntu as a platform, security, we take very seriously. Encryption anywhere, we can use encryption, updates, latest and greatest updates, kernel patches, Livepatch for the kernel. (coughs) Livepatch for the kernel is particularly interesting from a security perspective because it enables us to address security vulnerabilities without rebooting systems, and that's really important in a containerized environment, where you're not just running one or two machines, you're running potentially thousands of machines or containers or applications, and being able to update one single kernel with a Livepatch, without rebooting any of them, that's what security people are excited about when we talk Ubuntu kernel and security. >> (mumbles) Ubuntu being deployed into Internet of things, or to what extent is your roadmap going in that direction 'cause we're seeing a lot of new development going into the Internet of things, to deploy artificial intelligence and deep learning algorithms and data, down to the edge, and so-- >> Yeah, it's beautiful, I mean, that edge-to-cloud story is something that we've got a very clear view on. We produce an OS, an Ubuntu OS called Ubuntu Core, is a read-only operating system custom-tailored for IoT devices. That's the OS, it's the same Ubuntu but rolled and managed and updated in a different way. Applications fit onto that device in the form of snaps, or Docker containers, frankly. They're a little bit different in the way that they're implemented, but we have a new packaging system that's well-adapted, well-tuned-- >> A snap is more, something different from a container, how? >> It is, it's a form of a container. It's less than a container, but it uses some of the same container primitives. It's, frankly, it's an archive and a set of security profiles that wrap that tarball, essentially, and the way it's executed in a very secure manner, so it's wrapped with AppArmor profiles, it only has access to certain parts of the system, it contains its own dependencies, but they're contained in such a way that they're protected from the rest of the system. A lot of that sounds like Docker, and it is similar to Docker, but Docker provides a little bit more of that machine experience. Docker will include a file system, it'll draw an IP address sometimes, or defroute traffic, whereas a snap actually runs directly on the underlying OS. It's more tightly linked to that OS. In terms of linking back to the machine learning, that happens in the cloud. Inevitably, IoT drives more cloud adoption because those little IoT devices, they've got so little processing power and storage by design, that information needs to go somewhere, and it goes to the cloud, where something like a TensorFlow, running in a Docker Swarm, or a Kubernetes, or some combination of those two, are really crunching the interesting problems. >> First, Google recently made a big to-do about federating more of the machine learning algorithms all the way to the edge device, so, the world is going in that direction but I hear you. That's, they're very constrained-- >> Dustin: We hear a lot about the edge. >> To run the algorithms that pull power on the edge device, but it's coming. >> Yeah, for sure. >> Great. >> Stu: All right, so Dustin, I heard Kubernetes and Swarm, you guys, agnostic to that, support all of it. >> Dustin: We are. >> What do you guys code on, what do you hear from customers? >> Yeah, so we're very proud of our position here. I'm here at DockerCon, supporting Docker. Docker Inc. is a close commercial partner of Canonical. We, Canonical is authorized to resell Docker Enterprise Edition, Docker services, Docker support. We've got mutual customers who buy that directly from Canonical, and we support Docker and Swarm and Datacenter on top of Ubuntu, and that's a great story that brings us from the developers who are running Docker on Ubuntu on their Macs and Windows machines. John, I'm sure, was talking about Windows and Docker. But when they put that into production, we've got the wherewithal to support that. We offer Kubernetes as another platform. I've spoken with some really bright, just last night, with a really bright cloud architect from a major Internet service provider, and their role is they set up Docker Swarms for their internal customers, and Kubernetes Clusters for their internal customers, and Cloud Foundries, and OpenStacks, all inside of this big telco Internet cable giant, and it makes sense, and they can do all of that, and do all of that on top of Ubuntu, because it's the platform that can offer whatever they need for their customers. >> All right, one of the other announcements in the keynote this morning was LinuxKit, so, I got a little bit of a preview before the show, and I don't feel that it was Docker trying to punch at the providers of Linux, and it didn't seem to come off that way in the keynote, but for those that hear at a glance, oh, wait, LinuxKit developed with a bunch of, you know, seems like mostly hardware companies plus Microsoft and Docker. What do you guys see, how do you look at that? >> It's genuinely fun for an open source engineer to put together a Linux distribution. It's like the thing you want to do, and customize it and tailor it, and the beauty of open source is you can absolutely do that, and so, what I saw from LinuxKit, I too got a little preview, it seems it comes out of the part of Docker that also works on unikernels, Alpine, to an extent, and they've built a container-optimized, or Docker-optimized OS from Docker, so if you want Docker all the way down, it sounds like LinuxKit is a solution that they're working on, still working on. I'll say that Ubuntu, containers are in our DNA, we built a kernel and we built a security system around containers for quite some time, and we continue to optimize that, and we work directly with Microsoft, Google, Amazon to ensure that the Ubuntu that's running in those public clouds is ready to run Docker and other container systems out of the box, and very consistently, in a way that looks exactly like the Ubuntu that's running as the bash shell on the Windows desktop, as the Ubuntu desktop itself, as the server that you might run in any one of the public clouds. It's a very consistent experience. We do tune that and tailor that, but it's in ways that ensures portability. >> All right, so Dustin, you talked about kind of the history and how long people have been using it. Production should not be a question. It's just where, what, how you're doing this. What things do you still see us needing to mature, or what excites you about this going forward? >> Yeah. The management, honestly, and that comes back to security. Ensuring that running those containers at scale, you're doing that in a secure manner. Minimal is part of it. We hear that quite a bit, that, "I want a minimal image, I want a minimal host." That is an important part of it. It's, we have to be a little bit careful that we don't go so minimal that we end up creating a bunch of snowflakes, special unicorns where every container image is a little bit different, every host is a little bit different, because it's more minimal than the previous one. That actually creates more security problems, so I think thinking that problem through is, it's one of the most important problems that I think through, or I'm working on right now, and I think others are interested in working on as well. >> All right, Dustin, you've been way too pleasant through all of this interview, so before we end up, as an Austin local here, I have to ask you the divisive question. Your favorite barbecue place. (Dustin groans) >> You know-- >> Jim: Your favorite bar band, too. Keep going. >> Okay, yeah, I mean, you can't go wrong with the award-wining Franklin's barbecue or the gas station Rudy's, we love those. My favorite's a little hole in the wall out close to where I live. It's a trailer that's been serving barbecue out of that trailer since 1997. It's called Bee Caves barbecue. Those guys, they put together some fantastic barbecue five days a week. They sell it until they're out, and then they close up the shop and they go fishing, and it's, you got to get there early, and when they're done, they're done, so I-- >> Yeah, is there a connection between people that make barbecue and people that put together Linux distributions? It sounds like a lot of the same thing. >> Maybe so, maybe so, yeah. I've got a smoker out back. I like to smoke meat as much as I can. >> Absolutely, all right, well, Dustin, really appreciate you joining us. Welcome to the >> Stu, thank you, Jim. >> Stu: CUBE alumni list now, and we'll be back with more coverage here from DockerCon 2017. You're watching theCUBE. (bright electronic music) >> I remember--