 Good morning everybody. My name is Martijn Barke. I work for VMware as one of their cloud strategists. And this is the session about large-scale mission critical enterprise deployments with OpenStack. And I'm going to talk first about how VMware Integrated OpenStack works. And afterwards I've got Julius Hariban. He works for Medeus, so he's one of our customers and he's going to talk to you about how large-scale mission critical deployments actually work, because I can talk about it from a product perspective, but he's actually doing the real work and I think that's a valuable lesson for you guys to learn. But first of all, I wanted to start. So I'll first give you a story about VMware and OpenStack. So what's the story? Of course, there's a little bit of an introduction to what VMware Integrated OpenStack is, and then we'll go into the journey that basically Julius went on with Medeus and VMware Integrated OpenStack. So that's the thing that we're going to talk about. Last but not least, basically a next step for you guys to learn more about VMware Integrated OpenStack. And of course there's a little bit more time for Q&A. That's the thing that we really want to do if you've got questions for either me or for Julius around the topic of VMware Integrated OpenStack. But let's start. So this fits into a broader perspective, a broader vision that we have from a VMware perspective. What we're trying to achieve from a VMware perspective is consistent infrastructure, consistent operations across the private cloud, the public cloud, and the edge. That's the end goal, because a lot of customers are nowadays using data centers, but it consists of a multitude of things that they want to achieve. It's out of these three infrastructures, it's public cloud, but still there's a large amount of customers that are still looking for OpenStack. That's the thing that we need to achieve. From a VMware perspective, we're trying to accommodate all of them for them to be able to manage and to host the applications, because that's in the end what needs to happen from an enterprise perspective. You need to be able to host any type of application and you need to have choice of what infrastructure you want to run on. Someone to move from a private cloud to the public cloud, but OpenStack is still a very valid thing. And from a VMware perspective, this is more or less the vision that we have for a VMware cloud. It basically consists out of a multitude of infrastructures you see at the bottom, but it's all to host either virtual machines, containers, or Kubernetes. And that's what we typically talk about, and what we're trying to achieve from a VMware perspective is to create that consistent infrastructure for both virtual machines and containers. From a VMware perspective, we see then that you need to manage all those resources. If you're either running it on private cloud or the public cloud or on the edge, you need to have a consistent layer of management to basically manage all those different resources. And that's the goal that we see from an IT operations perspective. You need to be able to manage all those resources effectively to be able to provide that as a service towards your customers. If it's either developers or line of business, they need to be able to host that to basically provide the applications. So we're looking also for consistent operations. And in the end, it's to manage both the existing applications, sort of traditional applications, but we now move into this world of containers with Kubernetes and that kind of stuff. So it's really to create a consistent developer experience. This is our picture. This is a picture that we provide from a VMware perspective. And then the question is always from, okay, but VMware, OpenStack, how does that play well together? Well, from our perspective, that's to be... It's a little bit of a black thing that was not the intention. But we actually love OpenStack. I think we've been one of the key contributors of what we're doing from a VMware perspective. It's basically... We've been doing this whole development from an OpenStack perspective for the last couple of years. We've been contributing to OpenStack. And we're basically providing to create a stable platform via our products. So VMware integrated OpenStack as our product. But we've been trying to add content to the OpenStack community for a long time now. We're actually trying to solve something. Because OpenStack, building that, building the cloud, that's not that complex. But it becomes more complex because you also need to monitor it. You need to troubleshoot it. You need to scale it up, scale it down. And you need to upgrade it. These are a lot of challenges that we see from a VMware perspective that are happening with a lot of customers. Customers are moving towards OpenStack. And they want to be able to do all those things. From our perspective, how we try to help is to create a faster time to value. That's actually what we're trying to achieve. And we are trying to simplify the complexity. Because we see a lot of customers struggling to deliver an OpenStack environment. Somebody makes a decision. And it takes about six to nine months to actually realize and get an OpenStack environment. The solution we have for that is basically something that we call the best of both worlds. So we see VMware integrated OpenStack as a solution that you host OpenStack on the best platform that we have available, our software-defined data center stack. Our software-defined data center stack consists of three things. Everybody hopefully knows. A data center consists of three things. Compute, storage and network. So the software-defined data center stack that we have from a VMware perspective is vSphere, which probably most of you in the room are familiar with. Storage perspective, we've got vSAN. And we've also got NSX, our networking, software-defined networking solution. And we combine that with a solution that out of the box provides you the OpenStack as an automated solution. And we've been doing this for quite some time now. We actually acquired a large OpenStack environment when we did the acquisition of NYSERA a couple of years ago from the NSX environment. Out of that acquisition came the product VMware integrated OpenStack, an automated solution to deploy OpenStack on top of your vSphere environment. That's basically what we're trying to do. But the result we're trying to achieve with VMware and to get an OpenStack is simplification, is ease of use, and to repurpose what you already have. So you're existing infrastructure, a lot of people are using vSphere. There's no point in moving away from that. We're trying to repurpose it and to build an automated solution on top of that. That's the thing we need to do. And also one of the key things, the knowledge that you already have from your existing infrastructure, we're trying to use towards that OpenStack environment that you're trying to build. So what is VMware integrated OpenStack? So it's basically the combination of OpenStack together with VMware integrated OpenStack management server that we provide. And that creates VMware integrated OpenStack. Of course, from a VMware perspective, we combine that with all the nice products that we have in our software defined data center stack, the combination creates you an OpenStack environment on top of our battle-tested environment solution from vSphere, vSan, and NSX. So where do we see our customers using this solution? So basically they use it for four things. Either it's an enterprise automation solution that is trying to provide with OpenStack, they trying to deliver a developer cloud, they're trying to enable the edge computing, or as we now have, we're trying to empower an AV solution, so a telco cloud. From a VMware perspective, we are moving into all these environments and trying to provide you with a consistent infrastructure on which you can provide OpenStack as a service. VMware integrated OpenStack is then the solution that makes it possible in an automated way to deploy OpenStack in all these three or four use cases, in an easy way, in a simple manner, and with a quicker time to value. That's actually the thing that we're trying to achieve. So how do we do that? Well, here you see an overview of the OpenStack components and how they integrate. So you see, for example, that we've got drivers for Nova, Neutron, Glanz and Cinder, and for Keystone, to be able to integrate into our software defined data center stack. That's the key goal that we're trying to achieve with VMware integrated OpenStack is the combination between what you provide from an OpenStack environment, and what we are trying to provide you from a software defined data center stack. The combination is actually the thing that we are trying to achieve a management solution that provides you for your integrated OpenStack, more or less out of the box. You install a server, sort of VMware integrated OpenStack management server. And that basically deploys OpenStack on top of your software defined data center. That's the key goal of leveraging the infrastructure that you already have. The assumption here is that you're running it on vSphere, and that you can then get the combination of VMware integrated OpenStack with the software defined data center stack. The latest release that we have actually is not 5.0, we just released 5.1. It's based on Queens. That's the thing that we're trying to put into the package. And it deploys automatically on top of the vSphere environment. That's the key goal. So what are the core differentiators from a VMware integrated OpenStack perspective? Well, simplified installation, it's a seamless upgrade, and we are trying to provide intelligent operations for OpenStack. So basically you need to look at it from if you have a vSphere environment, you use VMware integrated OpenStack. Within a couple of hours, you've got an OpenStack instance running, and we provide an intelligent operations framework so that you can manage and monitor the whole solution. And the best thing yet, if there's a new new solution coming up new OpenStack release, then we will provide you with an upgrade package to move from one version to the other. We battle time we test that in our environment, we provide that as a service towards our customers. You download the package into your VMware integrated OpenStack environment and at that moment, we can upgrade the solution from version to version. That's the key goal. So it all starts with this simplified installation method. So what we provided an OVA, so that's a packaging format for vSphere. You download that OVA, you import it into the vSphere client and that will give you a VMware integrated OpenStack instance. When you click this solution, you'll be able to provide it with configuration. And with that configuration, it will then automatically deploy, in this case, queens on top of your vSphere environment. That's the key goal. The virtual appliance is something that you can download from the My VMware website if you've got the correct licenses, because of course this comes with a licensing model from VMware. The other thing that we're trying to provide is this seamless upgrade experience, not only from an OpenStack perspective, but also from a Software Defined Data Center perspective. We take care of all the solutions and that's the key value that we're trying to provide from a VMware perspective is to be able to upgrade your Software Defined Data Center, but also do a hitless experience, an upgrade experience for the OpenStack piece. And that's where we saw a lot of customers struggle to move from one version to the other. We don't want people to stay on a previous lease because it's a hard thing to upgrade. No, you need to be able to upgrade whenever you want to do that. And that's where we provide more or less of an update. You can download it. You can upgrade whenever you want to do that, to be able to move towards the new version of OpenStack whenever you want to do that. We provide you with that package. We already tested that package into our own environment and that will give you then that hitless, zero-touch experience that we're trying to achieve with this OpenStack environment. And the other thing that we're providing is not only the OpenStack intelligence, but we're trying to provide you an intelligent operations framework so that you can manage and monitor your OpenStack environment. So we've got four solutions that will help you. The first thing is VBLI operations and infrastructure monitoring solution that gives you insight of what's happening inside of your software-defined data center and in the OpenStack environment. We've got a management pack for VBLI's operations to be able to monitor the solution that gets into it. The other one is VBLI's loginsides. So of course a lot of logs come out of that. We've got a logging solution that gives you insight of what's happening from a log perspective so that you get an insight of what's happening inside of your environment. VBLI's operations is that monitoring tool that provides you a dashboard. If something goes wrong, you can use VBLI's loginside to troubleshoot and to get to the road cause analysis as soon as possible. So there's a combination between two. The other two things, VBLI's network inside, well that provides you from a networking layer and inside of what's happening. And that's a key thing when you're moving into the OpenStack environment is to get an insight of what's happening in your OpenStack environment from a networking perspective. VBLI's network inside will give you that insight and give you a visualization of how traffic flows through your OpenStack environment. And last but least, we've got the VBLI's business component which gives you insight from a cost perspective. It's not only from an operational perspective that you want to have insight of what's happening in your OpenStack environment. You also want to know about the costs and that's what we provide from a VBLI's business perspective. So this is on the operations side. But from an OpenStack perspective, we've got a longstanding commitment. We've been starting in 2015 with our Viya 1.0 solution. At that time it was based on Icehouse. But you see with every release, we've upgraded to a later version. And this gives you the ability to easily move from one version to the other. From Icehouse to Kilo, from Kilo to Mitaka, from Mitaka to Okata. And we're now at Queens and we'll probably move to the next version when it's there. Of course, this is a little bit of an outdated slide because it's now 5.0 and we're currently at 5.1. But you'll hopefully see that besides all the latest features that we provide from an OpenStack perspective, we also provide you a lot of value on all the other things. We currently also have Kubernetes on top of our VMI integrated OpenStack solution that you can use out of the box to give you your customers Kubernetes. We provide management in the security and metering, the management and perf, and of course we've got an advanced networking. The true value from VMI integrated OpenStack is you combine it with NSX. That's the true value because we have NSX as our software-defined networking solution to provide you a software-medic approach towards firewalls, layer 2, layer 3, load balancers, and integrate that into an OpenStack environment. That's the true value when you really trying to achieve the full potential out of VMI integrated OpenStack is when you combine it with our NSX software-defined networking solution. But maybe Julius can talk a little bit more about that when he's speaking about the yesterday's use case. So let's release, well, our solution is based on Queens, that's the latest one, and of course all the enhancements that went into Queens. We provide that as a package to our customers which they can use to upgrade their current solution and to be able to move towards this new version at a seamless and hitless experience. That's the thing we're trying to achieve in an easy way, in a consistent way, move towards the new version of OpenStack on top of a vSphere platform. The true value is that it simplifies the way we operate OpenStack. For that, we've been releasing customers that basically had a vSphere environment in a couple of hours. They were able to deliver OpenStack functionality towards their developers, their lines of business, and that's the key value of VMI integrated OpenStack. Besides that, it also requires no OpenStack PhD. It's something that we provide out of the box. I already said, you just need to download the OVA, import it into the vSphere client, put some configuration in it and it will then deploy it. It's an easy way to get OpenStack up and running on top of your vSphere environment. And again, it simplifies OpenStack into operations, not only from a VMware perspective, but also the way we monitor, manage, and operate the solution which can be handed out to developers to get an insight of what's happening inside of your OpenStack environment. And of course, last but not least, it's a single vendor support. So you call us if something goes wrong in that VMware integrated OpenStack environment, you dial our global support services and they will help you troubleshoot the solution. That's the key value that we are trying to deliver with VMware integrated OpenStack. Of course, it's nice to hear everything from me and I'm just here to tell you about all the noise features that are in VMware integrated OpenStack, but I think the key value of the key experience comes from customers that actually use it, so that's why I want to hand it over to Julius, who's going to talk about his implementation with Amadeus. So hello everybody, I'm Julius Heriban. I'm here for Amadeus data processing, or in general, we are a big service provider when it comes to a travel industry and I joined Amadeus roughly 16 months ago, so all I will present the key belongs to our company engineers, developers of Amadeus But I hope I can give you some insight, because I have some experience, not only for this, but as well for other clouds, which in the past five to ten years, I helped to build including things like the starting of the OpenTelecom cloud and some other VMware based clouds and outsourcing cloud services in general. So basically what is Amadeus about? If you came over to Berlin by plane, there's a very high chance that you use services of Amadeus without knowing. We are B2B business, we are focusing on services to airlines and we serve more than half of all airlines worldwide. We give services to airports, we give services to railway companies and so on and so forth and the highly transactional business, low latency requirements, consistency databases and so on and so forth. And that's actually the biggest challenge of Amadeus, to grow and to deliver high quality, always on services in an environment where the pace of change is everything and you need to deliver software in a kind of private SaaS model to businesses in this critical area and to serve the customer experience of all of you when you travel. Some numbers just to illustrate it are on the screen and roughly we process more transactional second than Google search. We have some to call it there and friends just to give you an impression. It's different type of transactions and we have big business which is around distributing the travel contents or if we go to Expedia and a lot of other sites and are looking for what is the cheapest fly from A to B and in the past you went to a travel agency and they gave you something like three offers and you took one and that was 10 years ago. Today you go to that page, look and gather kind of unbelievable amount of all sorts of potential combinations and you pick the cheapest that change from three look to one look and what we call the book ratio to today which is something like 100,000 books so request selects for some content and you click only on one and pay that 100,000 more than it used to be 5 to 10 years ago. That's the gross ratio we need to process when it comes to transactions and it's exponentially going and on the end hopefully in a transaction where you buy an airplane ticket or where you visit the airport and you go to the machine which is scanning your backcode and these transactions are processed by the systems of Amadeus. Not only Amadeus but a lot of them. I wrote some illustration pictures for some of our environments and this environment is not focused on the airline or air in general business but it's focusing on the so-called hospitality business so behind is a big hotel chain which is served directly in the US. We succeed to to be great for this hotel chain just few weeks ago at the 10,000 hotel milestone which is a great one for us and the whole environment is in a kind of twin data center DR environment to deliver really a 5.9 availability and it's hosted 100 percent really 100 percent on OpenStack. More inside of that and you see again these two locations you see a stack of some of the proven VMware technologies like vSphere NS6 and you name them. Then there is the biolayer which establish infrastructure as a service then we consume it on or not consume we produce pass level services on top of it by the means of an open shift we have a user and deliver of that and then we produce really the SaaS application on top to ensure that actually the business can can serve the trend. We have a similar but a little extended environment which more focus on previously explained the shopping experience which is in the meantime much more than one or two data centers in the meantime extended to public cloud because this is a gross transaction is such heavy and exponential that single physical data centers are not any more capable of serving it. So we have a hybrid cloud strategy which very much inter-corporate pieces of various public clouds and of course big pieces of private cloud produce on our own and together in a kind of federated fashion overlaid with a unifying pass layer that's serving the customer needs or the needs of the customers of our customers. My more detailed view on this when we when we look on a public cloud piece is some kind of a region construct where you have in all of regions availability zones and the availability zones you have you have what we call pot a kind of bundle of various components including something like couch base for synchronizing the data including all sorts of nabacanching systems which which calculate what is the best way from A to B and so on and so forth and there is a central place where you have the master data which get replicated over to multiple locations in a kind of onion ring of multiple caching engines and all that together deliver from various places the transaction and the data the customers needs and basically that piece is glued together or put together establish what we call a globally distributed amadeus and it's initiative where we outgrown our own data centers and really embrace public clouds to establish a hybrid cloud production for us out of various locations technologies and so on and for most of what you see we use or we focus on using open technology with dance and terraform and some other orchestration tools we rely heavily on open technology and comes to our own data centers but want to combine it not to build just another cycle but combine it in a way that we can transform the kind of infrastructure of us to a cloud experience for our consumers and that's actually a sweet spot when we speak about what Martin was presenting and we selected and I was not there when he was selected something like three and a half years ago selected the viral product really as a clue between the traditional IT world which which works yeah you know how it works yeah and the new world of DevOps and RSE teams focusing more on the productivity about the CICP pipeline and development and so on and to glue these things together and there are challenges with that but there are some advantages and one of the challenge is that actually it's hard to find skills and people who understand both the MP ecosystem and the opens tech ecosystem at once and so if one of you are these blessed individuals yeah please come after the speech to me I can offer you a good job it's one of the challenges and the advantages is that we actually succeed to transform some of the operation teams and some of the investments we've made in have a licensing infrastructure in processes and you can imagine that such highly transactional system it's really high SLA and proper change controls and and proper security we are PCI DSS audited environment because we process credit card data and so on and so forth so security is one of the critical pieces to maintain and deliver where we have definitely a sweet spot compared to some other technologies because there are already a long time in the enterprise making its own journey so so to have a clue in between these words and hopefully match them in an effective way and that's probably a challenge when not only Amadeus is facing but most of you and there are of course other challenges like things like upgrades and so on and we are in a heavy process to upgrade some of our some of our legacy yards environment which is still yards but which was at some point in time frozen to to non jeopardize huge business invocations onboarding of customers so we need to revamp that and maybe in half a year I might have a good dropout story to tell how how it worked to upgrade the kind of four digit host calls to queens and so on that the last releases which we've deployed on the newer clouds of us because we have in the meantime multiple of them in multiple locations are very promising so we grow exponentially so and that's basically what I wanted to explain feel free to a question it's not a sales pitch it's not for me so feel free to ask and I'm happy to have a proper peer exchange reasons why I came here okay first question did you describe the production environments only or also the same do you use the same stack for development and integration environments yeah that's a tricky question so so the use the use in some of our deaf environments as well as other flavor of opens nevertheless as we progress on the on the stages and we have something like 11 phases between a few deaf on a on a desktop up to production to ensure the code quality integration itself is maintained there is a cut between okay what's what's a just enough deaf focusing and what's already a QA and then integration or even user acceptance test or similar and that line is drawn very much around how can we reduce the PCI gss code when they want some security and flexibility needs and it's not about technology and so so I haven't shown here the deaf cloud but it's of course part of the story come on guys questions comments remarks I hope it was interesting and I hope if some of you are on the same journey like we are we can have a peer exchange help each other with experiences not to get caught in the same types and we from Veeam we're also happy to talk to you then that's the whole thing I unfortunately sales but I do think that that the use case that that julius sanat we're trying to simplify stuff it's not only about the plumbing so it's not only about a compute storage and networking it's in the end and hopefully that you've seen that in the presentation that julius gave it's all about the application and what we're trying to provide from a Veeam my perspective is that solid base where you can host that application on top of and as you've seen it's not only a Veeam or only stack because this is a combination of open shift together with vio from VMware on top of a software defined data center and of course there's also the challenge of the multitude of multiple clouds that you then need to manage and that's the key value that we want to send out from a VMware perspective vio for us is a simplification of running open stack on top of your vSphere environment to be able to do what you're trying to achieve that's a little bit of the difference of the the thing that we're trying to provide to our customers and hopefully we've shown in this presentation the combination of a customer together with what we do from a product perspective that we're trying to simplify and make your life a little bit easier not so much that you can sit back and relax but in the end we need to provide services towards our customers in this case it's the developer building the application to do processing on top of that for the traveling industry the last thing i had on this one you can use over already showed it a little bit if you want to go and test this we've got a hands-on lab that you can do if you want to integrate it open stack and test it yourself within a half an hour one and a half hours two hours you will run through the product and hopefully you will see then that it's pretty easy to set up and that's the key value and again if we still have questions feel free to do that we'll be here and if you have any follow-ups please do so the next session is actually one of my colleagues and they're deep diving into what we do with via within NSX we have a networking perspective so if you're interested in that what we do with NSX please remain seated and they will show you that so thank you for attending this session if you got any questions feel free to ask them thank you