 Live from Boston, Massachusetts, it's theCUBE, covering OpenStack Summit 2017. Brought to you by the OpenStack Foundation, Red Hat, and additional ecosystem of support. And we're back. I'm Stu Miniman here with my co-host John Troyer at the OpenStack Summit in Boston, Massachusetts. A little bit cloudy, but a lovely day here in Boston. Happy to welcome back to the program Mark Shuddleworth who is the founder of Canonical. Mark, pleasure to see you here. Welcome to my home turf. It is great to be here and cloudy is kind of appropriate, I think, under the circumstances. Yeah, absolutely. I mean, spring in Boston, you never know what you're going to get. The OpenStack Summit, sometimes, you know, we're not sure what we're going to get. So, I want to start, you know, been a lot of changes, you know, happening at your company. You've got a new role change coming on. Why don't you bring us up to speed onto the company in general, and then we'll dive into it. Sure, so the three legs of computing, personal computing, data center computing, or cloud computing, and the Edge, the IoT world, which is neither data center nor personal. Of those, I had dreamed of Ubuntu as sort of going mainstream in all three. Clearly Ubuntu is the de facto standard now for cloud computing and the data center, and also, I think, arguably for the Edge. Where we failed, and I feel responsible for miscalculating effectively, is our push into personal computing, phones, tablets, PCs. And so I think the desktop remains really important to us in support of developers, who are really the lifeblood of free software and open source and IT innovation. But as a business, we chose to focus on these two, the cloud, where we are very strong, and IoT, where I think the story's only just beginning. But again, you know, we're at the center of everything you've ever read about self-driving cars, for example. That required some changes in the business, and those are sort of at an emotional level challenging changes. But from the point of view of the people who depend on us, who are really our customers in cloud and at the Edge, that's better focus on the things they care about. Okay, and you're going to be stepping back in as the CEO role. A lot's changed in seven years. Maybe you can comment as to what it was like the last time you did this and what your expectation is now. Yes, so I have my training wheels on as the CEO. We've had the great privilege of having Jane Solberg, who was CEO previously stepped into the CEO role. And under her watch, you know, we've had seven years of pretty extraordinary commercial growth. My focus has always been first the community of Ubuntu and then the product of Canonical. Jane really built the business. And she's put us in a great position for the next step. After seven years, she felt that was the right thing to be doing. So I will take the reins of something that commercially is very different from Ubuntu when she took it on. And it's exciting, but I'm also mindful of the fact that there are a bunch of disciplines that she's mastered that I now have to take the reins on. All right, so Mark, open stack in general, there's many that watch there and says, does anybody make money on it? So I'll start with yes or no, yes? Yeah, yeah, we do, yeah. And in fact, you know, we've always been kind of contrarian and counter cyclical. There was a time in open stack when the meme was open stack will dominate everything and everything as a service only matters if it's inside open stack. I called that bullshit as a service and the big tent that will collapse. And I think now we've seen that, right? In fact, now the time, now the vibe in the air is one of fear and uncertainty. Actually, I feel more confident than ever that the pieces that matter of open stack really matter, right? Really, the big choice that a CIO has today is between renting CPUs from the public cloud. And that's a great choice. That's an amazing option. We work to make Ubuntu on Amazon, Azure, Google, SoftLayer, Rackspace, Oracle Cloud, an amazing option for the CIO. But then there's always the own the cause option and this is where open stack has to play. I keep reminding folks here, don't worry about Amazon, they don't actually compete with open stack, right? And they don't want to compete with open stack. They're really good here in the public cloud. Open stack's only competitor is itself, right? And so it's really about this community's focus and desire to focus on the important thing, which is virtual machines, virtual networks and virtual disks, right? So we are now seeing, I think, that resolve, that focus emerging. Jim talked, I think, eloquently about the fact that the line between the Linux and the open stack is very blurry. There really isn't room there for other players. So essentially this boils down to people who choose REL for their open stack and people who choose Ubuntu for their open stack. The data are very clear. The substantial majority of deployments do choose Ubuntu and we're proud of that and pleased to invest in that. We consider that a successful business. And we think, you know, there's ample room for growth there for ourselves as well as other players like Red Hat. Well, Mark, I think here at open stack summit this year in Boston, we see five, 6,000 people, actual implementers here in the hallways going past us, not just vendors. So we're through that part of the hype cycle. Of course, now the thing that's confusing people or the thing that's consuming cycles in the conversation is containers. Docker versus Kubernetes, we heard a lot of Kubernetes this morning, but actually containers as a technology, aside if you're after you rip off the labels, super interesting in and of themselves. How is Canonical and Ubuntu looking at containers, LXD, Docker, the new open source projects around container run times? How are you all approaching that and their relationship to open stack? So just the numbers first, if you look at the various kind of artifacts that people using containers leave around, you'll find the vast majority of containering is done on Ubuntu, right? So whether you're interested in Docker, whether you're interested in Kubernetes, whether you're interested in Mesos, whether you're interested in LXD, you'll find the majority of that's on Ubuntu. And so that's really important to us. We don't want to essentially prescribe to people that Docker is the right way to do things or Mesos is the right way to do things. We simply want to say, look, this is a fantastic new way of operating software. We want innovation, we want competition, we want to be a platform on which you can consume all these kinds of containers. So that's the first thing. And I feel pretty successful there in the sense that just the vast majority of containering right now is being done on Ubuntu and that's good. The second thing I'd say is look, what's really emerging is that we have two kinds of containers. We have containers which are essentially processes on demand. So if you think of Docker, if you think of Kubernetes, to a certain extent Mesos, what you're really managing there is a collection of processes. They have nothing to do with underlying machines. In fact, the connection between the app, the process and the machine is sort of tenuous, right? And then we have infrastructure for managing machines and really those are quite separate and we can compose them together very easily, right? So OpenStack doesn't need to be anxious about Docker because Docker runs just great on OpenStack, right? Docker runs just great on VMware and on bare metal and on public clouds, right? So you can really separate in your minds the way you think about infrastructure for machines and then the way you think about managing processes on demand, right? And those two can fit together just as well as you like. The only slightly confusing piece for people is in fact our kind of container, the kind that we lead, which is not to say it's just a thing that we do because no one else really was interested in it, which is kind of containers for legacy workloads, which is LexD and those are really containers that are designed to look like machines, right? So they sit alongside Docker, they sit alongside Rocket and Kubernetes and all of those things. They're just a way to kind of get the performance and latency benefits and density benefits of containers on old workloads where you may have thousands of VMs and you want to move that to containers without actually re-engineering the application. So that's what LexD will do for you. But other than that, really, you can think of it as infrastructure managing machines. You've got bare metal machines with mass, you've got OpenStack, you've got public clouds and you've got LexD, those are all machines. And then processes on top of that, MISOS, Docker, Kubernetes, they're all great. And those two don't really fight with each other at all. Mark, can you give us any guidance on when it comes to revenue in the cloud space? We were teasing out with Jim a little bit, some of his eight-figure deals, I think it was, that have it, a lot of some of the telcos. People look at the cloud space and they're like, is it a million dollars, is it 10, any guidance you can give on cloud in general for you? Of course, everyone likes to slice and dice their numbers and no one likes to get sliced and diced. But I would say we crossed a couple of big milestones, I think publicly, and news is out that we crossed $100 million in bookings, which is a great milestone for us. And in terms of growth, our private infrastructure business grew 70%, public infrastructure business grew 90% in the last 12 months. We really do feel now that we have competitors that we respect in Red Hat, the great company, great product, RAL. We do feel that the game has changed, which really opens people up to saying, okay, maybe I should be multi-sourcing my OS. And we do now see consistently that we're closing multi-million dollar deals with companies who are very happy customers of VMware and Microsoft and Red Hat, but they now see a rationale for getting commercial support for Ubuntu from Canonical for their enterprise infrastructure, both on cloud and on-prem. In terms of customers, do you see a typical profile as we're looking forward with OpenStack here? What is a typical Ubuntu OpenStack? If you look at the OpenStack data, for example, 54% of the production OpenStack are in Ubuntu across all sectors, so there isn't a particular sector there. You've got the mind-share and the math. I'm wondering kind of what are the applications you're seeing and what are the use cases? We really sort of blazed the trail onto Telco first, and I think now you're seeing other vendors talking about Telco as a specific focus. The Lighthouse Telcos using OpenStack are all essentially on Ubuntu. AT&T very publicly. There are two others here in the US that we consider now to be multi-sourcing, which is great for us. In Europe, the leading clouds are on Ubuntu and in Asia as well. Actually now I'd say there's a lot of work to be done shifting Telco operations towards OpenStack, but the base principle of OpenStack as what they call the NFVI, the network function virtualization infrastructure, what the rest of us call cloud, OpenStack is that and Ubuntu is likely the majority of that. So now I think the focus shifts towards some other industries. In media, we see dramatic disruption of media. You guys are part of that. This is an incredible operation that couldn't have existed 15 years ago, but now look what you've done. Netflix is part of it. All on Ubuntu. Big, over-the-top operators now are starting to use Ubuntu as their operational weapon essentially to do media at scale, whether that's transcoding or delivery or clever things at the edge with devices. Ubuntu tends to be there. We see finance emerging. So looking at the payments providers, that seems to be where the finance industry is moving to Ubuntu. Certainly the guys who are doing stuff at scale, PayPal and others, see Ubuntu as a very important new piece of their infrastructure and we're very happy to support them. eBay's talk publicly about their work on Ubuntu with containers and all sorts of and OpenStack and other infrastructure pieces, but also mainstream banks now moving to Ubuntu. And then retail. Retail is a more complex environment, but actually the interesting thing in retail is IoT because if you're managing hundreds of thousands of fridges or you're managing hundreds of thousands of stores, there are a lot of things, a lot of smart things that happen in there from security cameras to smart doors to stock control, infantry control, and the control of HVAC, refrigeration and so on. A lot of that is going smart, which means it's becoming apps on Ubuntu effectively on those devices. We've reached a certain state of maturity in the core OpenStack. You talked about how Ubuntu is being used a lot in the edge. What do you see as some of the real challenges we have to face there? Is it security like always? Is there other pieces where your company and your partners are working at trying to solve that? I love this insight that the internet, which was designed to withstand a nuclear strike, can't withstand an attack by rogue refrigerators, right? And it's really true that we have to think of the software stack inside all of these boxes that were sticking on the roof or sticking up above the ceiling board or sticking underneath the floorboards. We have to start thinking about that as critical infrastructure, whether you're a government or a large institution, or just someone who manages a building, right? If that infrastructure is compromised, then in some sense you're compromised, right? And so bringing the ability to be very precise about the version of the operating system that is on this thing and that thing, and have control over that, and then being very precise about the version of the apps on top of that, that's now getting serious consideration. It's a super hard problem. I think there's some very elegant technical breakthroughs that we've made, snaps, for example, are a really beautiful way of kind of learning from Docker, but instead of focusing on kind of cloud type applications, focusing on these edge type applications where you want to deliver exactly the right set of bits onto exactly the right device at exactly the right time. It's fun stuff. It also, one of the reasons I really love what we do is because innovators are fun to work with. And in cloud, the innovation is kind of esoteric and abstract. Have you ever actually seen a Kubernetes, right? Can you relate to a Kubernetes? They're kind of a machine learning. Have you ever seen a learning machine? You know what I mean? It's amazing, fantastic stuff, but it's all in your brain and your mind. Whereas with IoT, people can relate to buildings and refrigerators and drones and self-driving cars and they can see business opportunities for themselves in making those things better through software. And so that's what's a lot of fun for me is just is connecting to entrepreneurs and innovators in a very sort of like practical way, right? Starting with a Raspberry Pi or an Intel Nook and then just showing something amazing. It's pretty exciting times for IT. You talk about all those interesting new technologies. I'm curious, when you talk to business leaders, where's their state of mind? They're excited about the technology. They're worrying about being greatly disrupted. What do you see as some of the big challenges? So for example, on OpenStack, I describe people's state of mind as the same as it was about the internet after the dot-com crash, right? It's not like people stopped wanting to connect to the internet, right? And in the same way, the institutional imperative to be able to both rent CPUs and own CPUs, that fundamentally hasn't changed at all, right? But what's been taken a lot is a lot of the puff around and a lot of the fluff around what it means to actually own CPUs, right? What that's got to feel like for the business is the way it feels for Facebook or Amazon or Google, right? Which is at racks show up, they're automatically inventory. They're instantly assigned to Windows workloads or CentOS workloads or Ubuntu workloads or Rail workloads, right? And with no humans being involved, you know, stuff's getting on. Now I'm owning CPUs or renting CPUs, but the experience is the same, right? It's dynamic, economics are different and I'll choose what economics I want, but it's instant, it's immediate, it's frictionless effectively. OpenStack on-prem has to fit into that. CIOs want to consume OpenStack effectively inside a box, right? Which is not to say a black box, but it is to say inside an economic box. You know what a thousand cores on Amazon are going to call you, they'll cost you, you know what they'll cost you on Google or on Azure, and you need to know the same thing on OpenStack. So that's really what we're doing for OpenStack is giving the business the ability to consume OpenStack with very predictable operational costs. Whether that's through a fully managed service, BootStack, right? You know, it's literally, it will cost you so many dollars per day per server, however big you want it. If you want to make it smaller today, you can make it smaller if you want to make it bigger today and you get everything you need, storage, compute, upgrades, operations, everything's just like a cloud, right? So that's one way to do it. Another way to do it is just to train people to achieve those same kinds of predictable operational characteristics themselves. So that's really what I think is going on. The mind of the CIO is, with regard to cloud is maturing. They're starting to understand that you won't get fired for buying from a public, a brand name public cloud provider. All of the big banks are signing those contracts. It's great. You also need to be able to operate your own private infrastructure with very predictable economics. And if you can do that, then you've put your business in the position that it needs to be in. Absolutely. No number of the keynote speakers that we saw from the users are ones that I've seen in the public cloud shows. They're here. We understand it's a multi hybrid cloud world. Mark Shuttleworth. In the keynotes, they talked about one of the largest open stacks in the world moving to fully managed services. That's on Canonical. That's Canonical's BootStack. Well, Mark Shuttleworth, really appreciate you joining us. We'll be back with lots more coverage here from OpenStack Summit 2017. You're watching theCUBE.