 Good morning. How's everybody doing? Awesome. Thanks for joining us today for IBM Client Day. Who is here for the first time at OpenStack Summit? Wow, that is awesome. It's really neat to see, we've been involved in OpenStack since the beginning, and it's really neat to see from Summit to Summit the growth in attendees, and then the growth in folks who have had no history with OpenStack to date. So I'm Jesse Proudman. I'm the CTO of BlueBox, which is IBM's OpenStack hosted product. I like to start kind of every conversation about IBM Cloud with this slide. This slide talks about the IBM Cloud strategy. You've probably seen it around the venue over the last couple of days. And it was this slide that really made me interested in IBM about a year ago when they acquired BlueBox. So if you think about the OpenStack marketplace, if you think about cloud computing in general, at the end of the day, everybody in the room is trying to deliver solutions. They're trying to deliver solutions to their customers or to their internal users. And those solutions need to be built on some kind of software. That software can be built internally. Many companies are doing that, but bought as software as a service. But that software needs a platform to operate on. And so at IBM, that platform is BlueMix, and BlueMix is comprised of containers, cloud foundry, and a bunch of services that can be used to build composable applications. But that platform needs an infrastructure to run on. And at IBM Cloud, that infrastructure that we've selected for the future state of our product line is OpenStack. So that's not necessarily novel. We all kind of recognize that pattern and trend. The OpenStack piece for IBM is a great foundation. But what's really unique about the IBM strategy is this notion of locality. So public cloud obviously is a dominant force in the industry. And we have public BlueMix. We'll have public OpenStack capabilities. We have those in beta today. So that's not particularly unique. But what is interesting to me, and what was interesting a year ago, was this notion of a consistent experience across public, dedicated, and local, and delivering that as a service instead of as software. So what do I mean by dedicated and local? Dedicated is a single tenant hosted installation running in software. So we have 25 plus data centers around the globe that we can install this stack dedicated to a single customer in. And then local means in a customer's data center or site of choice. And so the key here is that both that dedicated and that local option is delivered as a service. We're getting rid of this notion of shipping software and we're trying to bring that experience that you get from a public cloud perspective into the world of private cloud, of your data center. So that's what's particularly interesting to me. And if you look at this space, I think there have been a lot of entrance in the last year and a half, and there's been a lot of interest in this notion of consumption as a service. And it's a great way for organizations who are very interested in OpenStack and who want to get the power and benefit of OpenStack to try it or to interface with it without often dealing with a lot of the struggle and the upgrade cycles and the technology sort of debt that exists with the OpenStack platform itself. So that said, we have IBM client day today. We have a bunch of our distinguished engineers and best folks here today to share a bunch of content with you. We're going to start this morning with OpenStack for beginners, which for the audience in the room sounds like it's a great place to start. Then we'll move on to talk about the OpenCloud. So Jason McGee will get into more detail on that public dedicated local strategy and how those all work together to create a seamless experience for you. We'll move on to talking about IBM BlueBox, so the dedicated local OpenStack product working with CloudSoft, which is an application deployment tool set. So doing some conversation around how that operates on the platform and how it utilizes some of the capabilities with the software private network to do multi-data center configurations. Mark Shuttleworth will be joining us this afternoon to talk about Linux 1 and Ubuntu. So that's the mainframe that runs Linux. Then we've got Kyle Messery, who'll be talking about one network to rule them all. So this is OVS, OVN, and Courier. How do we get integrated networking across bare metal VMs and containers within OpenStack itself, which I think is one of the key drivers of folks consuming OpenStack. We've talked a lot about containers. You've talked a lot about bare metal. All those pieces exist. Now, how do we get them operating in a consistent network domain, which is a little bit harder than one may think? Then it's 4.20. We'll be doing a networking event in here. We've got a raffle for an iPad and some posters that were done by the IBM Design Center here in Austin. They're pretty awesome posters. So there are little tickets that you can fill out and drop in a bowl at the back. And if you haven't done that yet, you can go grab one of the tickets. They'll be by the door and put that in. So we'll draw that ticket at 4.20. We also have beers and soda in the room. So I think it's the coffee break in general. There'll be coffee in the hall, beer and sodas in here. So come join us for that. And then we'll end the day talking about urban code, deploy, and its integration of heat and patterns on top of OpenStack. So again, 4.20, come join us for the posters and the beers. And now I will turn it over to Tyler to talk about OpenStack for beginners. Thanks, Jesse. So when you look at OpenStack overall, it seems really complex when you're new to it. There's all these projects, there's different working groups and the foundation and companies and distributions, and it seems really unwieldy. But when you really start to look at the pieces, it starts to make a lot more sense. So the basic level, OpenStack is a cloud operating system. The idea is it's a layer, a control layer that you can use to integrate and deliver components, whether it's compute storage, networking, now containers, DNS, key management, all sorts of other projects, but it really comes down to just supplying them as a service in an easy way to consume. So it's all open source, obviously. It's all focused on providing particular cloud services. Currently there are 54 official projects, which is a change last year. There was a governance change to a thing called BigTent, and that's when that changed. So now there's 54 official projects. All of these projects or services communicate via APIs. So it was a key design decision right from the beginning, was we can't use back channel ways of communicating. They have to do it via the APIs the only way it's going to scale. It was originally started between a collaboration between Rackspace and NASA on the first two projects were the compute and object storage projects just about six years ago. And to go from basically two projects and two companies six years ago to 54 now is pretty impressive. So not too long after it was started, there's concerns about it only be two organizations running the whole project. So the foundation was created. So that's now the OpenStack foundation is a totally separate entity that is a nonprofit that operates and manages the OpenStack community. And it's written, this used to say all. Now it's predominantly in Python. I think this is a good way to set kind of table two is what isn't OpenStack. So it's not a hypervisor. What are you using for virtualization? Oh, we're using OpenStack. Well, okay, that's the control layer, but it supports a wide number of hypervisors including VMware vSphere, Hyper-V, KVM, Zen, and on and on and on. It's not a free VMware replacement. This is where I see companies getting themselves into trouble. They're like, hey, we pay a lot of money for VMware licenses. Can we just do this free OpenStack thing instead? It usually doesn't end well. And it's more similar to AWS than VMware. So the idea that the instances are mostly ephemeral is that we can delete them without worry because we're scaling them up and down and you're not concerned about any one individual instance. It's also not a product in and of itself. So if you want to use OpenStack, how are you going to get it? If you're going to get it through a service provider like BlueBox, a distribution provider like Morantis or Canonical, and it's not a single distribution there either. So there's a pretty large number of distributions and ways to consume OpenStack. And also, you can just get the raw bits directly off GitHub if you want as well. And it's not a single open-source project. So that's when people, oh, we're using OpenStack. Well, which of the 54 projects are you using? Because you don't have to use them all. It sounds daunting that there's 54 projects, but you can pick and choose to put together the solution that makes sense for you if you want containers or you don't, or you want object storage or you don't. It's not a storage platform. It's a network virtualization in and of itself. So the components that do the network virtualization use plugins to support things like OVS and OVN or commercial products like from Juniper and Cisco. But in and of itself, it's not a network virtualization platform. So why does it matter? Why are we all here this week and come together every six months? It's one of the only open-source projects that provides the full comprehensive cloud services framework. So not just VMs, not just some storage. It's all these other services. Some pretty significant cloud service providers are using OpenStack as their platform. And what we really see this is this is the next generation after Linux is OpenStack will have that really big impact of bringing open-source to customer environments. This is a Google Trends that I clipped not too long ago, but this is just basically search on all those different open-source cloud options. And you can see OpenStack starting in 2010. It's now pretty much taken over the space. So how did we get here? As I mentioned, 2010 started with NASA and Rackspace. In 2011, the very first public cloud launches built on OpenStack in 2012. That's when the foundation was created. 2013, Grizzly. So if you're not familiar, each of the releases get a name and it's alphabetical. So the current release that we're working on now is N, which is Newton. So the Grizzly release added block storage and networking. 2014, the OpenStack marketplace opens. So if you go to the foundation's website, you'll see there's a lot of full information on all the different ways you can get OpenStack. The big change last year, as I mentioned, was going to the Big Tent model. So there used to be a totally different model for bringing new projects into OpenStack. So we switched this new model last year, and it's been pretty successful so far. Last year also, from a size perspective, CERN announced that their OpenStack cloud hit 150,000 cores. So this year, we just had the Mitaka release just came out just before we got here. So Shamal's going to come up and talk more about the community aspects of OpenStack. Thank you, Tyler. So thanks for that overview and history. So OpenStack, as Tyler explained, a little bit of what its purpose is, what it's used for, what it's not potentially used for, and misconceptions, if you will. But really, OpenStack itself is part technology and part community. A lot of people that participate in OpenStack are here for the community aspect of it. You'll find very helpful people, Tyler alluded to working groups before, and whatever your interest is, whether you're an enterprise customer, a telco customer, you're in academic, scientific research areas, etc., you'll find people with that interest to collaborate with and work together to achieve whatever the use case is that you have. So some stats on the community itself, just phenomenal in terms of growth and size at this point. There's over 589 companies based on the last annual report, and over 170 members are from over 170 countries. There's about 34,000 individual members in the foundation. And to give you an perspective, this number at the last annual report was somewhere around 30. This number is from the annual report that just came out, and I believe I heard Jonathan at one point say that we're closer to probably 40 or 50 even at this point. So very tremendous growth rate. Massive developer community, over 5,600 developers, and by the way, this is Lifetime, so this is not disreleased, this is overall Lifetime of the project, and over 260 commits. So these are not lines of code, these are actually commits of code. So very massive scale from a community perspective as well, and the community is very helpful to each other. From a release perspective, as Tyler mentioned, generally the releases are named in alphabetical order. The naming scheme is based on the city where the design summit happened. So generally when you pick a place, like Mitaka for example, was when the Tokyo design summit was happening, so you pick a name that represents the location where the software was built or designed. And there's generally two major releases every year. One of the other changes that's happened recently is projects don't necessarily have to release every six months, so while bulk of the projects do release every six months, some projects have the capabilities to release more frequently. So the newer projects do actually tend to release faster, but then the more established projects in the course, the services that have been consumed the most and adopted the most, do tend to follow the six month release cadence. The other important thing to note is the support structure within the community. The community generally supports up to N-1 from a community perspective. Vendors and product developers, they can actually support greater than the service distributions and companies out there that'll support you, you know, still on Juno or Kilo. But from a community perspective, once a release is cut, there's an EOL date established with it. So for example, Mitaka just came out, the EOL is TBD. The previous release was Liberty and it's going to be End of Life in November when the next release, Newton, comes out. But what this also means is for the first six months, you can actually put in critical bug fixes as long as they're back ported from, like you have to commit them to a new branch first and then back port them. Or if there's security fixes, they can actually go more than 12 months even because security obviously, if there's a vulnerability, we have to address it. Now, I'm not going to go over all the different services, whether it's block storage, compute, et cetera. But what I will say is for the most part, most open stack services kind of follow this architecture. There's an API layer that generally provides, like, create, read, update, delete type functions against the infrastructure resource that you're pooling. Below that, you have something that schedules and decides, okay, which actual host or storage provider is going to fulfill the request that just came in. And then below that, you have the service which is interfacing with the infrastructure itself. And then usually that's through a form of a driver or plug-in and there's a provider. And as Tyler mentioned, between services, it's usually RESTful APIs. We use a message queue and we send, like, remote procedure calls over the message queue more or less. And so even in the service layer, they try to decouple each of these functions as much as possible so you can scale them individually. And as well as between services, it's decoupled because it's all RESTful calls anyway. The other thing I want to kind of highlight here is, I mean, since most of you are first-time summit attendees and some of you might be researching open stack, there's two resources that came out in the last summit called the Project Navigator by the Foundation. It's a phenomenal resource because if you go in Project Navigator, you see, you know, the various services, you get a description of what they do so you can tell that Nova is a compute service. But then there's also seven or eight criteria, actually eight criteria that we've defined of does it have an install guide? Is it packaged in a lot of the distributions that are available? Is the release team release on regular schedule? And so basically we use these attributes to kind of give a maturity score to the service. So this gives you a good idea of, you know, Nova, for example, I think meets seven out of the eight criteria or maybe eight out of the eight. The newer projects will be less, but that kind of shows you that, okay, I can do this service, but it is newer, so you kind of know what to expect when you're getting into a service. But likewise, there's links to the relevant docs there, and then there's also additional information. So if you go on Keystone, for example, you can click there and you see who the project team lead is, which is Steve Martinelli, and you see, like, you know, which most active, who are the most active and all of thoseestens based in a user setup space, et cetera. It's a very useful tool as you're looking into researching and consuming OpenStack. And the list of capabilities. So now as I described service architecture, the list of capabilities is ever-growing as well. And so the way to kind of look at it is there's services that provide like infrastructures and service components, then there are services, because OpenStack Clouds can be massively scalable, there are services that actually help you monitor and manage the Cloud itself, better provisioning tools, or even testing the code for stability and whatnot. And then there's advanced services that kind of give you the experience of consuming IaaS but providing a value at service on top of it. So at this layer, you have things like being able to do databases as a service, or being able to leverage the other IaaS resources to DNS management, key management, et cetera. And through all of this, because services kind of talk to each other and they authenticate through tokens, Keystone is a common component used both for user authentication as well as helps out with service-to-service communication. And then from a development perspective, since there are so many projects, instead of building the same function over and over in different projects, anything that's kind of a common need in the community, they have a common set of libraries that they've defined called Oslo which captured that need. So basically, we reduce duplicity of code as well as reduce potential errors because everyone's using the same function calls more or less. And the other thing that I'd like to mention is just use cases in general. OpenStack is not built for one specific use case, it's not just for test dev. People are using it for all sorts of environments, test dev, QA production, and for all sorts of workloads. And there's actually, I don't have the URL here, but there's actually a website where if you go to opensack.org slash user dash stories, you can see public references in all sectors that are using OpenStack. So I'll give you a really good idea of what types of companies with which backgrounds and what size are consuming OpenStack. Furthermore, Tyler and I, along with the Enterprise Working Group recently, just authored this book which is coming out at this summit called OpenStack Path to Cloud. And this is actually a second part of a series that the Enterprise Work Group has been working on. The first part was a business primer on OpenStack. So it was about how do you do the research and justify how to look for ROI, TCO. So making the business decision to go with cloud and OpenStack in particular. This second book in the series is more about you've made the decision to go OpenStack. So what are your choices? As Tyler mentioned, the different cloud consumption models you have, there's different deployment models, whether it's public, private, hybrid, community, et cetera. And then there's also the decision of which workloads do I start with to maximize my success probability of getting funding, getting more users on board, et cetera. And so this book contains all of this. It's an e-book, so it's available at OpenStack.org slash Enterprise. Tyler and I also have maybe like four or five copies. So if anyone wants one, we'll be glad to give one. But again, as you get started as a community, you know, it's good to kind of understand services and research, you know, how do you make the business case to help your OpenStack deployment be successful, along with the technical finding someone who to work with that has the technical capabilities to make you successful as well. And then also, as I mentioned in the community, getting involved is important as well. So there's tons of ways to get involved in the community, whether it's joining the foundation, joining mailing lists, attending meetings on IRC, helping with code or even just doing reviews of code. So tons of ways to get involved. But with that, I'm going to pass it on to Brad, who's going to share how IBM actually is involved with OpenStack. Thank you. Neither is fancy devices. Hi, everybody. So I wanted to talk to you all about how IBM has been involved in OpenStack, particularly in contributing. One of the interesting things about an open source community like OpenStack is, well, how do you have influence? Well, do you show up with a fancy title or a big thing of money or, you know, fancy degree? None of that really works. OpenStack, as a community, you have influence by rolling up your sleeves and becoming a contributor. You become a contributor by starting to learn, do some bug fixes and maybe do some documentation fixes and get to know the other contributors. I learned this firsthand. I started contributing to OpenStack, you know, at the end of 2012. I hid the fact that I had a fancy IBM title and I had a PhD and I rolled up my sleeves and was junior programmer, learning the ropes of OpenStack. I think I went about eight months before people realized I was an executive at IBM, technical executive, and then they looked at me funny for about two weeks, and then they got over it. But that's what you do. If you want to have influence, nobody cared that I had a PhD, nobody cared I had a fancy IBM title. It was, well, Brad, how are you helping the community? How are you helping the software? What can you contribute? So with that, if you look at IBM, since the early days here, we've been ramping up our contributions, going all the way up to where we are in Mataka now with 42 core contributors, you know, over 200 technical contributors, over 500 people at IBM helping to push OpenStack in some fashion, go to market capabilities, business development, what have you, and making key contributions in all of these. We've really been ramping up and we've got a lot to be proud of. Number one, it commits and reviews to Keystone. Number two, it commits to Nova. Number one, it commits to Senlin. Number one, it commits and reviews to Barbican, Refstack. We've been ramping up our leadership. We've got nine technical leads, so key projects, Keystone, Glantz, Nova, Refstack, Heat Translator, Senlin. These people are elected by their peers in the community and they've worked their way up for many a year to have those leadership positions and it helps us to drive the changes that customers want. That's what's so important to us, and I think it's a great thing to do, that's what's important is to drive the enhancements features that we need and that's why we invest so much in OpenStack to be able to do that. So if you look at there, lots of commits, lots of contributions. Where do we spend time contributing? If you look at where we do a lot of the main projects, security. We have a lot of enterprise customers, so security and integration with things like LDAP and Active Directory and Federated Identity. These are things that we've been driving since the beginning. I remember when I first started working on OpenStack, the identity component of it, which is called Keystone, didn't even have a secure connection to LDAP. So if you're gonna connect to your LDAP, you didn't even have a secure connection. It's kind of a problem. I actually put in the TLS connection myself between Keystone and LDAP. So little features that enterprises typically like to see in something we wanna make enterprise strength. If you look at storage, IBM's been huge in various capabilities of storage. If you think what we call migration, retyping. So if you wanna use the block storage and go from one backend and IBM backend to another vendor's backend, we call that retyping. We've done a lot of work in the migration, supporting migration and also replication. We've embraced the database as a service capabilities of OpenStack and we'll talk a little more about those. Containers, we'll talk about in more detail, but containers and networking and how those two work together are other areas where IBM's been huge because we and our customers see how important it is for OpenStack to provide you an infrastructure that allow you to run containers on top and allow you to have some really nice seamless networking options. We've also been working about core networking issues and I'll talk a little bit about that on another chart. Got lots of cores working on the dashboard, made the dashboard extensible so that operators can add more features to the dashboard if they need it. Adding features like AngularJS prove the performance as well. Another area we've been huge is interoperability. So folks want OpenStack deployments to be interoperable and we've set up a project called RefStack and lead it that helped drive where folks can send in the results and we can validate. Yes, these things are interoperating as we've said they would, passing the tests. So if we wanna go into more details about where we're contributing, networking. So anybody looked at OpenStack networking yet? Yeah, not too many. So here's what you're gonna see. Been doing a lot of work with these things called OVS, OVN, improving the scalability for L2, L3 networking capabilities that OpenStack provides. Doing a lot of work and providing monitoring troubleshooting tools. As you do this advanced networking and the networking capabilities that OpenStack offers, you gotta start getting comfortable with these advanced networking topics and we're trying to make it easier to learn those, debug those. Another area where we've seen a lot of pickup, anybody from Telcos? Anybody here from Telco company? All right, which Telco company from? Which one? Excellent, how about the other person behind you? Nope, anyone's asleep. So if you look at it, we get a huge pickup from the Telcos in an area that they call network function virtualization. And they're seeing OpenStack as a great way to do all this network function virtualization. And we've been putting the pieces in place so that the folks in the N of V world can embrace the OpenStack world. Now what does that mean? If you look at what the Telcos are like, they like a standard based approach and they're actually using one called Tosca with creating these things called the standard profiles for all their network function and virtualization. Eventually all that stuff has to be translated and mapped down to stuff that OpenStack does, namely heat and hot templates. We actually lead some projects that allow that translation to occur so that all the things that they love to specify can then map down to OpenStack. So we've been doing a lot of work there and helping out the N of V related projects in OpenStack. One of them's called Tasker, Tasker rather, to help that area. Storage and database, right? So again, I've mentioned block storage, we've been huge, volume, migration, retyping, replication. Been doing a lot of quality enhancements in those areas to make sure that stuff works the way we claim it's gonna work. Obviously we've been improving our own drivers with replication capabilities. We've also been doing some work with the object storage of OpenStack, which is called Swift, not to be confused with the Apple language. It's, you know, when you're IBM and you do open source projects, because that's all we do these days is open source and you've got OpenStack Swift and then you've got Apple Swift, which is also open source, it gets quite confusing. So this is the one that's part of OpenStack, you know, adding your features there for container synchronization. We've also been doing a lot of work with improving the database as a service option of OpenStack. It's a neat feature and we've been adding support for CouchDB, which is an open source database that's very popular as well as our own DB2 and providing capabilities there that allow those databases to run seamlessly as part of an OpenStack environment. So we've been having a lot of fun with the Trove team driving those kinds of enhancements. You know, lately in security, there's a new project for storing keys and secrets called Barbican, our folks have been adding features to that to make it better and extra capabilities there. The main component for security is called Keystone. It does your identity, authentication, access management for all of OpenStack, it's what you use to integrate into LDAP, integrate into active directories or an identity manager. We've been doing a lot of work there and most recently quality improvements as well as adding the notion of domain-specific policy support. So if you have different domains, different groups, different what-have-you that need different security policies, a lot of work is being done in that area to make that more seamless and more fine-grained security policy capabilities there. Containers, anybody interested in containers? Any hands? Anybody? Got to be more than that, it's a hot topic. Don't be shy. Containers are huge, we've got folks working in the container project, which is Magnum of OpenStack and that improves the ability of OpenStack to support things like Mesa, Swarm and Kubernetes. And then there's also, containers are so popular, there's also work being done to improve the networking capabilities for the containers so that you have a container and it can get an IP address and it runs seamlessly to what else you need and the rest of your environment. There's a project called Courier. So our folks have been working to make sure that overall story of the containers and the networking is a seamless story. Is there more work to be done there? Absolutely, we're just gonna keep pounding away at it. You can see the list there of things that we're doing. But we're really trying to bridge those worlds and make that a seamless experience for folks that love the containers and love the OpenStack. Compute, people familiar with the compute part of OpenStack Nova, we're doing a huge amount of bug fixes. Our big contributions there are the leader in bug fixes and improved scheduling support. So improving, it's a very established part of OpenStack. It's one of the main things you'll use and we just continue to improve the quality. We also do some incubators. There's a policy for auto-scaling project called Sendlin. We've been having a lot of success with telecommunication companies in China using that to improve their scaling. So a lot of work being done there as well. And some of us actually even write some books. And some of us actually bring more than four or five copies to the conference. So we actually have a book signing today and we don't have four or five. We do have four or 500. Please come see us one to two myself and the other authors, Steve Martinelli and Henry Nash, Keystone cores, will be at the booth. And if you're not familiar, we'll cover Keystone, which if you're new to OpenStack, typically the first thing you gotta figure out is the security, how you're gonna integrate into your existing LDAPs, active directories or your identity options. And so that's typically the first problems you hit and fortunately we got a nice Riley book to get you started. We'll be there one to two, come sign. Who wants this one? Anybody? First hand one up. I'm so happy I didn't bust his nose open and the stitches would've been terrible. And then if you wanna read more, there's some places you can read more. There's some, you can just Google out there for my name or whatever. I've got some OpenTech articles that you can look at and different things you can find. But it's usually pretty easy because I've been blasting them out on Twitter for the last three or four weeks incessantly. All right, well we'll be happy to take any questions. Any questions at all? Anything I know? We're gonna have our next speaker up. It's gonna be talking about the cloud market and opportunities and gonna be talking about our modern cloud stack. So please stick around for that. And I think that speaker is gonna get set up as we speak. So don't be shy, always come talk to us.