 here from Sargillai, who's Senior Vice President General Manager of Converged Cloud at Hewlett-Packard, and someone else that many of you probably know, Brian Aker. He's an HP Fellow. I'm not sure what that means, but it sounds pretty cool. I'd like to be a Fellow. I don't know, but it sounds pretty awesome. All I know is he's worked on a few other projects in the past that have done pretty well. He was one of the folks that actually helped create MySQL and a little website called slashtot you may have heard of, so he's done a few cool things. I think it's great that he's helping out to make OpenStack successful. So let's bring him on out. Good morning. Wasn't that NSA thing cool? I don't know, it's going to be, how do you come up with that, you know? I once had to go after DreamWorks when they were doing a 3D movie, and I got through that okay. So how many of you guys were at the party, our party last night? We had a big party, you know? We spent a lot of money here. Yeah, that's right. That's right. Thank you. Thank you. Hopefully you recovered enough. You know, we didn't want to be the first ones up because we wanted you guys to recover by the time we show up, so. Anyhow, Brian and I are going to talk a bit about for the next 20 minutes or so how we're exhilarating in cloud with innovation from OpenStack. However, before that, I want to talk to you a bit about my personal journey into the cloud. Some of you who may have worked with me before know that I'm actually a networking person. I've been doing networking for about 20 years. I got into HP about three years ago when HP acquired 3Com. And at that time, I became the CTO of networking at HP. So I'm actually an engineer. I know that's shocking, and some of my people argue with me about that, but it is true. Ask him about SSID sometime. It's actually fascinating. Yeah, if some things fall apart here in the presentation, you'll realize I'm an engineer. I'm a working person. And if you're using any of the wireless here instead of paying attention to what I'm saying, that's if it's working, it's my stuff. It's not working at somebody else. So I got into network. I was working on networking and I came to HP. And as you know, HP is one of the largest enterprise companies providing data centers and solutions for data centers for most of the Fortune 500 companies, Fortune 1000 actually. And we were looking at how to build networking for clouds. And I was digging around, and I noticed that we had this open flow protocol that we've been working on since 2008, but we didn't want to tell anybody about it because we're the quiet company. And then we were looking at different things we can do. And we started dealing with SDN. I was working with the cloud services guys. We're trying to see what to do with SDN. Other people in the industry were working on SDN. And so more and more, I got involved in cloud. In fact, I was consumed by it because it was really interesting problems. And I like interesting problems. And so eventually, they basically said, why don't you just come and run the cloud piece? You know, networking is fine. Run the cloud piece. And on that particular day, I wasn't really thinking about life balance. And so I said, okay. And so that's how I ended up in the cloud. And the funny part is that for those of you who know networking history, we've always represented the network as a cloud for the last 20 years. Diagrams always kind of looked like a cloud anyway. So I guess I've been doing this for 20 years. I just didn't know what it was. So that's, you know, that's how I ended up in cloud computing. And you know, it is, as you guys know, it's the most exciting thing happening. OpenStack is very exciting. So it's actually a great move. So who's HB in cloud? Why do you care? Well, first of all, we throw great parties. Is that not true? I mean, come on, every, you can come to the OpenStack convention without the HB party. Okay. This is recorded. I could get in trouble. But that's fine. We also have about 200 people here from HB. Thank you very much. And some of these folks have actually contributed code as well, I hear. And some of code, like this guy. But on a more serious note, HB actually has a very long history of supporting critical open source initiatives in the industry. You know, for some of you, you guys remember SCO, remember those guys? 2003, right? Linux was just making its way out. Things were happening. SCO starts to sue everybody that's doing Linux and tell them not to do it. And people are still concerned. Now it seems obvious that Linux is the way to go. But people are still concerned. It's not clear. You know, newspaper, all the press is writing what's going to happen. HB comes along and says, hey, you can use Linux, we're going to demify you. It doesn't matter if you're getting it from Red Hat or from Suisse or whatever, we are going to push Linux into the enterprise and we will demify you. Forget about SCO. And for ones of you who remember, this was actually a huge deal at the time because we had this point of inflection where Linux was actually being adopted at a really heavy rate. Now, probably yeah, sure. Many of us had been using it at that point for many years. But at that point, we saw the industry actually suddenly adopting it and replacing proprietary unixes with it. So here was at this point where it could have been shot down and any proprietary, you know, unixes could have like perhaps searched forward. But suddenly you see HP coming in and saying, like, we don't care if it's actually Red Hat or Suisse or whatever, we're just going to broadly indemnify you. And this was done by Martin Fink, who's the current CTO at HP. So this was a huge deal at the time. And it actually meant something where HP was in the same way when it looks at OpenStack, looking a little bit into the future and trying to figure out exactly what bet should actually be made. Which because what did it actually lead to for HP? Well, the good news is this really helped the industry and it also helped HPE. In fact, today we ship more Linux servers than anybody else in the world. Okay. In fact, during this presentation, we're probably going to ship at least 30 of them. Many of them probably to some of you guys. So it was a good bet for HP. And it was a great bet for the industry. Because, you know, we believe that when you have these communities building open source projects, you get the faster innovation. So moving on to OpenStack, in 2011, we were looking at HP as what should be our next generation platform for cloud. And we were looking at some internal solutions and some external solutions. There was a bunch of different things going on in the industry. But we zeroed in on OpenStack because it had a very interesting, extensible architecture was being run by Rackspace. There was some very interesting things going on with it. And we decided that this is something that probably could become with the right support and the right ecosystem, the right community. This could become the Linux of cloud. It could become the platform. But in order to make that platform successful, it needs to have an ecosystem. So the first thing we did is we said, you know what, we're going to put our whole might behind this and we launched OpenStack Public Cloud to show the industry that this was not just one vendor, this was a whole ecosystem. Which you think about it, at the time we had a number of open source projects that were actually cloud pieces. But all of them were generally backed by what? One vendor or by essentially no one. And at that time having Rackspace had done the actually creating OpenStack initially. But to really be an ecosystem, to really be something that was just beyond one player, somebody else also had to adopt it. And this is where 2011 I was actually at a conference where the HP people had approached me and were like, no, no, we're actually going to do OpenStack. I'm like, really? You're really going to do OpenStack? And they're like, here, look at the secret website that we've been actually building internally. And so they had made that bet and we're trying to figure out how could we actually make this more successful. And really, in the ecosystem, we have to have multiple players. Which as we know today, compared to 2011, HP stood up and actually started working with Rackspace. But this came out, so where we then moved to 2012 we are. So 2012, working with folks like Rackspace and others, we said, well, how do we take OpenStack to the next level? You know, looking at our experience in Linux, looking at our experience, making sure that you can build good community projects around open source. What is required? What kind of framework is required? Because you know, once you start all these big companies getting involved, if you don't have it set really well, it can get messy. Everybody thinks they own everything. So working with Rackspace and others, we created the OpenStack Foundation. And one of our biggest contributions there is, you know, we have a lot of lawyers at HP, we're a big company, we have lawyers. And so we actually spent a lot of the time helping write the charter. Thank you, Eileen Evans, for the foundation that put it in such a framework that it would make it easy to have this work the way it works today and help OpenStack grow and become what it is. So how does HP help the ecosystem? Besides throwing parties? Guys, again, it's important. You know, first of all, you know, the OpenStack ecosystem is a big thing. It's partners, customers, vendors, designers and so on. So there's a lot of work that has to be done there. But first of all, it starts with people. And we have a lot of people involved in OpenStack. In fact, I believe we're like... About fourth, Ken's largest contributor right now to Grizzly. Fourth and Grizzly. Not too bad. But as you guys know, there's a lot more to making OpenStack work than just contributing code. Is that true, Brian? That is true. So first of all, let's look at actually one of our big contributions. This is something that I personally have interested in for like five plus years and coming out of my experience with MySQL. In fact, it comes as my experience with MySQL when one particular release in about 2003 where someone's like, yeah, we launched. We put those software up on the site. I was like, did you check this one platform? And they're like, no, we just tar-balled it up. That's a problem. So I became interested in actually how do we make continuous integration actually works. And what you're seeing right here is a graph of actually looking at... It's truly actually innovation. So within inside the Grizzly period, thanks to a member, like, 12 member CI CD team. Yeah, only 12. I was told there were more. We can hire more. I'm sure Monty, if you're out here, right? Somewhere out there? Yeah, you can... I think that's a green light. We can hire more people. So, I know it's for people on stage. I don't read emails. So what you're seeing here is 7,000 successful commits within inside the Grizzly period. And that right there, look at that's innovation. That's more people coming and actually working on OpenStack. This is out of also, by the way, a total of 21,000 plus actual commits that were attempted. That's good news. That means some stuff didn't actually pass. And that's great. And as we're accustomed more to, like, gating trunk. And, okay, so why do we gait trunk? So hardened security. Think about this. How many people go and review code before it actually ends up in the trunk? That's awesome. How many people, like, also everything in the trunk is being tested? We're being installed. It's getting bigger. Notice, by the way, I said CI and CD. Continuous deployment. The piece that we're moving forward on. We're not only trying to test, you know, did the smoke test did work, but did it actually deploy correctly? Because the only way for the software to have so many people actually contribute to the software is to make sure that what, it actually always worked. The trunk is always gated. The source code always works. I don't know how many of you are, your experience with other open source projects over the years, but, like, within OpenStack and with, you know, the software that essentially came from, you know, well, came from different pieces to actually form OpenStack. That right there, that's a new concept within OpenSource. For the ones that you haven't really done much in other projects, you're probably taking for granted this, like, nice infrastructure system that's doing all this testing and make sure is that people are reviewing things and all those other things. And so that right there gets us into an operational enhancement. We know this stuff can actually be eventually upgraded. We know this stuff actually all works together. All of these are, like, very, very core elements in what really is required to create another stack. So where does HP actually also involve itself? Triple O. This is something that NtDocomo began in many ways in beginning to actually do the bare metal install components that started off in Grizzly. And then we came and actually added more developers to work on things like how do we build the image tool? How do we make these pieces actually function from day one as a stack, not just as a separate set of components that you install on whatever you happen to do, but in actual as a stack because the key piece of this is that this is the replacement for LAMP. You know, how many years we actually sat around with people trying to, like, figure out different letters and, like, go, oh, maybe replace the L or maybe replace the M or maybe replace this piece. And that was really kind of what people, like, focused on, which was unfortunate because the real key was actually to replace a stack from end to end, actually building that piece. And, like, triple O and actually doing the bare metal install is what begins that piece. So it takes a bunch of the different software and starts building the actual ecosystem. And there's great talks this week about triple O, about heat. This is the bare bones kind of beginning orchestration. How do we orchestrate bare metal pieces together so that we can do something more than just install on a single computer? I mean, people in their audience have actually, like, installed DevStack at this point. DevStack is awesome. It gets people to the hello world of beginning to do things. But how many people have tried to take DevStack and then said, well, could we just make it work a little better and somehow get it to work across multiple machines and all that? DevStack is a beautiful hello world. But it's not that next piece. That next piece has to be written so that we can actually, and it has to be built in to open stack. Because otherwise, it's just an assorted pieces of parts that really are not going to go together as a particular ecosystem. There's more talks on heat. I think there's another one today. Like, you can go in here and people talk about this a little more intelligently when I'm going to in this short a period of time. But this stuff is really neat. And it really is starting to become the glue and the fabric that ties all this stuff together. So what do we do beyond that? Like, we're not just actually working on pieces that where we're either leading or, well, and we're a member of. But we're actually still working across the entire board. So things like NOVA affinity scheduler, we worked on this piece. Sender backup we did. We actually increased the tempest test by 28%. This is something that's kind of unique to a fairly large company is that we can actually spend time and developers and realize that this is a very big bet or a very long period of time. And the thing that I've actually preached inside of HP is that the best work we can often do is actually not even on the feature side, but is actually making sure that like we can provide developers for like things like quality of experience and so forth. Something that is something that a large company can actually spend time on and resources. So a lot of our work has to go in those areas. But we're trying to like, you know, we end up end up touching lots of pieces of it and so forth. This is a slide I actually did at a similar conference here about a year ago in this conference where I started talking about a little bit about like a bigger view of like where OpenStack can be, perhaps, because we see in the beginning, you know, Rackspace was smart enough to realize, okay, so we have file system. We need something for actually for compute. So we see Nova glance with kick in. And from there, what do we see? Brian, what is that networking called something else once? No. Yeah, yes, it was that. Yeah, there was some kind of a quantum event that occurred. Something happened. As you know, everybody has to go through the slides and go, okay, Brian, you didn't say, okay, no, we didn't say the word. Good. I'm people who even understand that joke here. Okay, well, so much. Well, educational for some people probably like have to be so we'll have another one. We'll see if you can find it. So anyway, these are all these pieces we see. And this is the beginning of this is the continuation of the ecosystem. So how does the stack grow? And this is the next slide that I had actually done time pointing out that like really as we go forward, like there are so many more pieces that have to be done. So like little factoid, HP public cloud, we counted to figure out how many discrete systems it requires to launch a public cloud today. For us, it's 32 systems that actually have to be done like we can identify 32 independent systems that we actually need to launch a public cloud. And so some of these are some of the pieces you'll see in there. Things like, you know, what else do you need? Well, you need blocks storage systems. OK, good sender. We need billing systems. We need alarming. We need messaging. This is just some of the pieces. There's so much space in this ecosystem right now going forward for other components. So what else? What are we doing in this? Moniker DNS of service. This is entirely open source. You can go download it, use it to your heart's content. Open source DNS is a service piece. We launched it with public cloud in February. And it's now been available for, well, let's see. It's actually been open source for a little bit before that. And I think we're heading down the track of trying to get this to be included directly in open source, sorry, an open stack. But to run a cloud, what's the key component you need? You actually need DNS as a service. Absolutely required. Other things we've been working on? Relational database as a service. This was a project that we started by a rack space a while ago called Red Dwarf, which probably also will eventually. We're going to have a problem with that. Yeah, that's really eventually going to be a quantum problem. It's probably the last time you hear that name. Yeah, not like the word ever got used for anything out there. So, or far out there. Relational database as a service. So this is actually providing, yeah, somebody got the joke, great. So anyway, so like this service is actually providing databases to the service because really in the end of the day, how many people want to like be bothered to like, did I back up the database? Did I like make sure, oh, did replication fail? Oh, replication failed. What should I be doing next? All of these pieces that for the most part, people don't really want to manage this stuff. They want it managed for them. So we're also trying to look at like, what are the other pieces that are coming along that need to be managed that really is just speed bumps as regards to most developers to find that path towards writing application. So this is another service that we've actually worked on. And this is, we pretty much work on, I have pretty much hand in hand with a rack space to get this done. And it's going to be great. So, see how many more people we can get to actually add to the developer base for it. Then we have other things like cloud systems, which is... So one of the themes of this conference has been getting OpenStack to users. And what better way to get it to users than actually to get it on a platform that users already have? So for those of you who are familiar with cloud system, cloud systems is HP's market leading pre-integrated cloud solution for both enterprises and service providers. We've deployed it with other 1,000 customers. Most people who run private cloud pre-integrated our running cloud system. And so these are normal, typical customers. This is not only the leading edge. This is your Fortune 1000 companies who are running cloud. This is what they're using. And so in this conference we announced Cloud System 7.2, and guess what? 7.2 has OpenStack on it. So all these 1,000 customers in the enterprise can now go and use OpenStack. They don't have to download it. They don't have to configure it. They can just go and use it on 7.2. Think about this for a second. There's what, more than 1,000 paid customers. These are large enterprises that are actually using the software today. And as their upgrade process happens, they're suddenly going to actually have access to Nova Keystone and Glance. That's a really powerful concept. So this isn't like OpenStack having to push itself into an environment. This is OpenStack being pulled into an existing environment already, where suddenly this whole set of new tools and abilities will be enabled that's based on OpenStack. And this is something that we think about as much different than, say, us launching a public cloud or a group going into a company and saying, well, we'll install OpenStack for you. This is actually existing environments that will have it already just pulled directly in there. That's an audience that probably isn't sitting here at the moment. They'll be here next year or the year after that as they start pulling in. But that's a pretty fascinating thing to actually watch. This is actually some of the future we're going to see around, actually, OpenStack, as it kind of goes into different areas. So as you can tell, what Brian's talked about and in terms of how we're putting this and some of the platforms we're shipping, we're very focused on OpenStack and getting OpenStack to the mainstream of enterprise. When we talk about OpenStack, we're not only talking about public cloud. We're talking about cloud period. We're talking about private cloud, managed cloud, and a seamless operation between them, which is really the promise of what we call Converts Cloud. We believe with OpenStack, we can unshackle the industry from proprietary systems that are stifling innovation. We know that with OpenSource products, if you have the right community and the right support, innovation will move much faster. And that is what we're betting on. We are betting on full on OpenStack. So we talked a bit about some of the products we've released, talked about public cloud, talked about HB Cloud System 7.2. But as you can see, we have much larger plans and we have a lot of things that we're working on. In fact, I can tell you that over the next 12 months or so, we are going to be putting out OpenStack on all our key platforms. Remember, we ship one of the next server a minute. Think about that. We also have enough marketing people that can go in and edit our slides and we don't notice. Yeah, I forgot to mention that. There's things we couldn't put, but they're coming later. But if you go to our booth, they can say things that we can't say up here because it's recorded. So go to the booth and see. That engineers tend to run their mouths, so just start asking. And some of the things that you can see in there, for instance, like the cloud messaging. So that right there is pretty fascinating. So we had an internal project doing messaging. And then suddenly what we see, we saw Marconi come out. So we actually were able to shift internally pretty much quickly from actually an API that we had been working on looking at to going, huh, how can we actually make this with Marconi? So as of this week, you can actually go to the HP Cloud site and use messaging, which is currently the draft Marconi API. So the API that we're looking for, for like having OpenStack actually have a messaging, RESTful API. That's actually what is there at the moment. We'll be shifting it as required to make sure it matches the Marconi API. The source code's a little problematic because unfortunately developers who wrote it were actually all Java developers. So that's not really gonna go into OpenStack very well. Though they did told me that they were like, okay, Brian, we will actually learn Python. If we can open source this, we'll learn Python. Like, excellent, that's great. So anyway, and you can see the other things, like we talked about DNS, like monitoring pieces, which is where we're trying to figure out how to tie into what OpenStack is doing and the other things we talked about. CDN is actually open source, something nobody actually talks about. But the CDN software itself is, and obviously Compute is Nova and the rest. Notice there's actually already drivers for things like 3PAR that are being built for OpenStack. That's pretty fascinating. Load balancer right there, code underneath that, it's actually a Python-based implementation. The upper layer is Atlas, so it's the Java API component that Rackspace had done. But underneath that whole thing, I forget the name of the project off the top of my head, that load balancer is built via HAProxy plus some additional components. That whole layer right there, that Python layer, all of that's open source. And we're actually looking at trying to figure out how to like take that component and place that underneath the not-to-be-named Quantum project going on to the future. So really, all of this stuff is pretty much open source and all of it is trying to follow exactly what OpenStack is doing. So we're trying to ride the level of the ecosystem because, well, as anybody realizes, like this OpenStack continues to grow, that ecosystem is gonna become fascinating. And you're gonna have a lot more people trying, a lot more people actually working to build projects and solutions and everything else within it. So to summarize, two years ago, we made a bet. We made a bet on OpenStack as the future platform for cloud and with your help, it's becoming a reality. As communities help, this is becoming a reality. This is it, this is happening. This is the platform for cloud. It will be open source, it will be innovation. It's not gonna be proprietary and it will unleash innovation on cloud and the industry. And so we have big plans and we appreciate your support in pushing them forward. Thank you very much. Thank you very much. Thanks, guys. So next up.