 All right, good morning. Hi, everybody. I'm Robert Cathy, and I'm here to talk about OpenStack Momentum, what's happening and why it's important, or if you get the reference, how OpenStack got her groove back. You know, I want to start off by talking about how we all love open source. We wouldn't be here if we didn't love open source, but sometimes this is kind of what open source feels like. You have an idea. You want to solve a problem, and you want to solve that problem with your friends. You go talk to them, and you share ideas. Maybe you get a pizza, and you invite some friends, sit around the table, pull out a dry erase marker, and come up with a solution. You start getting really excited about that idea. So you take it out to the community, and lo and behold, you find out another group of people with a pizza and a dry erase board have been working on the same problem, and they've come up with a different way to solve it. Maybe different technologically, maybe different architecturally, but the ideas might not be compatible. So this process of conflict, this process of advocating for how you want to solve the problem is core to what makes open source work. And when you, you know, you can try to persuade people, but at the end of the day, this conflict is just going to be a part of the process. And of course, it's really not like that fist fight. It's more organized than that. Conflict is a good thing. Conflict requires that you advocate for your idea, and in that process, the best ideas usually come forward. And a good analogy here is a rock polisher. You take some relatively ordinary-looking brown stones, you pick them up and put them in the can, and you close it up, and you run that rock tumbler for five or six hours, and lo and behold, what you find when you open it up is these beautiful polished stones. That process inside the can is kind of like what happens in an open source community that's functioning properly. There's friction, there's conflict, ideas banging against each other, and what comes out is something that's better than what came in. So the process might not be pleasant, but if it works correctly, it provides value. Now, that all may be true, but at some times, it kind of feels a little bit like you're in a tunnel. You engage with your group of people trying to solve a very specific technical problem, and you can get kind of lost in that process sometimes. Do we go forward, do we go backwards? Which way do we go to try and get where we're trying to get? And of course, as your world is kind of small inside the tunnel, and you're hearing from the same people over and over, maybe you're hearing the same ideas echoing off those walls, and you can lose sight of a broader perspective, and the fresh air that comes from being outside that tunnel, and once, of course, once you get to the surface, things feel a little bit better, and so that's what I wanna invite you to do with me today is kind of come outside the tunnel, come up to the surface, we're gonna look around at a broader perspective of what's happening, not just with the open-stack project, but with the open infrastructure community broadly defined. The specific things I wanna talk to you about, innovative technologies are happening within our community. These applications that are being built on top of these innovative technologies are delivering real value in production. Businesses are relying on them. There are massive scale deployments of these technologies pretty much everywhere you look. If you notice from the keynotes on Monday, it was one user after another user after another user who are running 100,000 cores of these technologies in production, 200,000 cores, or even brand new technologies, or relatively new technologies in open infrastructure, like Cota containers, Baidu's running tens of thousands of cores in production. There's a growing market opportunity here, and I'm gonna argue with you that the growth rates that are being estimated for open infrastructure are comparable to the growth rates that we see in hyperscale proprietary public clouds. That's a good thing. And when you look at open-stack specifically, you're seeing a highly trusted piece of software that's running real infrastructure in production for business critical workloads. And then you also see technologies like Bare Metal that are making the promise of delivering technologies like Kubernetes on premises much more palatable and easy to deploy and manage. So the community is what's driving this. That process of those rocks banging into each other, of conflict is what is making all of this work, and I'm gonna show you how the open-stack community is open-source done right. But first, I don't know if you get this reference, but Conor McGregor is a mixed martial arts champion from Ireland, and during a pre-show press conference, a challenger was really smack talking to him, and he just turned around with his microphone and said this. So I am certainly not the Conor McGregor of open-source, and there are any number of people in this community who are more qualified than I am to give this talk, and some of them I put some names up here of folks that I've reached out to during the process, and I just wanted to say thank you. So if I've left somebody out, please forgive me. But alas, you don't get to hear from any of these great people, you get to hear from me today. So who am I? So as I said, my name's Robert Kathy. I have a consultancy called Kathy Communications. We've worked with the Open-Stack Foundation for a number of years, trying to help with getting the message right and understanding how to engage with communities and build the kind of conversations we need to build. We work with a number of companies both in open infrastructure broadly, infrastructure technologies, developer tooling, things like that. We've been a member and me specifically have been a member of this community almost since the beginning. This is actually my 15th summit, and I've helped a lot of different people put talks together, develop abstracts, build slide decks, but this is the first time I've actually given one. So be nice to me. So I want to go back to the beginning, not all the way to the beginning. It turns out we only need to go back about 10 years, and let's talk a little bit about how all this started. So if you're not familiar, and many of you of course are familiar, it's important to get this context of where all this started so we can kind of take this journey together over the next 30 minutes or so about how we got where we are today. When Rackspace got together with NASA, and Rackspace said, hey, we've developed an object storage technology called Swift, and you've had Anzo Labs, who is a contractor to NASA, build a compute environment called Nova. We could put these together and create an open source, an open source cloud infrastructure software project. Now curious thing in the United States when a government agency builds a piece of software, unless there's a specific national security interest or something like that, it is by law open source. It has to be available to the taxpayers who paid for it. So this worked out great. 2008, 2009 timeframe, things started. In 2010, the OpenStack Foundation was formed to create an entity to hold the intellectual property that represented the code and the trademarks, and also to create a governance process and a licensing model that allowed people to have confidence about contributing to OpenStack and consuming it. And the goal was simple. We wanna build an open source alternative piece of software with which you could build and operate a public cloud. So what happened? I mean, that was where we started. Where did we end up? Well, as we all know, you start on a journey and the road has twists and turns and it winds around corners and you don't always know when you start out what that road is gonna be like. But as you go through the process, as you go down that road with your friends, you discover that sometimes the destination can be more rewarding than maybe you were even thinking when you started. It's gonna be different, but it can be more rewarding. And that's what's happened with OpenStack and Open Infrastructure. It's really important to, as you look over the broad landscape, remember, OpenStack is the most widely deployed open source software for building and operating infrastructure clouds, full stop. 75 public clouds are running OpenStack today. Now of course, none of them are at the scale that Amazon and Azure and Google are at, but they're out there and there's very specific reasons why organizations might wanna build their public cloud based on OpenStack. But even setting that aside, the private cloud applications and the new applications and Edge and so forth are equally exciting to that. And some of those, you think these are things that we didn't anticipate when OpenStack started out. We weren't thinking about a bare metal provisioning environment management system. We weren't thinking about GPUs. So these are all things that happened along that journey that have gotten us where we are today. So there's some things we learned in this process. As we walk down this road. First thing, and this is really instructive and it's something that I think a lot of us knew intuitively, but didn't fully appreciate. And it's the fact that just complicated issues create an environment where things can be easily misunderstood. And more specifically there, infrastructure software, we'll talk a little bit about infrastructure software being hard in a moment, but there are a lot of details and complexities that had to be explained to a broader community about how OpenStack worked. And when you're talking about details, you're not talking about big picture. And sometimes that creates challenges and it did for OpenStack. Another important learning, change does not happen continuously. Change happens fitfully. There are fits and starts. There's go back to square one and start over. It's this, and another thing, change happens asynchronously with other change that's impacting what you're trying to do. These things happen to OpenStack. Another important thing that we learned along the way, value for business happens up the stack. Value is created where the application developer lives. And early on in OpenStack, the model was, well application developers are going to consume these infrastructure primitives using the APIs and Nova and Swift and Cinder and Glance and they're gonna go build their applications on top of that. Well then this different abstraction technologies came along and could sit on top of OpenStack or could sit on top of some other infrastructure software. The developer started to move away from the infrastructure. And it's important and instructive to remember that if you're running a business and you're trying to build a product in a service, the value is close to the developer. This is what I was alluding to a moment ago. Infrastructure is hard. People have said throughout time, well OpenStack is hard. Well yeah, it is hard, but infrastructure is hard. It doesn't matter who's running it. You can pay someone else to run your infrastructure for you and consume that as a service or you can learn it and have your own team that will go deploy that for you. You can hire, you can license a piece of software that can manage your infrastructure for you. You can have a managed service on your own premises. But it doesn't matter who's running it. It's difficult to do. Another important thing that we learned, and I think we knew this intuitively again, but if you look at how Amazon Web Services became successful early on, they did it by having a laser focus on the developer experience. When Amazon launched, you have to remember this. They only had three or four services. You could get EC2, you could get S3 and maybe two or three other things. Nobody was seriously running production workloads on early Amazon. They were using it for development and testing environments. And it was great for developers because rather than filling out a ticket with central IT and asking them to go buy a rack of gear and waiting for that to be provisioned and shipped in and then racked and cabled and lay down an operating system, six months goes by and you finally have your environment. No more. With Amazon, you take a credit card and in five minutes you have what you need in order to go do your work. It fundamentally changed the way developers did their work. And Amazon understood this and they focused on those developers, not the CIO. And then finally, we really learned that you have to be careful about listening to your own press. Technical reality is seldom reflected by what you hear being said by pundits and thought leaders and subject matter experts because going back to the beginning, complicated things are easily misunderstood. And if you're talking about details, you're not talking about the big picture. So these are all the things that we learned but probably most importantly, we learned that community when it's done right is a powerful tool for transformation within an open source project. And it can help you take the technical direction and the application directions in places that matter and deliver value for the people in the community. All right, so those were some of the things that we learned along the way in this process. There was a time in OpenStack where OpenStack was the coolest kid in school. Everybody wanted to be OpenStack's friend. And just like it's the case in high school, a lot of the people who wanted to be OpenStack's friends weren't exactly sure why they were there but they knew that something cool was happening and they wanted to be a part of it. So you saw summit attendants going way up and you saw media buzz being super positive about OpenStack but the key thing, there's nobody wanted to miss out even if they weren't sure how they were gonna monetize it, how they were gonna contribute to it and be a part of it. Now this is probably familiar to a lot of you. Just show hands how many people are familiar with the Gartner hype cycle. Okay, so Gartner put this together to try and simply and graphically explain the process by which new technologies are introduced to a market and then how users interact with those technologies to try to get to the point where they deliver some business value. Now this one I think is maybe last year and it's just emerging technologies. And so you can see some of the early technologies. These are the very early stages and as people get more and more excited about it, the expectations go way up and then time goes by and pretty soon people realize wow this doesn't work like we thought and you can see as of last year Gartner had augmented reality at the bottom of the trough of disillusionment. But then hopefully over time people keep working on these technologies and they listen to the people who are consuming them and it gets to the point where business value is created. So Gartner created this and it's just like any simplification. It's never really this simple in real life but it's a good way to think about how technologies are introduced and OpenStack followed a curve that looks a lot like this. So I wanna now talk a little bit about each of those major sections for OpenStack and how we went from OpenStack to open infrastructure and what happened along the way. Now I wanna do a disclaimer here and a qualifier. This is just for context. The history of OpenStack is one of those things people like to get beer and argue about and I would be happy to buy you some beer. The restaurant next door to the Yangtze Hotel has a deal where you can get 100 pints of Kirin for less than 200 RMB. So it works out to about 25 cents per beer and if you wanna argue with me later or have a conversation about the history of OpenStack I'll meet you at the restaurant and I'll buy you a beer and we can argue. So let's start with the early days of OpenStack. It was very straightforward. The message was you get compute storage and networking infrastructure primitives, API's. Your developers can go consume this in an open source software where you can control your destiny and you're not beholden to proprietary license software models and you can come and engage with the community and tell us what features and services you want throw some developers at it and we're all gonna go build this together. Startups were everywhere. You had Piston and Nebula and Cloud Scaling and MetaCloud and there's Gary down there. Hey Gary, I'm giving MetaCloud a plug. There you go. And it's important, you think about this for a minute. OpenStack was so hot. It was so popular that there were arguments about who got to claim to be a founder of OpenStack and these weren't friendly arguments. People got really animated about what their role was in starting OpenStack. So enterprise software vendors wanted to get in on the action and a lot of these software vendors and by a lot, I mean almost all of them made their money. Their business models were built around proprietary commercial off-the-shelf software that's licensed and you pay a license fee per CPU or per node that's running your software. So they didn't really understand open source and how you make something work with open source. Now Red Hat was an obvious exception to that. Red Hat obviously had a deep understanding of open source but most of the companies early on in OpenStack didn't and they had to learn that. That was a hurdle. The rallying cry that you heard pretty much everybody saying in the early days of OpenStack was we're gonna make OpenStack easy for you. We're gonna make OpenStack easy by doing one of several things. We're gonna take the code and we're gonna create a distribution on that that you will then pay us a support fee to make sure that it runs properly and do all the updating and so forth for you so you don't have to do that. Very similar to how Red Hat Enterprise Linux works. That was one model. Another model was we're going to run it for you on your equipment. Another model was we'll allow you to consume it as a service in a public cloud. So there are lots of different ways but the consistent message was we're gonna make OpenStack easy. And the other thing that was happening at this time is the community was forming. The community was trying to figure out how are we going to govern ourselves? How are we going to organize? How are we gonna make decisions about which pull requests get merged? Who's gonna lead the projects? How are we gonna decide who leads the projects? All of that had to be figured out in those early days. So by 2013 to 2015, the hype had really gotten heavy with OpenStack. And at the same time, some of the very real challenges were starting to become evident. We saw big names like Walmart and AT&T and Deutsche Telecom and BMW stepping up to say we're using this software to solve business problems. That was huge. As these new companies came in and started contributing to the integrated release cycle to two releases per year, as more companies started doing that, that early presence of rack space in the community while they were still there and doing great work to support the project by proportionately, it was getting smaller. That was a sign of health. Lots of companies participating and helping to build the software. Frustrations were starting to grow, however, over just how challenging it was to deploy and operate and upgrade OpenStack. Now the community was well aware of these concerns because they were consuming the software. They were using it. So every release cycle, new innovations were being brought into OpenStack to improve that process and improve that experience. So much was happening within the community and so many new projects were being created that we had to answer some fundamental questions about what is and is not OpenStack. If you're not running Swift, if you're not running Nova, is it an OpenStack cloud? So today we don't talk too much about these things because those issues have largely been resolved, but we had to put together a group called Defcore early on to try and deal with that because everybody wanted to say we're OpenStack and somebody had to say, well, are you really running OpenStack? Do you get to say that? So we had to deal with that. We also had to deal organizationally with what do we do with all of these projects and 20 and 30 projects that are now part of OpenStack? How do we organize that? And the compromise that was created was called the Big Tent. And it largely didn't work. It was just too unwieldy. But these were all symptoms of a growing community that had to figure out how to manage itself. Also at this time is when Docker came into the picture. And this is a really interesting example of one of those learnings I mentioned at the beginning about how stay close to the developer because that's where the value is. Well, Docker knew this. And by taking Linux containers and saying, we're gonna put an orchestration environment around that so that application developers can code on their laptop and push to production much more simply. And you don't have to worry about all those infrastructure primitives. There was a big value proposition there and Docker took advantage of it. This was also about the time you started to see acquisitions happening in OpenStack. So you had cloud scaling getting bought by EMC, now Dell EMC. You saw MetaCloud getting bought by Cisco and let's see, Piston got bought by HP and Plumgrid got bought by VMware. Bluebox got bought by IBM and probably one of the biggest acquisitions in the OpenStack space. And the other thing that started happening at this time is consolidation started happening. Remember those distros I mentioned? At one point there were close to 20 OpenStack distros. Now by comparison, I think there's something like 80 Kubernetes distros right now, a Kubernetes distros a fundamentally different animal than an OpenStack distro is. But there were a lot of choices out there. If you were an enterprise and you said, we'd like to do on-premises infrastructure as a service so that our developers can consume it and we wanna use OpenStack. You had a lot of choices for how you were gonna do that and those choices confused the market to some extent. This consolidation of people getting out of the distro business got it down to a manageable set. As we turned into 2016 and 17 however, things started to get more complicated for OpenStack. And as those storm clouds, if you wanna call them that, began to collect on the horizon, we knew things were gonna get interesting and we knew that we were gonna have to make some choices. We could stay the way we were and let the inevitable happen or we could adapt. We could try to find a new environment. We could look for new opportunities for what can be done with Open Infrastructure. All right, so as things started to get more complicated, what happening? What are some of these gets? Kubernetes got interesting. Kubernetes became something that was so exciting to this Open Infrastructure broad landscape that people started to say well, so if I've got Kubernetes, do I need OpenStack? Do you, what's the relationship between these two projects? And from a technology standpoint, how do I deploy them and use them together? Of course, Kubernetes is very useful if you're in a public cloud setting where someone else is managing the infrastructure for you. So early on, that was a big challenge that we had to work through. This was also at about the time that proprietary public hyperscale cloud became about more than AWS. This was at the time when Azure started to come onto the scene in a strong way as well as Google. Enterprise CIOs finally got cloud. They finally started to understand, oh wait a minute, this isn't just an annoyance, this isn't just a headache for my purchasing department who has to deal with developers using their credit cards to get infrastructure, shadow IT, this is actually a point of differentiation for my company. If I can figure out how to deliver software faster, I can build competitive differentiation that's actually sustainable long term. So CIOs got on board. Pundits got down on private cloud. This was also at about the time that people said everything's moving to public cloud. There's no reason to run your own infrastructure. Why would you do that? That's commodity. Focus your differentiation, focus your engineering talent further up the stack where the application developers live. And this was also the time that the enterprise hardware and software vendors were gonna get focused further up the stack and support developer environments and support hyperscale public clouds. All right, so at this point it was time for OpenStack to get on with it and figure out what was gonna happen in the future. We had to make some choices and we realized that those choices were gonna have implications for how people use the software and interacted with the community. But this was a case where the community was gonna lead the way. The community was gonna make the difference. So if you, we're now getting close to the present day and last year was when we started to see some really interesting things happening. So these other projects that came into the OpenStack Foundation, these projects were very carefully curated and selected because they played an important role in open infrastructure. Now with one exception and that exception is Zool, these three projects, Coddic Containers, Starling X and Airship were new technologies that were contributed to the OpenStack Foundation to solve a specific set of problems. In case you don't know, I'll put in a plug for Zool here. Zool was the continuous integration environment, multi-repo gating that was built to help us build OpenStack. So the community created Zool as a way to solve a problem that there wasn't a good solution for out there. And what's happened is Zool has become a really great tool for large organizations who are trying to build software globally with big development teams. It's perfect for microservices the way so many applications are built today. This was also about the time that the community realized we need to be talking about open infrastructure not just OpenStack because there's other technologies that need to be a part of this. And we'll talk about that even more broadly. But these six areas were for purposes of just getting organized and focused were the areas that the community decided to focus on and the foundation. Another thing that happened is the community really embraced the notion of open collaboration, cross community collaboration. The focus was on the user. The focus was on what do we need this application to do? What do we need these infrastructure services to do to support the applications that are important to my business? And we don't care where the technology comes from. We don't care. We can have arguments about the licensing model. We don't care where they're hosted. We don't care if they're not hosted. We simply want the best technologies. And today when you come to an open infrastructure summit you don't just see five projects being talked about. You see 30 projects talked about. You see 40 projects talked about. And they get together and figure out how do we integrate these technologies to solve specific problems. The other thing that happened at this time is when OpenStack became more than VMs. The OpenStack started life as a VM. The compute instance you got with OpenStack, NOVA, was a VM. And with containers, the notion of running on bare metal is really attractive. Of course the issue with bare metal, if you're not running in a hosted public cloud environment, the challenge there is you have to really work at it as an infrastructure team to make sure that those bare metal instances you're providing to your developers are secure and fast and reliable. So this was at about the time that the Ironic project finally got to a level of maturity that you could say I'm gonna deploy Ironic and perhaps you might not deploy any other elements of OpenStack to run your cloud. Or to run the compute environment in your cloud. So OpenStack became about more than VMs. And the other thing, the four opens really came out as an organizing concept, as a philosophy to say this is how we're gonna run our community. When you participate in an Open Infrastructure community, these four opens are gonna guide how we do business together, how we interact, what's important to us. So as the calendar turns into 2019, you find that this Open Infrastructure community is massive, it's global, it's 100,000 plus members across 200 countries and 700 organizations. Another thing is you see a community that's achieved a measure of balance. And if you go back to 2015 and you were to go into a summit in 2015, you would see that the vast majority of the participants in that summit had non-technical roles. They might be marketing, they might be IT leadership, they might be executive leadership, but they weren't engineers, they weren't cloud architects, technical. Today, when you come to a summit, it's just the opposite. You see more technical people who understand how to design, build, and operate the software than you do people who have non-technical roles. And as you walk around and talk to vendors who participate in these events now, you hear consistently saying things like, you know, there may be fewer people here, but we're having more conversations that are of quality. We're talking to people who are actually trying to solve problems, and that actually makes it a more pleasant experience. So that balance is useful. You also can't deny, in 2019, open infrastructure means performance. Some of the biggest brands in the world are running software built on open infrastructure technology to support important critical business applications at massive scale. You see a community that is highly engaged. In the last release that pushed last month, the train release for OpenSec, there were more than 25,000 changes accepted. Now, what's important to notice, to note here, if you dig into that 25,500 number, you find that about one third of those changes were submitted by people who were new to open infrastructure and new to OpenStack. That's a healthy community because it's not just the same organizations contributing over and over. You have a balance of people who are committed to the software and continuing to improve it, as well as new people coming into the community who are contributing upstream. You see opportunity in open infrastructure. 7.7 billion dollar market is estimated by 451 research to be where we're gonna be in just a few years, products and services in OpenStack specifically. This doesn't count all of open infrastructure. If you add Kubernetes to this, it, the number almost doubles. Oh, and by the way, let me go back to that for a second. This 30% compound annual growth rate, and that's again a 451 research number, that's similar to, it's in about the same ballpark as you see what Amazon's growth rate, compound annual growth rate looks like now. Now, obviously it's a much bigger number than 7.7 billion dollars, but if you're looking at growth, opportunity is there. And this is a community that is delivering consistently landing releases on time with the features and services that the users need in order to go support those business applications we were talking about earlier. All right. So open infrastructure starts with OpenStack. It all started with the community, and I wanna talk a little bit more about that now. Okay, so, hang on just a second. Feel like, nope, sorry. All right. So where we are, and the journey that we're on now is very different than what we thought it was gonna be or very different than many of us thought it was gonna be, but we have this powerful community that's helping the software adapt and be applied in new ways, in ways that we didn't imagine when we started out nine years ago. The ability of the projects within open infrastructure to integrate new technologies is key to this. You see a global footprint and you see very interesting problems being solved at massive scale. These are a couple of places, if you wanna learn more about the Four Opens, you wanna get news updates about OpenStack and open infrastructure, those are some places you go. I hope that you have a great, have had a great week here in Shanghai. It's been an awesome summit and we'll see everybody next June in Vancouver, I hope. Thank you.