 Live from Vancouver, Canada, it's theCUBE. Covering OpenStack Summit North America 2018, brought to you by Red Hat, the OpenStack Foundation, and its ecosystem partners. Welcome to theCUBE's coverage of OpenStack Summit 2018 here in Vancouver. I'm Stu Miniman with my co-host, John Troyer, and happy to welcome back to the program, fresh off the keynote stage Mark Collier, who's the Chief Operating Officer of the OpenStack Foundation. Thank you, Mark. Thanks for having me back. Thank you for having us back, and thank you again for doing the show in Vancouver. So... Oh man, such an amazing place, like this convention center. I don't think it's fair to call it a convention center because it's like a work of art, you know, for us and everything. And it's my second time here for this show, and I think, you know, kudos to your team because you have good enough content that people aren't just wandering around, taking pictures of the mountains. My wife's off, seeing the well-watching right now, but everybody else here, they're engaged, and that's what you want in the community. Yeah, definitely, I guess you have to make sure you don't lose their attention to the whales and the sea planes, but so far, everyone seems to be getting down to business. Yeah, I think it would be fair to say that there's some transitions going on in the marketplace in general. And at this show, I noticed when I got the invitation, it's like the OpenStack slash open infrastructure summit. Got a big track on edge computing, got another one on containers. We've been talking about containers for a few years at this show, which, you know, really interesting to talk about. I mean, the edge stuff, we were talking about it, you know, as NFV and the telcos and all that stuff in the past. What is the OpenStack summit these days, Mark? Yeah, I mean, I think that it's evolving to reflect, you know, what people are doing with open source when it comes to infrastructure. And so, you know, we call it open infrastructure, but basically it's just a world of possibilities that have been opened up by, of course, OpenStack, but also many, many other components, you know, some of which came before, like Lennox and things like that, and some of which, you know, started after, like Kubernetes and there's many other examples, TensorFlow for AI machine learning. So there's kind of this like embarrassment of riches these days, if you want to automate your infrastructure in a cloud-like fashion, you can do so many more things with it. And OpenStack, you know, solves a very specific, very important layer, which is that kind of traditional infrastructure as a service layer, compute storage and networking, but once you automate that, it's proven, it's reliable, you can run millions of cores with it, like some of our users are doing, you want to do more, and that means layering other things on top or sort of connecting them in different ways. And so, just trying to help users get something more out of the technology is really what we're about, and OpenStack becomes like an enabler rather than kind of like the whole conversation. Yeah, one of the things I always say in this industry, sometimes we just don't have the right expectations going into these environments. You know, when I think back 15 years ago as to what we thought Lennox was going to do, oh, it's going to crush Microsoft, it's like, well, Microsoft's still doing quite well, and Lennox has done phenomenal. We wouldn't have companies like Google if it wasn't for the likes of Lennox. And in open source, you've got a lot of tools out there, so while there are the CERNs and wall monitors of the world that take a full OpenStack distribution and put out tons of cores, I've run into software companies where when I dig into their IP, oh, what do you know? There's a project from OpenStack in there that enables what they're doing, so I've seen it a lot of shows, they're like, there's companies that are like, yes, I want it, and then there's like, oh no, there's this piece of it I want, there's that piece of what I want, and that's kind of the wonder of OpenStack that I can do all of those things. Yeah, exactly, I mean, I think we've talked before about calling it composable, open infrastructure, and making OpenStack's always been architecturally designed in terms of the goals around it to be pluggable, so from the beginning you could plug multiple hypervisors kind of underneath, and you could plug different back ends for storage and networking. So that sort of concept of being something, an integration engine that plugs things in is part of the OpenStack kind of philosophy, but now you see that the OpenStack services themselves are sort of, you can think of them as microservices, and like if you just need block storage, you can use sender, and that may make sense for some specific environments, and are you running OpenStack? Well, you are, but it gets a little bit fuzzy in terms of like, well, you're running all of it, or part of it, and the reality is the things are not as simple as a binary yes or no, it's just that the options are much greater now. Well, Mark, that has been some of the discussion in the community over the last few years, the core versus the big 10, and now, of course, with all this interoperability, conversations with both OpenStack participating in other communities and other communities here today, this week, I mean, what's the current state of that conversation about what is OpenStack and how does it interrelate? I think you kind of touched on it with this composable idea. Yeah, I mean, I think that basically it's kind of like, OpenStack is as OpenStack does, so what are people doing with it? And that tells you kind of what it is, and what people are doing with it, there are a lot of different patterns. There's no one specific deployment pattern that everyone uses, but probably by and large, by far, the most common pattern is OpenStack plus Kubernetes, and so when you talk about the interop piece, this is a really good example where OpenStack has evolved to become a better citizen, I guess, of open infrastructure by having more reliable APIs, kind of being a target that tools that build on top can rely on and not sort of have to worry about the snowflakes of different clouds, and there's still more work to be done in that area, but we talked about OpenLab, which is an initiative this morning that puts together OpenStack, Kubernetes, and other pieces like Terraform and things like that, and it does constant end-to-end testing on it, and that's really how you make sure that you know kind of what combinations work well together, and sometimes you just find bugs, and it turns out a couple of changes need to be made upstream in Kubernetes or in OpenStack or in Go for Cloud or in Terraform, and just if you don't know, then the user kind of with the sum assembly required model finds out, and they're like, I don't know, it doesn't work, it's broken, well, is it OpenStack's fault or Kubernetes' fault, and they don't, they just want it to work, so if we can identify that upstream, we can fix it. You're saying OpenStack has become more of a stable layer that other folks can build on them. Yes, exactly, yes, it has become a much more stable layer. Which means there wasn't a whole lot of flashy storage networking compute up on stage, actually. A lot of the talk has been about it. Yeah, that's a really good point. I think it's just, it's really proven in that way, and one of the things that was highlighted was like virtual GPUs, right? So if OpenStack's like designed to be pluggable, what do people want to have as an option now? And in terms of compute storage and networking on the compute side is they want GPUs, because that gives them AI machine learning much faster. If they're Bitcoin miners, like I'm sure you all are in your basement, you know, that they're going to want GPUs. And it was really interesting is that the PTL was like the technical leader of the NOVA project. You know, Gaik got up and talked about virtual GPUs. I was back in the green room and like three of the other keynote speakers were like, oh man, we are so excited about this VGPU support. Like our customers are asking for it. One of the, the guy, Mohammed from Vexos who's the CEO of Vexos, he said, our customers are demanding this Queens release, which is the latest OpenStack. And we were kind of surprised, like they just really want this Queens release. So we asked them why and they're like, well, they want VGPU. So, you know, that's kind of an example of evolution in OpenStack itself, but it's an extension enabler for things like GPUs. And that's kind of an exciting area as well. Yeah, it's interesting, cause in previous years, it was the major release was one of the main things we talked about, Queens. As you mentioned, other than the VGPUs and that little discussion, spent a lot more time outside, talked to a lot of the users. You talked about, you know, the new tracks that were there. And something I heard a lot this year that I hadn't heard for a few years was, you know, get involved, we're looking to build. And I was trying to think of the sports analogy and maybe it was like, you know, okay, we're actually building more of a leak here and we're looking to recruit as opposed to or is it rebuilding, you know, what exactly kind of OpenStack 2.0 is in the future? Yeah, that's a really interesting point. You're absolutely right. And I, you know, can imagine or can remember sort of talking to some of the speakers as they're working on their content. And I don't think I've totally picked up that that was a big trend, but you're absolutely right. That was a major call to action from so many different people. I think it's because, you know, when we think about what we are as a community, you know, I talked about how we're a community of people who build and operate open infrastructure. And you know, it's really about solving problems. And if you want to, if you're as open community collaboration, you want to solve problems, you can't be afraid to stand up and say, we have problems. And sometimes maybe that feels awkward. It's like you're, the tech 101 is like, get up and say you solved all the problems and you should buy it today. You know, it's online or downloaded or whatever. And I think that we just realized like the magic of our community is solving problems. There's always going to be more problems to solve. Now you're putting more pieces together, which means the pieces themselves have to evolve and the testing integration points have to evolve. Like just a new set of challenges and sort of saying, here's what we're trying to solve. It's not done, help us actually is more, I think, true to kind of what the community is all about. I'm wondering, do we know how many people are at the show this year? I don't have the exact count. I think it's around 2600, something like that. Yeah, it's so fair to say it might be a little bit less than, you know, last year's North America show. Yeah, it is a little bit less. And what are you hearing from the users? What are the main things they come for? You've got the new tracks, you've got the Open Dev conference co-located. What kind of key themes can you get from the users? Yeah, I mean, I think one of the things that we found is that we have twice as many cloud architects this year than a year ago. So I think, you know, there's always this period of time where conferences or attendance is driven by curiosity. Like I've heard about this thing, what is it? You know, it's the cool factor. It's hype curves and all that stuff. I want to learn about it. And at this point, like people know what OpenStack is. You know, we've got tons of ways you can learn about it. There's local meetups. There's OpenStack days all over the world. There's content of videos online. Like there's, it's just not like a mystery anymore. Like the mystery draws in kind of the people that are just poking around to learn. Now we're at that point of like, okay, I know what it is. I know what it's for. I want to architect a solution around it. So seeing twice as many cloud architects, I think is an interesting data point to think about how we're shifting more towards people are not asking if it's proven. They're like, it's been proven for whatever, two, three years, however, you know, the perception is. But like the technology is just very, very solid. It's running infrastructure all over the world, the largest banks and so on. So I think that's kind of how things are shifting to what else can we do and put on top of it now that it's a solid foundation. Yeah, I wonder, you know, sometimes there's that buzz as to what's going on out there. There was a certain large analyst firm that wrote a report a couple of months ago that wasn't all that favorable about OpenStack. There's others that, you know, watching on Twitter during the keynote and they're like, they're spending all their time talking about containers. You know, why isn't this just part of the, you know, cloud native con, KubeCon show? What's the foundation's feedback on? You know, what are you hearing? You kind of, you know, what's, you know, your core, you know, deliverables and you know, why this show should continue in the future? Sure, I mean, I think that what we're hearing generally from users and seeing in our data as well as from, you know, analysts like 451 and IDC, those are a couple of different reports coming out like right now or just came out that Jonathan mentioned this morning, I think is adoption just continues to grow. And so, you know, I think people are not looking at just one technology stack and maybe they never were but I think there was this kind of temptation to just think of it, is it containers versus VMs or is it Kubernetes versus open stack? And it's like, no one who really runs infrastructure thinks like that because they might have thought it until they tried it and they realized these things go together. So I think the future of this conference is really just becoming more and more centered around what are the use cases? What are the technical challenges we're trying to solve? And if we, to the extent we're getting patterns and tools that are emerging like the lamp stack of the cloud, so to speak, how can people adopt them? So you think about cloud just taking all kinds of new forms, edge computing, those are the kinds of things I think will become a bigger part of the conference in the future. I do like the open infrastructure angle on this. I mean, as infrastructure folks, right? You know that that storage compute network doesn't manage itself, doesn't configure itself. Doesn't provision itself. And so a lot of the app layer things should rely on this lower layer. And I thought last year in Boston there was this kind of curious open stack or containers conversation which seemed odd at the time and that's clarified I think at a number of levels from a number of camps and vendors. Yeah, I agree. I think we have done our best from our point of view from the foundation, myself and others that are involved in our community to try to dispel those myths or tamper down that kind of sense of a rivalry, but it takes time and I do feel like there is kind of a sea change now. There are just so many people running in production with various container tools, predominantly Kubernetes and an open stack that I think that sort of myth that one's replacing the other is like it's hard for that cognitive dissonance to last forever when you give them like the hundredth example of like somebody running in production at scale. Like they must be doing it for a reason and then people start to go, well, why is that? And I did like the comments you did make about cloud is not consolidating and simplifying, right? Even at the Dell Technologies World Show, Michael Dell got up and talked about the distributed core, which is a little bit of oxymoron, but the fact is compute is everywhere, right? And it's not only, it's on the edge, it's on telephone poles, it's in little boxes in our, you know, they're going to be on our walls, in our walls, right? And this open infrastructure idea can play everywhere. It's not just about an on-prem data center anymore. Yeah, I think that's a big part of why we started to say open infrastructure instead of cloud, just because, I mean, you know, I guess we spent 10 years arguing over the definition of cloud, now we can argue over open infrastructure, but to me it's a little more descriptive and a little less kind of, I don't know, a little less baggage than the term cloud, so. Yeah, definitely differentiates as to where you sit in the marketplace and one thing I definitely want to give the foundation great kudos on, the diversity of this show is excellent. Not just that there was, you know, a welcome happy hour and there's a lunch, but look at all the PTLs, the project leads that are there, a lot of diversity, up on stage. It's just evident and it's just something kind of built into the community, so, you know, great job there. Thank you, I'm very proud of the fact that we had just some excellent keynote speakers this morning and, you know, that's always something that we strive for, but I feel like we got closer to the goal than ever in terms of just getting broad representation up on the stage, so, and some amazing leaders. It's always nice from our standpoint, because we always say, give us your keynote speakers and give us some of the main people making things happen and it just naturally flows that we have a nice diversity, not from gender, from geography. Yeah, absolutely. From, you know, various backgrounds, so that's good. All right, want to give you the final word, you know, takeaways that you want people after the show or maybe some things that people might not know if they, you know, didn't make it here for the show. Yeah, I mean, I think, you know, the number one takeaway is it's all about the people and we want to make it about the headlines or the technology and even the technology is about the people, you know, but certainly the operators are not, like I said, you know, logos don't operate clouds or infrastructure people do, so, getting to meet the people, seeing what they're doing, you know, like Adobe I mentioned, their marketing cloud, they have 100,000 cores of compute and four people operating it, so, if you got the right four people and the right playbook, you can do that too, but you got to meet those people and find out how they're doing it, get their recipe, get their playbook and they're happy to share it and then you can run at that kind of scale too without a big team and you can, you know, change the way you operate. Yeah, I know I said it in my last question, but the last thing, I know there's been a big emphasis to not just do the two big shows a year, but the OpenStack days and other events globally, give people, how do they get involved and, you know, where can they come find out more? Yeah, that's a great question. I'm glad you asked because, you know, there are so many ways to get involved and of course it's online, it's IRC, it's mailing lists, you know, 24 seven, but there's no substitute when it's about the people from meeting in person. So we have the two summits a year, we're also having an event which is called the PTG, which is really for the developers and some of the operators will be coming this fall as well where we're having it in Denver, but the summits are the big shows twice a year, but the OpenStack days are really important, those are annual, typically one to two day events in, you know, 15 plus countries around the world. One in particular that is going to be really exciting this year is in Beijing. We've had that for the last couple of years, huge event, but of course, others throughout Europe and Asia, Tokyo is always an awesome OpenStack day and then there are quite a few in Europe as well, so that's another way you can get involved, not necessarily have to fly around the world, but if you do have to fly around the world, being in Vancouver is not a bad spot, so. Yeah, absolutely, and boy, we know that there's a lot of OpenStack happening in China, so Mark Collier, thanks again to the foundation for allowing theCUBE to cover this and thanks so much for joining us. For John Troyer, I'm Stu Miniman, back here with lots more of three days wall-to-wall coverage here from OpenStack Summit 2018. Thanks for watching theCUBE.