 Good morning, everybody. Thanks for coming. I'm really excited to be here. Some of you might know I'm actually from this area. I was growing up here and unfortunately, I needed to move away for business reasons that quite a while ago. So I'm particularly thankful that the Lennox Foundation brought up all these conferences to Düsseldorf this year. This is really exciting because there's an interesting fact I must have been missed while I was away. It seems that there has been a cultural revolution going on in Germany and now even in Düsseldorf people are wearing Lederhosen. That's a really exciting thing. So I'm quite amazed. Now, today I want to talk to you about the building blocks of Amazon Web Services and how this relates to open source. Obviously, Lennox and open source project played a big role in that. So a question I do get very often is how did an online bookseller end up a large cloud vendor? How did that happen? And to understand that you really need to go back to the mid-90s when Jeff Bezos founded the company and he was really after founding an online retail company rather than a cloud company. But you quickly realized in order to be successful in the internet business, you needed quite a sophisticated IT infrastructure and back in those days it meant spending a lot of capital buying stuff from a few well-known brands that offered proprietary and often very expensive solution stacks on top of that and that made it fairly hard to get started for a startup because most of your capital was sunk into IT infrastructure. And another problem was the fact that if you didn't use your IT equipment all the time, like in the case of Amazon, we have this online retail business and IT usage flocked to age subnavingly over the course of the day and even over the course of the year and when you're buying all these expensive things and realize that if you want to offer a great service to your customers, let me try to get this away here, you need to provision your IT infrastructure for the peak time plus some extra to be safe because you want to provision a good user experience and if it happens to be like in our case the Christmas season where everybody seems to be buying a lot for a few weeks and then the rest of the year your IT infrastructure sits idle largely, you are in a situation where it's not optimal. Plus the other thing was he was trying to sell books, not become an IT expert really, but to get started you needed to spend a lot of time and resources again in creating this IT infrastructure and that made it very hard to get started. So those were some of the lessons we learned. To address this idle infrastructure problem we thought about what can we do to offer this infrastructure to other people as well. So the idea of allowing other sellers to use our infrastructure was born. So we needed to create an infrastructure that we could use to safely host other sellers business on our platform and that was kind of the first service model that we ventured into. That took off quite a bit and we continued that. So the next target group we were focusing at were developers, software developers often have a similar problem because they need hardware resources, particular for testing and integration cycles for a short amount of time. Often it's hard to predict when they need it and once they're done, it's sitting idle. So the traditional IT resource acquisition cycles didn't work well for them. So with these groups we found already a few folks that really started to love this idea of the cloud, what we now call it. And Amazon took this further. So on March 14, 2006 we made this press release. We announced the first public service, S3, the simple storage service. This was our first public cloud service and we took it from there. A cloud service as we know it today, somewhat different from how we started. Obviously there are a few basic building blocks that you need. You want to be globally available. That means you need to have IT infrastructure in a bunch of regions. We currently have 10 regions work-wide and for reliability reasons you want to have fault isolation zones, which we called availability zones. We currently have 52 peering points that connect our infrastructure to the rest of the internet and then there is this storage block that I already talked about which was the first one. Once you offer storage services to the public you inevitably end up needing waste and control access. Security mechanisms, audit mechanisms are a very natural consequence of that. Once you store your data in the cloud you realize, wouldn't it be great if I can not only store it there, but also process it there? So this is when we started to develop EC2, the eclastic compute cloud service. So now you have your data in the cloud and you have your servers in the cloud. When that grows you quickly realize, oh, I need networking. I need a way to model my virtual data center like my original data center that I may or may not have had in the past and this is one of the most critical pieces to get right. Networking is often overlooked in the cloud business and this is a key technology to get right, especially if you want to future-prove it for scaling. You never know how big you grow and there are a number of challenges when you want to scale your networking infrastructure. Now what you see on screen right now is what many people consider a classical cloud business. Amazon didn't stop there. Remember the heart lesson that Jeff learned when he said I want to become an online bookseller? This, what you see there in an online book business, there's quite a big gap still. So what is really important is to have a lot of these standard services available that make your life easy and help you to apply your websites. So we started to add database services because it's not just a raw object storage. You want to have a high-level API that often is offered by databases. When you got this, you want to be able to analyze this data. So you need an analytic services and looking at typical web applications, you will see a lot of standard building blocks that keep repeating from application to application. So we added those app services. These are things like orchestration things, queuing services, and even basic things like email sending services or search for your data. Now a key audience that we always want to focus at is developers. So if you want to be successful, you want to have a very good development environment. So you make it easy to manage and deploy your things. So for the cloud, this means you want to have things like container technologies like we have with elastic beanstalk, but also DevOps is a key thing. So we developed something that's called Opsworks that allows you to manage your various deployments in an efficient way. And finally, if you're infrastructure growth, there are templating technologies like we have with Cloud Foundation that helps you to grow if you want to scale out, have patterns that you can replicate over regions or availability zones and scale up and down if needed. So about a few years ago, that was a pretty good view of what the cloud... Identity management across multiple devices and sync services to store and sync your data across all these devices became pretty much a mandate. You wanted to have push services and this drove another wave of innovation. Now, you could see that we're done now. We've helped many of our customers to move their data centers over. What's next? And the next thing we thought about was, if they want to focus all their energy on becoming experts at their core technology, at their product, offering services to their customer, running desktops is probably not part of this. So we developed the virtual desktop service where you can now run your desktops in the cloud and then use simple devices. This could be classical PCs with a web browser or tablets or even cell phones to access your desktop that is running in the cloud. This simplifies the maintenance of these desktops quite a bit and helps again our customers to focus their attention on the things that they need to excel at their business rather than becoming IT experts. Now, this is quite a sophisticated view of clouds now because it's not just as foundational services but you need to offer a lot more. So from 2006, when we started to 2014, we've seen quite an impressive growth of adding features and services to AWS. I don't really want to dwell into all of these services here, but one thing I want you to take away from here is if you look at the number of new features or substantial features on new services that we have launched each year, you will see slight acceleration is getting more and more and more each year and that seems to be continuing quite steadily. How's that possible? And this is what we call the AWS Innovation Flywheel. Let me walk you through this because it's a bit of a complex thing to understand. Let me start with one of the circles here and that's the innovation cycle. If you can innovate and in order to innovate, you need a few things. The ability to really modify every parameter of your system is a key and Linux and open source provides some very good basis to do that. So open source technology allowed us to do this. Remember we started out with these expensive proprietary systems, migrating them to standard off-the-shelf commodity servers, running Linux and open source software instead was a big step forward. It helped us to lower our infrastructure costs and that in turn allowed us to reduce prices. Now reducing prices means it attracts more customers and when you have more customers it will drive up usage and more usage means you have a growing ecosystem. It's not only your internal ecosystem that's growing, it's also the external ecosystem that's growing. People start to do things with your platform that you haven't thought about. It's just again this idea of allowing people to do something and then carry it forward into places that we haven't thought about. We're very much encouraging people to think about it and we're also spending energy and starting to contribute back to open source projects. Linux kernel, the Zen project are just a few examples of these and I'll touch those in a minute again. Now more AWS usage will inevitably require more infrastructure and more infrastructure increases your economy of scale. You can just buy stuff at lower cost and that closes this flywheel. You're back to lowering your infrastructure. We've done this already on the software side now. We've got the hardware aspects of this and that keeps this flywheel spinning. You've got this richer circle going on and on and on and it's not even just that it keeps rotating. It also can accelerate. Any progress you make on any of these circles will keep the flywheel accelerating and that's how for example we were able to reduce prices 45 times since we started which is a significant achievement and I expect this trend to continue. Now let's talk about the glue that sits between all these building blocks and this is an operating system. We call it Amazon Linux although there's not much special about it. It's really an assortion of open source components that we put together and use this to build key parts of the AWS infrastructure. Obviously we wanted to have a certain amount of control over it. What it comes to updates, security is a pretty important thing and I'll touch that in a minute. Stability is important but also freshness. We wanted to control the rhythm of how quickly we enter new stuff into this and that really drove the decision to try to come up with our own version of Linux and while this is the foundation of Amazon Web Services it's not something that we just sit on. It's actually something that we provide to everybody inside of EC2 or AWS at large. So if you're starting an EC2 instance or if you're using some of the services that run on top of EC2 you are using Amazon Linux. You have the option obviously while Amazon Linux is an operating system that we offer to use it's not the only choice you have. Actually we are very much about choice. If you choose to run another operating system feel free to do so. We're actively working with a bunch of partners that offer Linux distributions so that they work well inside of the Amazon ecosystem as well. So that's pretty much a key for us. We do not want to lock you in. We want to offer choice. Now I have security on this slide. Let me touch on this a little bit more. We've seen in recent weeks how important security is. Being able to control the ecosystem at your own pace can sometimes be really helpful to save your customers. Inside of AWS we use a shared security model. That means we as Amazon take care of all the infrastructure and all the software that we run. And we try as good as we can to keep this secure. But you have some responsibility as well as an EC2 customer because the stuff that you control, the things that you operate you need to take care of the security yourself. Security is not a state. It's a process. And the better your process is the better you have your security. Now we try to help you as good as we can and an example is Amazon Linux. It comes with a very small initial set of packages. So we reduce the attack surface. Access allowed and we provide security updates in a very timely manner and have various ways to inform you about this. So there's a regular cadence of releases to provide you the latest and greatest stuff as well. So security is a key feature of it. Stability is another one that I talked about. So in an enterprise environment you really don't want to change too many things. If you make changes you need to do a lot of testing to make sure you don't accidentally break things. Nevertheless we are living in a web world and one of the things that the web world told us is that innovation is key. You need to be able to innovate fast. And I want to use an example here to show how we treat innovation. Docker, who have you heard about Docker? That's quite a while. But I'm pretty sure you can hardly leave this conference without having heard about it again. I'm not the only one talking about it. It's literally across the program everywhere. So we started to introduce Docker in our spring release this year and we've just updated it with the fall release. We have two releases a year with Amazon Linux. So we did an upgrade to version 1.2 and we continue to provide upgrades. Now this is a technology that moves fast. For others we have established technologies. You don't want to lose what is working for you. So that's why we often provide multiple versions of the same thing. So for example within Amazon Linux you get multiple versions of Ruby or Python or other languages installed in parallel so you can use them at your choice. Now while Docker innovates we're going to keep updating it because it's a nascent technology that will see a lot of improvement over time. It's not something that we only offer to you. We also use this technology in our own stuff. So it's available if you want to run it inside an EC2 instance you can choose to do so. But we've also included technologies like Docker already in our higher level services. Elastic Beanstalk is our container service so you can use it inside there already. And then I already talked about Amazon Opsworks. So it's another way to use Docker if you want to use it at a slightly higher level rather than doing all the work yourself. Now let me come back to the fact that we are in Germany and touch some of the contributions that Amazon is doing to open source. About last year we started to increase our engagement with the open source community. We became a member of the Linux Foundation. We became a member of the Zen project and we continued that this year with joining the Core Infrastructure Initiative and we also thought about growing our global developer footprint. So here in Germany we opened up the Amazon Development Center in Germany and we have two locations in Dresden and Berlin. At Amazon each team owns usually their stuff from idea to deployment so we create these centers of excellence. So Berlin is the center of excellence for machine learning a very important technology that is seeing a lot of growth there. The Opsworks team is also located in Berlin. And in Dresden we have the team that is doing EC2 instance development and Amazon Linux development. That was quite challenging for us. We started off last year and in Berlin we're already building out a new site because the old one got too small. In Dresden we had a similar challenge. We had two floors that we started with and I recently opened up a search floor to create more space for developers. So we're really serious about getting engaged with the open source community and having more visibility. If you want to help us feel free to catch me in the hallway or any of my colleagues that are here this week with that. Thanks a lot for joining here and I'm happy to see you through the course of the week. Thanks a lot.