 So good morning. I'm Todd Moore, my BM's VP for open technology. I've been involved in open source for a pretty long time. Back to the days of when we first found Linux and decided to start bringing it into IBM and through many, many projects. I was part of OpenStack before we had an OpenStack foundation. I worked with Mark and Jonathan to help structure what we were doing, write the bylaws, put things together, stand it up. And I've been part of the board the entire time. What we've done today here in the IBM track is to really just come and put together, hopefully, some exceptionally good information for those of you who are getting started here in the morning. And then in the afternoon, we're going to go dive in deep on microservices. And we've got friends here with us to help us go and do that too. Folks from AT&T, and thank you. Good. Appreciate that. Folks from AT&T and Materna, so we'll have real use cases that we can go and talk about here in the morning. And also along the way here, Jesse Proudman is going to come up. I think everybody knows Jesse from Blue Box, and it's OpenStack fame of some of the great presentations he's given here over the years. So Jesse will be coming in to talk about how we see the market and the possibilities around OpenStack and the things that we see our customers doing. And then in the afternoon, we're going to take a break around 4.30 and serve some adult beverages and invite in the folks. If you went to the keynote, you saw the interop challenge. Everybody go to the keynote. Who didn't go to the keynote? Everybody go? Good. So at the keynote, we saw 16 companies up on stage. There's actually 18 who did the interop challenge and all passed the test, but we had room for 16 and they came and became part of it. So those folks are going to come back in and you can ask questions. You can talk about the best practices that they figured out along the way about how to go and do that, the tools they used. I think Brad talked about using Ansible and the shade plug-in and the scripts that they used and how they structured it. So it was very easy to portably move an application from one cloud to the next. And you heard Don Ripper reinforce the idea that interoperability, portability, choice is really important. And those are things that we and IBM truly believe in. It's all boats float with a rising tide. We're creating a large market opportunity for everyone to come and play in and we've come together to go build OpenStack in a creative, collaborative, open way. So hopefully everybody who comes sees that and gets a chance to talk to those guys a little bit about how we've gone and accomplished that through RaphStack and DevCorp. And then in the afternoon, we're going to have some of our folks who are really deep into what we've been doing in microservices and start sort of down through the basics at the infrastructure level and then walking it up through how you can go about deploying and building microservices and then a look at the end through the eyes of DevOps, right? Because if you're going out there and you're building your new cloud infrastructure and getting going and you're using all the same old processes that you used before, the same sign-off and check-offs and same delayed deployment processes that you have, that's probably not going to get you the kind of benefit that you were looking for. So we'll talk to you about how to go about thinking about everything that you do and thinking about how to integrate into that DevOps pipeline. So hopefully you'll come in and out through the day and be part of it or stay the entire day, love to have you, and join the guys who did the interop challenge and have a drink with them as well along the way here. So thank you very much and let me turn it over to Shamil and Tyler and let you guys get going. Thank you. Thanks, Todd. So what we're going to do is we're going to start it off with a little OpenStack 101. How many people is this your first summit you've ever been to? All right. A lot of new folks. Welcome. Welcome to the madness. So this is straight off the OpenStack.org website. What is OpenStack? It's a cloud operating system, right? We can use the control pools, the compute storage networking, all through the data center, manage it through a single dashboard, single API. But let's kind of pull that apart a little bit. It's a group of open source projects. Right now we're north of 60 official projects under OpenStack Big Ten Governance and all of these projects or services communicate with RESTful APIs. It originally emerged from a collaboration between NASA and RACSpace in 2010. And the foundation is a totally separate independent legal nonprofit entity that manages, you know, puts on these great summits and handles all of the governance of OpenStack. And it's written predominantly in Python. So that sounds great, but sometimes an analogy helps. So say your fridge looks like this and you're hungry. One of the great options out there to get food, at least in the States, is a thing called GrubHub, where you can go online and they have food from all of these different local companies where you can get if you want burgers, nachos. You don't have to go directly to that company. You can go to GrubHub and pick out, you can even pick out a whole bunch of stuff. OpenStack really is the GrubHub of computing here where OpenStack is not providing the hypervisor. OpenStack is getting you resources from the hypervisor. So instead of those individual resources being considered all part of OpenStack, it's a totally pluggable architecture. So just like GrubHub may add a new barbecue restaurant, OpenStack may add support for a new storage array, a new networking technology or a new hypervisor just by adding new drivers. So OpenStack is that in-between layer between you as the user and the underlying resources where a lot of people think it's all just one component. And on that note, here's really, I think, what's important, what it isn't. So it's not a hypervisor and supports a pretty large number of hypervisors, including vSphere. It's not a free VMware replacement. This is something we see people take a look at OpenStack for and struggle with mightily. We're like, oh, we pay money for licenses, so let's just do this free OpenStack stuff and it's not like that. It's much more similar to the AWS model or Google model where the instances are much more ephemeral. So the whole pets versus cattle analogy, this is where your cattle live versus your pets. And the key thing is OpenStack itself is not a product. You can get it through vendors like IBM and others in the ecosystem, but the code itself is fully open source and available directly without having to go to a vendor if you don't want to. And because of that, it's not a single distribution, so there's no one open stack. Each vendor takes it and curates their own distributions just like Linux. So if you're comfortable with the Linux model, OpenStack follows that pretty similarly. And it's also not a single project. Like we talked about, there's north of 60 projects in the Big Tent now. You say, well, we're using OpenStack. What projects within the OpenStack community are you using? Are you using Nova or Neutron or Keystone or Magnum or Barbican? On that note, I would also like to add, there's actually a project navigator on OpenStack.org, which because there are 60 different projects that is really helpful, especially when you're starting into OpenStack, to go there. In the project navigator, you can see a brief description of each service within OpenStack as well as who the PTL is, along with some maturity indicators as well, as far as what's the adoption rate for the service, how many SDKs support it. So you can get a good sense of what is the service and how many people are using it in a very simple snapshot using the project navigator. And that's really key because I think a lot of us are used to the model of this is a GA software, this is beta, and OpenSource in general, what's specifically OpenStack doesn't follow that model. So is the project ready? Well, it depends. Does it meet your maturity guidelines? So for you, online upgrades are, that's a bare minimum for your particular deployment. Well, you may judge that for which projects fit that, which ones don't. It's not a storage platform. Obviously there's storage projects within there, including Swift. And it's not network virtualization. It supports a wide range of network virtualization technologies. We saw a lot of the cool stuff even yesterday with the NFV use cases around OpenStack. But OpenStack of itself, again, like Grubhub, it doesn't have any of the food itself. It's getting it from somewhere else. So it supports a wide range of technologies. One of the questions we talked about here, here's what OpenStack is, but why does it matter? Why is it important? Why are we all here? Why did we come to Barcelona? It's one of the only OpenSource projects that gives the full comprehensive cloud services, a framework to build a comprehensive cloud. Significant, as we saw in the keynote, significant cloud service providers are using OpenStack. And like Linux, we think it's going to have a larger impact on organizations in the future. So as Linux has quickly become the de facto server platform, we're seeing OpenStack be that de facto cloud platform. And really, it's over at this point. So there was, you know, in not a few years ago, there was a number of different options here. But we've really seen OpenStack as kind of taking that mantle of the open-source cloud option. So I mentioned before, it started in 2010. NASA and Rackspace got together and decided to start this project. And it got going. And in 2011, the first public cloud was launched using the NOVA and Swift projects. And from there, that was when it hit a key growing point, which a growing pain point, which is, you know, people were uncomfortable with a single company doing it, and that's when the foundation was created. So the foundation was created and OpenStack was moved there. And that's really when it started to take off. So following those twice a year release cycles, Grizzly added block storage and networking, the OpenStack Marketplace, that's another great resource on the OpenStack.org webpage. So, you know, if you saw the interoperability demo this morning, you can also go in there, look up providers, and see if their distribution or managed cloud offering passed all of the RefStack tests. You can see all that there. It's right in the Marketplace. So you can see where your options are, where you want to go to get OpenStack. 2015, last year, was when the BigTent model that we talked about earlier was introduced to bring all these other projects into OpenStack. And another big one last year was CERN crossing 150,000 cores. And then we heard yesterday that they're north of 190,000 cores. So very large use cases, very large scale for OpenStack. And then this year, you know, we just, just not too long ago Newton came out. And we're going to talk a little bit more about the Newton release. And I would say that, you know, the growth is still staggering. Like if you look at every summit, just as we saw in this room when we first asked how many people are attending the summit for the first time, on average 50% plus people are new to summit every, every time. So there's definitely more and more people also, you know, join the community, becoming members and participating. And so the pace is phenomenal. Yeah, and this just continues that you can see the number of companies, individual members of the foundation, as well as, you know, north of 6,000 developers contributing and as of now over 300,000 code contributions. So it's a project of staggering size. And it's just the only way that's possible is just all the people in the community contributing in many different ways. So I mentioned the twice a year release cycle. Basically, the way this is organized is the names of them are in alphabetical order and they're based on where the design summit location was. So that's why we're on Newton, the next one's Okada. And as we go through each one, anyone that's a member of the foundation can get a vote on the, and nominate potential names for releases to a year. But projects can release independently. So Swift's a common one that releases off cycle. But projects can release whenever they want, but it's captured together in those twice a year releases. The stable branch is the support schedule for, you know, what happens, okay, Newton's out now. What happens to, I'm running Mataka. You know, what happens to me now? I'm running a N minus one release. So it's all the what happens past there. So that first six months, all the bug fixes are back ported. When you get to six to 12 months, that's when it's critical bugs and security patches. And then once it's over a year old, it's basically just security fixes. And a key distinction here is this is the upstream support model as well. So, you know, there's different models for support as well, but this is what the community has to adhere to. Yes, that's a key point. Obviously, the individual vendors then can extend that however they want. And that's some of the values we can provide as vendors to the community is supporting specific releases for longer, packaging, pulling patches, back, things like that. And on that note, the other question we get is these different ways. We talked about there's distro vendors and all these different ways to get open stack. And this is an analogy that we like a lot is coffee, right? So if you like coffee like us, you can get it a number of ways. And the easiest way is to go across the mall and walk up to the Starbucks, place an order, and you'll get your latte handed to you right there. And that is, you know, what we do at IBM with our Bluemix Private Cloud is it's fully managed open stack. So we do all that work for you. If you're looking for a little bit more control, that's where you get into, with your coffee, you can go to get your Keurig machine. You're still limited. There's different pods you can get, but you can't get anything you want. But it's packaged for you, so it gets a lot easier. You control the strength and you have some more control over it. So that's more like the distribution vendor. So they have their own packaging and how they manage things, but it gives you more control of your environment. And then, obviously, there's always the do-it-yourself option. So you can roast your own beans, grind them, French press, and get exactly the exact coffee you want. That's really the do-it-yourself open stack model. So as you're going in that direction, you're getting more flexibility in exchange for you're doing more of the heavy lifting. So it's really finding, as an open stack user, it's finding the model that works for you and your company to be the most successful. The way all the services work, this is a pretty straightforward framework. Each service is a little bit different in each project, but this is kind of a high level. They're pretty similar from this perspective. So you start always with an API on the top, and inner service communications always be that API. There's a message queue that connects that API to a scheduler of sorts. So say we're talking about Nova, and you want to make an API request to spin up some instances. The scheduler is taking that request and figuring out which resources are available to be able to service that. The underlying service, there's always a database on the back end to keep track of all the resources that that individual project has, which VMs you've created, all those things. If it's Neutron, which networks you have. And then that talks down to a driver, to talk to that individual piece. So if you're using Nova, and you're using KVM, the Libvert plugin, if you're using VMware, Zen, Hyper-V, each of them have their own plugin, and then that's how it talks to the underlying provider. So like we talked about earlier, the Hypervisor OpenStack isn't a Hypervisor. It works for all these different Hypervisors. In that case, that's when it talks to the specific Hypervisors. So because of that, it gives you the ability to use different things on a daily plane, and in some of the project cases, multiple at the same time. So you could support multiple different underlying components. And because of that, the key piece is the interoperability. And now we've been talking, we saw it in the keynote, we're talking about it now and throughout. And Anashima, you want to talk a little bit more about interop? Yeah, so interoperability in OpenStack, a simple way to explain it is to use an example of standard power plugs, basically. So if you think about it, really, your programming interface is the same as having a power interface as well. So the OpenStack API can be considered the plug, which is defined, falls into a defined standard. And because of that, regardless of what appliance you have, or in this case application that you've built, as long as you're interfacing with the plug in a standardized manner, you can plug in, basically. And at the same time, because that plug is standardized, it doesn't matter whether you're plugging a laptop or a toaster into your home or office, which in this case could be different OpenStack clouds. So effectively, by standardizing on the interface, we are able to accommodate different types of appliances, and vendors or providers can actually do different things. You can have different cord links, you can have different types of cords, et cetera, as long as the plug is the same. And at the same time, you can take your laptop from home to your office and vice versa, because you're plugging into the same standardized plug. Yeah, I mean, that's really the goal of, you know, having that Open API open standard is avoiding that lock-in. So, yeah, you don't have a laptop that only plugs in at your house that you can't plug in at the office, and that's really what OpenStack brings you, is that capability to build your apps and then run them on any compatible OpenStack cloud. Even the number of projects we talked about, north of 60 already, this is just a small subset of the projects that are out there, and their names, some are pretty clever. But there's all the type of things you would think of in a cloud. You know, if you want DNS as a service, key management, you know, a service catalog, all these different, in the database as a service, there's all these projects in OpenStack that can satisfy the needs of your particular cloud and what you need, and you just decide which of those projects you want to deploy. If you want, you know, if you want to do bare metal, there's the ironic project. You know, you pick those things. Generally, people deploy the core, you know, the Keystone, Nova, Neutron, kind of core, but it's up to you as an individual user which projects make sense to you. Some of the popular use cases for OpenStack, web hosting, high throughput computing, we heard yesterday from CERN running public clouds like we heard today from the number of European public clouds running on OpenStack, web services and e-commerce type workloads, big data, databases as a service, video processing and content delivery, and a big one now we're seeing is containers. So containers are great, things like Kubernetes, Docker Swarm, but what do we run those on? So we see a lot of people using OpenStack to deploy the resources to run those containers. And beyond just the use cases, there's also, you know, a multitude of markets and organization types that are using OpenStack as well. So you've got everything from, you know, as we've seen through the keynotes and probably through the summits so far, you've got research organizations using OpenStack, you've got telecom using OpenStack, enterprise, etc. And so depending on your market and your needs, there's actually working groups within OpenStack as well which align to that. So if you are, you know, in the academic or research industry, you can actually participate in a scientific working group and, you know, get together with your peers to kind of discuss what are your common needs and stay for enterprise, but it's just not use cases. There's also, you know, the markets in OpenStack are very diverse and broad as well. Yeah, absolutely. And to kind of help navigate that and figure out, well, you know, we talked about different ways you can consume it and manage versus distributions versus do yourself different projects you can deploy. How do we figure out... OpenStack sounds great. It's going to mean our needs. How do we figure out how to get there? So the foundation has put together a number of e-books. There's two available. If you go to opensack.org slash enterprise. The first one is talking about the business cases of OpenStack. And then the path to cloud one is helping you figure out how do you get... how do you get going with OpenStack? What kind of team do you need? You know, helping you figure out how to pick the model that works for you. The e-book is there. We actually have some hard copies available after the session. So if you want to come up, we have some of them available here. But if you don't get to grab one, like I said, they're on the website. And there's also a link to Amazon to get them from Amazon. This all sounds great. Well, how do we get more involved? It's not just, you know, twice a year coming to a summit or once a year. There's so many different ways you can get involved in OpenStack. And the key thing is if you're not already, which I think you have to be to buy a ticket to the summit, but join the foundation, allows you to vote for... not just release names, but members on the different committees and things like that. There's tons of mailing lists for all those different working groups or individual projects that you're interested in. And you can see on the OpenStack site there's a list of all of them. All of the meetings or most of the meetings for the projects and working groups and things like that are held on IRC. So the nice thing about that is transcripts are kept. So even if you miss a meeting or it doesn't work with your time zone, it happens to be 2 a.m., you can go back and see what happened in that meeting as well. If you're interested in contributing code to OpenStack, there's a really good guide on the OpenStack.org site to how to get your environment set up to do some coding. And it's not just code, right? One of the most overlooked things I see is documentation. So even if you don't know any bit of Python, documentation could always use some help, and it's pretty easy to contribute. And the next summit that you attend, if you come a few days early, there's also Upstream University, which happens generally a day or two before the summit actually begins. And in there you'll get core contributors and people that have been participating in the OpenStack technical community as well, actually walking you through how do you do a commit to OpenStack, what are best practices for commit messages, how do you interface and get buy-in for your blueprints or your code changes, etc. So it's very worthwhile if you plan on developing within the OpenStack community to attend UpstreamU as a good starting point. Yeah, and you can always get started with something as simple as code reviews. So you can see other people's code, look through it, give comments, give a plus one or minus one on it, and just even to get introduced to the community. So there's a lot of ways to be involved, and it doesn't have to be just coding, there's a lot of different ways you can influence the community and the direction of the projects and give back to the community without just writing lines of code. So we're going to talk about Newton specifically, what's new with Newton and what some of the models are. So Newton is the leading release of the OpenStack platform and really in Newton the community came together and really innovated across the board whether it was Enhance-Masonova which is a virtual compute, Ironic which is bare metal, Magnum or Courier which are container-related, and really to define OpenStack as the one platform to manage all three. Whether it's virtual machines, bare metal containers and also as I mentioned earlier there's a wide variety of industries and markets that are participating in OpenStack as well and really as those needs from user surface, OpenStack tries to accommodate those so we're running more workloads across more industries and that continues to be a trend release after release. And then finally through Newton we've also made it a lot easier to manage these elements so there's key enhancements to Horizon which is the user interface as well as enhancements to APIs such as micro versioning in certain projects that allows APIs to be more flexible so they can be incrementally updated with new features as well as allows potentially APIs to maintain some level of backwards compatibility as they move forward as well. And a key thing there is even something as simple as the command line interface. Previously each project had their own command line commands and they didn't all exactly follow the same semantics and things like that so that can be a challenging user so the OpenStack Python client is now on the 3.0 something release. It's bringing all those together under a single command set which has made things significantly easier for users as well. And so obviously as we talked about there were probably over 2,000 plus contributors to this release, a lot of moving parts but we looked at one of the key release themes for Newton and it comes out to resiliency and user experience. And user experience is not just the operator experience so there's multiple types of users. So users are very loaded word within the community. User can pertain to people operating clouds, it can pertain to people developing applications to run on the cloud and it could even be considered the end user who might be doing self service through OpenStack provided applications as well. So across the board those have been key focuses and we've been ramping up on how to scale up, scale down for example in Nova we've been also focused on this concept called Cells V2 in Nova which allows better segmentation of the various layers of the Nova service to make it more scalable. Heat has also moved to a horizontal engine so Heat in the past had a single engine and now they can actually use a distributed engine model going forward and for bare metal there have been actually strides in Newton as well. The key two things for bare metal namely have been the introduction of multi-tenant networking as well as the ability to have multiple compute services managing bare metal to introduce a notion of high availability for bare metal as well. And then from a resiliency perspective the Keystone team did phenomenal work in introducing rolling upgrades as a concept within the project and Neutron has made strides around database migrations as well. And then from a user experience perspective Nova has allowed the capability to change some of the parameters without having to reload the service as well as they've taken the policy files in OpenStack define access to certain services and capabilities of the service itself and Nova has actually made the policy embedded in code now so as a user or operator you no longer have to make sure the file is set up properly the only time you modify the file now is if you're overriding in default so some of these things might seem little but at scale they provide a lot of value overall. And then with Newton the OpenStack was awarded the infrastructure initiative badge because of the great work by the OpenStack security team in terms of making strides in process around vulnerability management as well as the contributor stepping up and the velocity of security focus and changes has also increased greatly in the community. The other big thing is as we've been talking about interoperability you probably have heard DEF-Core reference a few times today and the DEF-Core working group has now been renamed the introp working group and really it's to reflect its strategic mission because introp really captures the essence of what that team is trying to do and the introp team is actually working currently on new guidelines as we've talked about intropability they're working on how do we expand intropability to include additional capabilities I've introduced in in Makata for example because they kind of go to releases through releases back and the same time a lot of the work that was done in the intropability challenge will also be fed back into that team so they can kind of do an analysis of based on these discoveries are there action items that we can take and help resolve to make intropability better for everyone? Yeah this is some of the other big trends we've seen with Newton the number of public cloud providers for a while had the growth had slowed down there a bit but now it's picked back up on the public cloud side and a really interesting one has been the amount of growth in China around OpenStack it's not just in the first place I think we saw a lot was the telco market but now it's expanded out from there Absolutely across all industries the market in China is growing phenomenally for OpenStack clouds as well as they just had their first OpenStack Day event as well which was a huge success so I can only anticipate that's going to be growing based on the adoption going forward From an IBM perspective as Todd mentioned he was involved in the project before it was a foundation before the OpenStack foundation was created IBM has been committed to OpenStack for a long time as well and so on this slide you can actually see our pace of ramping up contributors as well as our dedication to OpenStack through the releases and notably if you look at Newton in Newton we had 42 core contributors and we had the key areas of security as I mentioned security's been a big area that we've been working on as well as networking enhancements and then definitely interoperability is a top priority as well so from a networking perspective the team has been working on helping scale the OpenStack OVN support as well as helping with the get me a network feature and get me a network to describe what that is is effectively as someone who is provisioning an instance in the past you would have to make sure that your provider network or your networking is actually set up before launching the instance and in this case you can actually the operator can specify a auto allocation topology and you as a user when you boot up an instance if there's no network available and auto allocate is turned on it will actually provision all the underlying things that are needed to get you connected on the network security wise we've been working as I mentioned on solidifying the process around vulnerability management as well as making sure that vulnerability management as a tag is applied to projects to show maturity of a project in service as well and the key piece there is a lot of work IBM has worked on Keystone to provide capabilities for rolling upgrades and it seems initially disconnected but rolling upgrades for any of the projects are a key piece from a security perspective because if you have to take an outage to do an upgrade you're not going to do them as often and then you're not going to apply patches as often and that's a security challenge so the more often we can apply these patches and patch these services the more secure and the good news there is through the work that was done by Cinder through Keystone now through Nova there's a standard set of things that projects know now to do to enable rolling upgrade support so in the future we would expect more and more projects to be able to support a similar model hopefully we also worked tremendously in the OpenStack Heat project to enhance the capabilities of taking Tosca templates and translating them into the heat as well so we've introduced support for authentication we've introduced support for auto scaling as well when you do the translation and for Trove we enhanced the db2 guest image and the driver to allow for various benefits as well such as online backups and also being able to manage the configurations of the database through the Trove service itself finally just to give a good snapshot of where was IBM through Newton we had two of the top 10 contributors overall we had 42 different cores and we were number 5 in terms of both reviewing and committing code into OpenStack Newton release and here's a snapshot of all the different contributions that we made through this release so phenomenal work and velocity by the community and IBM is definitely trying to play our part in this evolution of OpenStack if you're interested in more about the Newton release not just what IBM is doing but what's new in it Brad Topol has a blog post going into the details of what's in the release and then also as always the foundation has a landing place on the website for the latest version the release notes if you will and here's the highlights here's what's new and all that type of information the work on the identity management stuff we have a great O'Reilly book that a couple people from IBM were involved with and we have copies that's available at the IBM booth so if you're interested in a free copy it's a great deep dive book on Keystone it's available at the IBM booth so that's that's it for us for now in just a little bit we're going to take a short break and we're going to come back in about 15, 10, 15 minutes with Jesse Pradman to talk about the IBM Open Cloud so thanks a lot everyone thank you