 So, I actually, I just had a question that he was talking about applications being the primary interface for his users. And you know, I think this is something that you see everywhere now. What is this? We got stuck in that roundabout at the Arc Day Tree on, so I'm really sorry we missed your keynote, man. See, who do we have? We have Alan Clark. I tried to get these stickers here as fast as I could. And gentle, Monty Taylor. I see. Getting the car with Mark again. So, you guys are a little late, a little late, but glad you made it. I heard your keynote was, you crushed it, man. I don't know. We were stuck. So, could you have allay parked that for us, please? Oh, sure, yeah. So, all right. Thanks, man. Thank you, Jonathan. If you could wash it too, that'd be great. I'll wash it too. So, we have a car on stage here. What is this? This is not just any car. This happens to be a BMW i8 supercar. And you might be wondering why we have this car here. We've been talking about innovation and really driving innovation, allowing companies to do amazing new things. And honestly, this is one of the most innovative pieces of machinery that has ever been created. And I can't think of a better physical manifestation of innovation than this right here. And we are also very fortunate to have the company responsible for this car here to talk with us today. So, please help me welcome the man responsible for the data centers at BMW Group, Dr. Stefan Lentz. Some may call this a concept car. It's lightweight, futuristic, and intelligent, but this is not a new car. This is every car we've ever built. Every BMW is created with efficient dynamics. It's our way of doing more with less, more power, less consumption. Joy is future proof. Good morning. It's great to be here, especially to be here with that car on stage. They are incredibly hard to get. And the OpenStack people sort of managed to get one on stage, which is wonderful. This car is everything we as BMW want to be. We want to be innovative. We want to be elegant. We want to be resource efficient. And of course, we want to be joy to drive with. And that's very similar to OpenStack. The story that I have brought you here is our journey from our conventional infrastructure to a private cloud and the role that we think OpenStack is going to take in that journey. Just briefly, who are we and what are we talking about? BMW's data center is a global data center. We run about 12,000 US images, half of them Linux and half of them Windows. We are over 60% virtualized plus we have about 20% of the servers running multi-project Oracle and web servers. So 80% or more of our infrastructure is cloud ready. It is an infrastructure as a service. It is a platform as a service minus the cloud. This data center is global and when I say global, it means we run every server, 98% of all servers worldwide out of operations team in Munich. When you are a project in BMW and you want a server deployed for like your application to servers in Munich, one in Spartanburg and one in Australia, you get the same server and the same infrastructure, exactly the same platform. We standardized brutally, I must say, brutally over the last couple of years our complete infrastructure basis and why, of course, we wanted to be cost efficient. Why can we run things from Munich, from Germany, from a high cost country and the reason is we automated so much that we say automation of our infrastructure. We are now about 25% more cost efficient than the best of our competitors in delivering infrastructure services. About three years ago we found out that this path of standardization of automation on a conventional level has sort of reached its peak or its end. We cannot become more cost efficient with conventional methods than we are now. And we will lose the advantages that we gain if we don't do something new. And the new thing that we saw on the market is cloud. So approximately like 2011 on, we thought of bringing all this automation, all this thing under one framework. And guess what we did? We developed a piece of software internally, our own, in Python, modular with a database and we called it our internal private cloud. And it was wonderful. Work perfect. I mean, we really got into something, we got new efficiencies, everything was great. And then we ran into trouble. Things wouldn't go like we thought, quality problems, all kinds of things. I'm not going to tell you why we got into problems now. I saved this for later. Because there's another story that happened approximately the same time when we first discovered that conventional automation and virtualization would not help us anymore getting more efficient and better for the company. I became a manager 10 years ago. Before that I was a system administrator, software developer, software architect, all kind of practical jobs where you do real work. Then became a manager. And as a manager, you get used to one thing and this is really interesting. People come to you and present you facts on slides. So they come with the slides and the slides are always the same. I mean, they contain bars and arrows and sometimes round shapes. And management buzzwords on one side. And the more you are in management and the more people forget that you had been the technical guy earlier, the more intense this gets. And I always ask people, I mean, if you come to me for a presentation, please give me facts, give me information, give me something I can work with. And it's always the bars and the arrows and the round shapes. And I had on OpenStack in particular on cloud management, I had approximately 20 or 25 management presentation over the year 2012 and 2013. And one of the company really did it. I mean, I was set in this presentation. I had told them specifically what I wanted to learn on OpenStack. And I had told them specifically that I want to have some idea how we can use it for our internal cloud initiative. And again, it was just these bars, round shapes and arrows and all that. And I said, this is it. I told to my assistant, no one will get any more time with me on a presentation on cloud management. And then the one room apartment was born. The one room apartment is one old server, seven or eight years old, that we got out of more or less garbage bin of our data center. And I installed Icehouse myself just to see what it was really all about. And in our next management workshop, I told the guys, listen, you have to reserve three hours for me because in these three hours, I will tell you what cloud is really all about and what's going wrong with our internal cloud project. And they reluctantly did this. After half an hour, the first guy said, why do we all get so technically? And I was just talking about what we have today and virtualization. So it really was a little bit difficult. But in the end, they all listened. And the sandbox, the sandbox, this one room apartment, one room apartment because everything, you know, like compute node, networking, control node, everything is in this one server. The sandbox helped us to understand cloud and to change direction. By the way, it was Ubuntu based, which does not mean that BMW has adopted Ubuntu now. That's just my personal favorite. Sorry, Susie. What do we see in OpenStack? After the sandbox, we immediately decided to build a semi-productive environment. So we have approximately like 16 terabyte of storage and 100 cores in a semi-productive environment. Semi-productive means we work with it in developing our own data center method on it, and we work with it to give certain user groups access to a cloud environment so that they can develop their things on that cloud environment. And that's the current status, 100 cores, something nice and small and reasonable. We see in OpenStack two major, two big advantages. First, we have an API and a data model to describe cloud and virtual instances that will become industry standard. So when anyone develops on that, it will be stable. We do not have to change whole tool chains. That's the one thing we see, big advantage, and that's the point where we failed because we were unable with our internal cloud with our own development to get that stable and good. And the second thing we see is it is open source and free. No one earns money on the growth of our company. No one comes with un-expected license fees after we have developed a lot of software around them, and we have experienced that a lot. So that's the reason why we think that OpenStack is the right choice as a core for our automation. We are not going to put really highly productive workloads on this at the moment, but this is just a statement for at the moment. We have online looks, virtualization, all the SAP, all the database, all the production control environments that built this. And this is a 60 seconds clock that ticks. Every 60 seconds your IT doesn't run. One car is not built. So we have that already on openXen, on SUSE, on Linux, and it works perfect. I think OpenStack can be part of that tool chain, but we have to do more development on that and bring more thought in that. We are conservative people. We want to build cars and not primarily develop software. The thing that bothers us most in OpenStack currently is the release cycles and the huge changes that we have from release to release. So if you are developers, then the thing I just want to give you is one message. Industry grade, we need more stability still in the future, but that doesn't prevent us to use it just right now as it is. Thank you. Well, Stefan, thank you very much for giving us that overview of what BMW is doing. Should we go park this somewhere? Let's park it. All right.