 That concludes our China spotlight, but next up we have Brian Stevens EVP and CTO from Red Hat So let's hear it for Brian. Come on up and let's hear what he has to say Welcome Brian everybody So I promise My daughter something so we got a bit of housekeeping. So if you can indulge me, please smile I've got a number of pictures sort of Standing here, but nothing of you guys you pretty good-looking audience Awesome. Thanks for that So Red Hat certainly a company that's been synonymous With open source, you know stands the reason we've got you know the better part of two decades Working in open source with the company was founded on the open source principles Not just from a code perspective But just sort of how we think and interact on a day-to-day basis inside of the company and with our customers You know what we've always said that you know never rest on your laurels And we always felt that as much as open source has been disruptive for our customers from the perspective of operating systems To us that just felt like the first part the very first chapter And so what we really think is that open source is such a defining development model that it's going to be disruptive across the whole IT space and I think that makes a lot of sense today when you actually talk to customers Surely talk about operating systems, but today the mission has gotten much harder It's actually how do they both manage and deploy not just infrastructure, but applications alike How do they do that at a level of scale beyond what they ever had to do in the past? How do they do that? Reliably how do they do that efficiently how do they do that nearly real-time without waiting on help desk tickets and Increasingly important is how do they do that repeatedly again and again and again? You know it's for those reasons why Red Hat is completely centered on open stack Not just because of the functionals the function the features that open stack is actually Focus on it's because behind open stack is you know looking around as one of the most vibrant development communities And for us, that's the best development model, you know in the world You know Red Hat's whole business model sort of thrives on developing software out and open in these open communities And while that sounds awesome It's also makes it very difficult because the traditional model of software development Means you can actually start to slow down the cadence of feature development and then you actually actually can go into a Qualification phase and hardening phase and then it ultimately bring that to customers I don't think the community is going to stand for Red Hat actually telling it to slow down because now we got to productize and make Some money so we've really sort of continued to hone inside of Red Hat You know our model our manufacturing floor that allows the innovation to flourish in the community But yet we can continuously take the technology inside of Red Hat Harden it make it work with our partners and then when we bring it to customers Not only does it have to work reliably we have to be able to sustain that upgrade over upgrade for many years to come And it's why that we usually don't talk about products inside of Red Hat We actually talk about subscriptions. That's the relationship that we want to have with our customers. It really is a continuum We're not providing them a point product We're providing them a continuum of both knowledge as well as listening to them so that when they actually have Directional statements of where they need the technology to go We can help influence those in the community and because of that That's why we have to be active in the community when you look at some of the stats for the recent releases of OpenStack You know Red Hat shows up, you know at the very top But that's not the yardstick that we measure our success on the yardstick that we use is can we then take that technology and Not just replace customer Proprietary technology in our customer environments. That's really not that interesting What's really interesting is can you allow customers and empower them to do things with open source that they never could do in the past You know for the last two plus years It's actually been fun to actually build a team, you know working on OpenStack, you know That's that's you know really the fun part of the job And I think we've been you know largely very successful at that We get a great group of engineers, you know represented here at the summit We also did something earlier this year because a vibrant development community is important But then you also want to get that technology as pervasively used as possible So we created a what we call a community distribution earlier this year called RDO And the goal of RDO is how do you take the software that's developed by thousands and get it quickly into the hands of millions? so RDO is really a Unopened version of OpenStack as it's developed nearly continuously upstream in package for enterprise Linux and today I think the users have shown that we've sort of serving them serving a need that hadn't been scratched in the past Because we're getting about a thousand unique downloads a day of RDO right now And then so the so the first leg is obviously development and communities the second leg is community distributions And then naturally the last part about that is building relationships with customers And as we look at that I think our initial thought was that of course you just want to bring OpenStack to Customers discreetly but what we found was that that the technology and the work that was going on on inside of OpenStack Was was creating new dependencies on Linux Unlike anything we'd ever seen before in the past three or four years Dependencies in the links kernel from things like virtual switching Namespaces and other containerization technologies and so instead of actually making OpenStack an a la carte subscription offering We said you know what we need a whole new version of Red Hat Enterprise Linux What we call Red Hat Enterprise Linux OSP or OpenStack platform that goes out together as a as a well integrated release That has all the advanced capabilities in the operating system as well as well that that OpenStack relies on You know so so technology is great, but but but you know customers don't just consume point technology from vendors They take technology from multiple vendors and expect it all to work really well today We learned that you know first and foremost with Red Hat Enterprise Linux, and it's a big part of why Rel is so successful isn't just because of the technology in rel is because of that it works with with nearly everything that a customer would like Our own mission statement, you know talks about Red Hat being a catalyst not just amongst the development communities It also talks about it's being a catalyst with our customers But also we in our mission statement talks about being a catalyst with our partners So we sort of get that as as job number one to make sure that we can take open source technology and make sure that it Works with with an ecosystem of thousands And so so what we're doing is actually taking a page right out of the playbook of Red Hat Enterprise Linux And now it applying it to OpenStack You know you know Red Hat's grown up in the in the past 20 years and you know in the very early days We're very feature-driven But now you're seeing us move towards really trying to understand customer problems and use cases So features have really become subservient to the problem that we're trying to solve And as we look in our in a customer environments of today the very heterogeneous very few customers are sort of greenfield It where they get to a clean, you know slate on how what they want to build Most customers that we see today are running virtualization typically VMware Increasingly rev and hyper V and they're also using both already using private cloud technologies as well as public cloud technologies like Amazon That was a big reason why we actually acquired a company called mange q and now named it cloud forms Is that we wanted a way that not come into a customer environment and say shut down everything that you're already doing And we're gonna bring you OpenStack instead what we needed to tell them is say we understand your current mission We understand you can't start from a clean You know a clean slate, but we're gonna actually bring you OpenStack But then we're also going to bring you cloud forms Which actually serves to knit all of their pre-existing virtualization environments as well as their public cloud environments together with OpenStack And it knits it together into a seamless open hybrid cloud that allows them to do you know Common common management on top from application deployment to resource management So you know a product strategy is great, but really customers aren't looking for point products from vendors anymore They're really looking for relationships That's why we had created a subscription model in the first place But any customer that you want to have a relationship with at a partnership level They're gonna expect a high degree of competency from you both in terms of competency of the current knowledge of code But as well as the ability to understand their problems and be able to reflect that back in the community that's why we actually You know participate so heavily in communities and speak so passionately about it But rather than me to take you through all that who better to tell you about how Red Hat has been working in the OpenStack community And our future plans and the very first engineer at Red Hat that put the first line of code inside of OpenStack Please welcome Mark McLaughlin Okay, good morning So I've been at Red Hat almost 10 years now and I still remember the words that persuaded me to join That Red Hat is 100% pure play open-source company But it was really only a few months later that this hit home for me We got the entire company together in Raleigh in North Carolina for an all-hands company meeting and on the first morning In a kind of dark packed auditorium like this one We showed a video for the first time the video is called truth happens and really when I left that session I was just incredibly pumped up and excited that you know open source is the fight for me and that Red Hat is the place to fight it from So this video means an awful lot to me personally But I think it's really relevant to where we are today with OpenStack. So let's see the video. Thanks very much I'm glad you like it So I guess the point of that video is that you know, we can have this righteous self-belief in what we're doing Even it when there are skeptics and doubters out there We can believe in the truth and inevitability of what we're doing So the video kind of represents a place and time for Linux back in 2004 And it's kind of hard to remember to think back, but you know back then Linux did have this incredible momentum Red Hat was out there proving its business model around Linux around open source and it was clear this was a real disruption happening happening in the industry But still it was still fashionable to be out there saying that Linux is just about hobbyists It'll never make it in the enterprise and it's just the hype And so in 2013 we see headlines like these about OpenStack and geeks like me we call this FUD fear uncertainty and doubt Sometimes it is just FUD sometimes it's those who are really challenged by what we're doing Silently trying to undermine our credibility But other times it's just a question of if you're not part of what we're doing If you're not signed up to this crusade that we're on if you haven't drunk the Kool-Aid Then it's kind of hard to believe in where we're going And this self-belief that I'm talking about it's it's not complacency right if you if you go along to design so much sessions this week It's not a bunch of people sitting around high-fiving each other You know, it's not a question of mission complete for OpenStack. We know we've a ton of work left to do Back in Portland during Brian's keynote. We showed a video called Faces of Grizzly And we had some interviews from from some of the contributors on the project And I can't tell you how proud I was to see you know Monty, Thierry, Russell and Chris and Doug Just really wearing their hearts on their sleeves talking about their excitement around their project They're you know real passion and commitment for the project It's these strength of our convictions and the strength of the community we've built that really gives us a unique opportunity in OpenStack But where does this strength come from? Well, I've been around open source for over ten years now And I spend a lot of time thinking about what makes a successful open-source company what makes a our open source project What makes a project, you know strong vibrant and diverse and It's really hard to define. It's no magic mix of ingredients that you put together and you're guaranteed success It's different for each project But I think in OpenStack, we've hit on a really interesting mix of approaches. We have open development and collaboration in spades You know, this isn't a project of cliques. It's not a project of throwing code over the wall or keeping code to yourself until it's perfect And maybe where that's most obvious is how often we see newcomers join the project and very quickly grow into be leaders on the project You know to make an impact on OpenStack. You just need to show up contribute be productive and constructive Make your voice heard and before long you'll be as much a part of the decision-making as anybody else And this is what governance is all about. I mean if you look at OpenStack governance structure You see a lot of you know, if we've got core review teams We've got PTLs the technical committee the user committee the foundation board of directors And it maybe it might seem like bureaucracy our process for the sake of process But really the whole point of it all is just a power in powering individuals to step up and be leaders on the project and it's really The individuals that make up this project is the magic of OpenStack It's it's us as individuals individuals with our diversity of interests, but our shared ambition for OpenStack That really makes OpenStack great But we're not just a project of individuals, you know, this is an open-source project with the kind of Corporate backing that you just don't see anywhere else Some people like to say that this is a mixed blessing for OpenStack, you know There's there's some assumption here that there is this tension between our Also as a community of individuals and our corporate backing But for me if this is a bad thing for OpenStack, then I don't know what the ideal is and I think back to when I was you know leaving university and You know, I'd really decided that I wanted to put as much time and energy and focus into open source as possible and Realistically the only way for me to do that was to work on it full-time Find a company that was willing to pay me to work on open source full-time And the corollary of that for me and why I think this is so great about OpenStack is that if you can bring together You know companies with shared interests shared goals for the projects and a willingness to fund lots of contributors on the projects, you know, then you're off to a great start And we've done that with OpenStack right we've brought together these companies and we've created the OpenStack foundation With the mission to protect promote and I think most importantly to empower the project and its community of individuals So we've built this strong community, but where's it going? Where is this going to go over the next few years? Well, we're all familiar with the the story of OpenStack's evolution today, right that OpenStack launched in 2010 when NASA and Rackspace Put out Nova and Swift, you know compute and object storage. These are the core infrastructure pieces in 2011 we added some missing pieces We added an image registry and identity service and a UI and then in 2012 Networking and block storage went out on their own as neutron and cinder And I think at this point for many people this was a nice tidy focused definition of OpenStack's scope and mission You know, there was this assumption for many people. I think that OpenStack would continue to focus exclusively on these core infrastructure pieces But for me that never made a whole ton of sense Firstly because these core infrastructure pieces aren't enough on their own to really build a viable cloud service And there's no clear bright line between the core infrastructure pieces and the other bits you need But most importantly if we built this community this awesome place for us to come together and collaborate You know, it's perfectly natural that we expand to start tackling other problems And so in 2013 we see two new projects have been added a kilometer and heat And I don't think you can really argue that these are core infrastructure pieces in the same way that our previous projects were But you know, who wants to build a cloud without monitoring and measuring and why would you not want an orchestration API in your cloud? So we have this You know actually I bring it back to the community all the time, right? I think about these projects Long been a part of OpenStack But more importantly that the developers working on the project that the teams Have long been a part of our community have built their projects from scratch like other OpenStack projects Share the same processes and tools same culture and They've even contributed to other OpenStack products. So given this community alignment and given this technical alignment How could we not be proud of really really proud of these projects and want to include them in our release? So for me when it came around to the technical committee Formalizing these projects as OpenStack integrated projects. It was a foregone conclusion These projects have long been a part an integral part of OpenStack. So yes This is an expansion of our scope But it's not growth for growth's sake For me, it's a very careful deliberate and measured group growth of our scope So what's coming next? Well, we know a few things about 2014 already We know that Trove the database service has graduated from incubation and will be integrated during the ice house release We know that ironic the bare metal service Marconi the queuing service and Savannah the Hadoop service will all be incubating during the the ice house release Now there's no guarantee that these projects will all Graduate after one release cycle, but if they did that will be four new projects added to OpenStack in 2014 And we know about other projects like Manila the file systems project Designate the DNS project are all making great progress all Collaborating together and are keen to be incubated And we've heard an awful lot about the newest project out there project solemn Which is a collaboration between rack space and red hats open shift and docker and others to build some paths capabilities on top of OpenStack services Who knows where that's going to end up in 2014? But for me personally the one that I'm really excited to see make progress in 2014 is is triple O So we first heard about triple O back in Portland when Monty Robert and the rest of the team at HP Started sharing their plans and their vision for triple O and things have moved really quickly since then Triple O has since been approved as the official OpenStack deployment program and Over the summer red hats started building a new service in this area called tusker Which has now been added to triple O and the two teams of joint forces For me, this is the big step forward that OpenStack is going to make in 2014 This is OpenStack the project taking upon itself the problem of deploying and managing OpenStack Now one thing about triple O is it is a little bit mind-bending And if it's going to be so important in 2014 I think it's worth taking just a couple of minutes to go over what what triple O is all about So triple O addresses a whole bunch of different really important topics But at its most basic the most basic thing that is trying to build as a first order of business is an OpenStack installer So what does this installer look like well one way of using it will be to put it on a USB stick like this So on this installer USB disk you've got all the familiar OpenStack services But you've also got Cilometer heat and and this new service tusker If you want to install a few racks of machines The first thing you need to do is pick one of those machines as your as your control plane as as the machine You're going to use to manage your OpenStack deployment You boot your USB stick and Immediately you have a fully running OpenStack cloud running just on this machine But we called that cloud and under cloud And the big difference with this under cloud is that it's managing bare metal machines rather than virtual machines If you talk to Nova in this under cloud and ask it for an instance you get back a bare metal machine And I think the crucial insight here is you know We have this service called heat which is there from managing and deploying multi-tier fallout scale out Multi-tier Multi-tier scale out fault tolerant applications, but what you know OpenStack is a multi-tier fault tolerant scale out application So why not use OpenStack or why not use heat to describe and deploy OpenStack So you want to deploy OpenStack you have an under cloud running, but the under cloud needs to know about your physical infrastructure It needs to know about the bare metal machines and the network topology Once it does that You're ready to deploy your your your OpenStack cloud So you talk to tusker in your in your under cloud and you ask it to deploy It builds this heat representation this heat template of Watcher OpenStack cloud is going to look like and then it starts talking to Nova to ask Nova to actually deploy the machines Next thing that happens is perhaps Nova might just decide to deploy the two controller nodes in one of the racks first so what happens here is Nova pixie boots these machines and Provisions some images from glance in your under cloud to these bare metal machines It then moves on and might deploy some compute nodes in the next rack And some storage nodes in the next rack And at the end of this process you now have a fully functioning multi-rack highly available production cloud And we call this cloud your over cloud. So you've got your under cloud which is used to deploy your over cloud So this isn't some far-fetched story. I'm telling you here If you go to the red hat boost today and have a look at the rack of machines that we have there We have a demo of deploying audio our community distribution on these machines using triple o So it works And I think the interesting thing that the or the MO or the or do demo starts to show is How flexible an architecture triple o is? So in this demo, we're deploying or do using standard or pms using standard pubic recipes We're using this USB installer we Were using a particular Deployment architecture, but this isn't all of triple o right that's not Triple o isn't prescriptive. It's intended to be flexible So you can actually use the triple o framework to do all sorts of different types of open stack deployments And I think the key thing is that we're trying to create We're trying to build triple o into this place where people can come to get together Where we all can come together to collaborate and not only an open stack installer Not only integration between hardware and software vendors But things like upgrades continuous deployment reference architectures blueprints Sharing operations knowledge. This is always in the scope of triple o. This is Triple o is the place where we can all come together and collaborate on these things So I'm really excited about triple o and I think it's going to be that the next big leap forward for open stack So we're down in this triple o rabbit hole now, so let's pull ourselves back out of there back to 2014 But what about beyond 2014? So I started with a video from Linux in in 2004 What about open stack 10 years from now? What about open stack in 2024? Does anyone really think we're going to add a handful of new projects in 2014 and then stop that this is now the scope of open stack and we're not going to expand beyond that I Really don't think that's realistic. I think open stack is going to continue to expand And it's going to become quite a broad umbrella of loosely coupled projects. I Think we need to get real about what this means for open stack and we to embrace the collaboration that's happening under this open stack umbrella We need to evolve our culture our governance and our processes as needed to handle this expansion But again, I'll say it again. We need to really make sure that this is a careful and measured expansion of our scope We need to prove wrong all of those who sees this expansion as somehow a distraction from our work on core infrastructure And so I talked earlier about the strength of our convictions and for me I often go back to our mission statement to remind me why we're all here So I really do think it hits on many of the key points Talks about ubiquity. It talks about an open source cloud Talks about public and private clouds large and small Massively scalable, but back to triple o again simple to implement So we've empowered our community with this mission statement And we will continue to deliver on it year after year Because we know we've built this great community this great place for us all to collaborate and back to the video again despite everything Truth happens So on behalf of myself and on behalf of red at thanks for listening here listening Thanks for being here and thanks for being the magic that is open stack Thank you mark and Brian that was a great great talk so