 And that's really what it's all about. That's really what we're here to do. So we've got some great speakers this morning and a lot of great content through the rest of the week. And next up, please help me welcome from AT&T, Sorab Saxena. Thanks, Jonathan, and hello, Austin. I'm excited to share with you the cool things we've been up to at AT&T with the help of OpenStack and you. Today, I'll walk you through AT&T's cloud journey in four sections. First, I'll share a challenge we face and our answer to the challenge. Secondly, I'll describe the AT&T integrated cloud, which I'll refer to from here on as AIC. It's important and some of our key achievements. Thirdly, I'll demonstrate a use case of communication service running on a cloud platform in production to show how it all comes together for us. Finally, I want to extend a call to action for the entire OpenStack community. So let's get started. The challenge. The rapid enhancements and advancements in communication services have resulted in numerous benefits for all of us, the users, and has in fact spawned many new industries. All of this hyperactivity has created explosive growth in network traffic. And just to give you some factoids, from 2007, our mobile data traffic grew 150,000%. Another data point, on an average day, 114 petabytes of data traverses our network. And if that isn't staggering enough, we expect our total network traffic to grow 10 times between now and 2020. The economic gravity of this reality says that we must transform our approach to building networks. And our answer to the challenge is to transition from purpose-built network appliances to open white box commodity hardware that is virtualized and controlled by AIC. We're also liberating the network functions from the same purpose-built appliances into standalone software components and managing the full lifecycle of this virtualized network infrastructure via AIC working in conjunction with both local and global controllers. Taking this approach prevents vendor lock-in and allows us to have an open, flexible, modular architecture that serves the business purposes off, scaling to meet the explosive growth at lower cost, increasing speed of feature delivery, and providing much greater agility. Our goal is to virtualize and cloud-enable 75% of our target network architecture using the software-centric approach by 2020. So let's dig further into AIC, a critical component of our answer. I'll use AIC acronym to describe three of its very key characteristics. A, its AT&T's global and distributed cloud. To achieve our 75% goal, we will need nearly 1,000 AIC zones across the globe. I, its integrated. It is AIC has a single code base for both enterprise and carrier grade workloads. C, it's obviously a cloud platform, but more importantly, OpenStack sits at AIC score, and that has catapulted us forward in our cloud journey. To give you a few more details, today, the AIC production platform is already leveraging 10 OpenStack projects, and before the end of the year, we'll be adding at least three more. The rapid evolution of OpenStack towards providing a full set of capabilities, both for cloud operators and as Jonathan touched upon, even for developers, is something we fully embrace in AIC and at AT&T. For example, in our latest AIC release, we introduced a new big tent project, Morano. This community innovation is a critical enabler in automating virtual function tenants onboarding across multiple AIC zones. Another example is Fuel. It is a new addition to OpenStack. However, we have fully adopted an industrialized fuel to automate the deployment of AIC zones. To bring to life these three key characteristics, and many more successfully, our team had to overcome numerous challenges. We needed to deploy OpenStack-enabled AIC clouds in days, not months, and do so at massive scale across the globe, support multiple site types that had very different implementations and tenant demands, designed for frequent hardware augmentation and software upgrades, and finally, such a large scale cloud platform demanded unique management and operational capabilities. Let me now play a video for you to show how our team automated the design, build, and manage phases. Automation was needed in all phases of AIC zone delivery. Design, deployment, and at scale management. In the past, cloud infrastructure design was manual. Excel spreadsheets with subnets, cabling diagrams, and rack elevations were emailed between teams. A single data entry mistake could translate into days of lost engineering time during deployment. Our cloud infrastructure design tool addresses our need for speed given AIC scale by automating the design of zones and in doing so, cutting the time for this step in half. These complex designs are subsequently consumed by our deployment automation. Next, we faced a chicken and egg problem with no existing solution. We needed to virtualize OpenStack, enable our control plane to be dynamic, and reduce its physical footprint. But OpenStack was not installed to help. Enter the virtual control plane. We succeeded in building and automating an under cloud that's entirely virtualized. So our OpenStack control plane runs inside virtual machines that we can easily move around, snapshot, roll back, and redeploy or upgrade. The subtraction layer allows the control plane to be the same everywhere, regardless of whether we're deploying in a single rack within a network office or a data center with hundreds of compute hosts. To automate the build of the zone, we take the cloud infrastructure design and feed that directly into a touchless seed mechanism that reliably bootstraps the zone and builds the virtual control plane foundations. Finally, our automation deploys services like monitoring, inventory, and security with OpenStack itself installed and configured by fuel. We selected fuel to set at the center of our OpenStack automation, but it would need to grow up fast to meet our service provider needs. Support for OpenStack controllers on multiple racks and network segments, just one of the significant enhancements developed to enable at scale deployments. Once built, we needed to manage AIC at scale. We are engineering day two operations into our fuel and automation framework. Capabilities ranging from minor setting changes to deploying new plugins and enabling new features across hundreds of AIC zones. So now that you have some insight of the solution in the design, build, and manage phases, let me share an example of an innovative solution that led to solving many complex operational challenges. On one hand, AT&T's massive scale needs independent cloud zones for the purpose of performance, resiliency, and security. But on the other hand, we would have an operational nightmare without a centralized management framework. To solve this, we developed the OpenStack Resource Manager, or ORM. Think of ORM as providing two key functions, a resource creation gateway, and a region discovery service. The resource creation gateway function provides APIs that allow the definition of new images, flavors, and user accounts, as well as the distribution of these to the required AIC cloud zones. For example, if a cloud operator wanted to create an image for one of many network functions and distributed only to a select set of AIC zones, ORM simply tags several zones with a group name, specifies the group name as the target, and then distributes the images to the group. Moving on to the region discovery service, it returns zones that are best suited to support the tenant's needs at the time of request. For instance, if you want to deploy a virtual network function for a mobile phone service workload, we can ask ORM to filter, down-select, and present back the zones that meet the criteria. Think of ORM exactly like an online hotel reservation system, hiding the complexities of thousands of zones behind a single pane of glass. Next, let's talk about the importance of integrating AIC with our large technology ecosystem. Before we enabled services and managed purpose-built appliances with many vendor-provided element management systems and homegrown business and operation support systems. In the software-defined networking world, these capabilities didn't exist since we are in uncharted territory. Hence, we needed to innovate and innovate fast to find new solutions. Let me discuss two key gaps and the associated solutions. The first gap was within the data center. We needed a local controller to build carrier-grade network functions. We worked with other open-source communities to develop the local controller and integrate it with Neutron APIs. The second gap was that we needed capabilities to enable and manage a plethora of very complex high-end communication services across the wide area network and integrate them with existing systems. Hence, we developed a new framework and a global controller called ECOMP, which stands for Enhanced Control Orchestration Management and Policy. I know it's a mouthful, so let me go back to the acronym. ECOMP provides a design time and execution time framework that tightly integrates all of our systems with the AIC Cloud as well as enables the wide area network. ECOMP and AIC are essentially our answer to making Cloud successful for service providers. Now, let me show you another video of a network-on-demand service using our AIC Cloud in production. Network-on-demand is a new service that we're creating in AT&T, giving the customer the flexibility to change services on demand. Customers today are reacting to an environment that's constantly changing, and they need their network to keep up with them. Today, if you look at what we basically put on a customer prem, you've got multiple devices, right, starting from the router that goes on the customer premise. If customer wants a firewall, that's another hardware device, if they want an application acceleration, that's another device, so it could have four or five or even more different devices. In today's environment, the customer has to phone in changes, and those changes often take weeks, if not longer, to implement. With Network-on-demand, we're giving the customer the actual interface into the network to make those changes in real-time. We've started with software-defined networking, which is the ability to really program the network. We want to go away from a static model of offering enterprise services. Using the new software capabilities available, we want to give the customer access to a portal or through an API, where they can dynamically make changes to existing services or add new services. With Network-on-demand, it takes weeks and months down to close to real-time in order to provision networks. The next thing is network function virtualization, which is our ability to abstract the software from the hardware, and once we do that, we can spin that up in any location in the network. Our Network-on-demand-managed internet service leverages our OpenStack-based cloud to virtualize network routers at AT&T sites. These virtual functions allow for the provisioning and management of networking components in minutes versus months. Bottom line, open-source software is the key driver for maximum speed and agility. OpenStack is a huge driver of this transformation. Our OpenStack cloud includes local control planes that are deployed near our customer sites to meet their service latency needs. These control planes or zones have automation templates and APIs that enable our developers to provide customers with the control to manage their AT&T network services. As we deliver new features in our cloud zones, every developer can reuse those capabilities to improve the services that AT&T offers. Where today we have a static environment in place that gets configured, when Network-on-demand are really putting the keys back into customers' hands. So as you can see, on Network-on-demand, it's powered by AIC and it's putting the controls in the hand of our end customers to summarize, I hope I've been able to communicate two key messages. One, at AT&T, we are fully committed to OpenStack and have been for a while. That's why we've chosen OpenStack to be the foundation upon which our software-defined network stands. And John Donovan, our Chief Strategy Officer, has repeatedly spoken of AT&T's commitment to open-source and specifically OpenStack. Second, we realized early on that making AIC and in turn OpenStack successful at AT&T meant the following. We needed a thoughtful approach to the end-to-end architecture. We needed innovation to match what our architectural principles demanded, like distributed yet centrally managed AIC zones. We needed automation of critical functions. One example, before automating deployments, it took nearly 10 months to deploy 20 OpenStack zones. With automation, we were able to deploy 54 AIC zones in less than two months. Well, thank you for that. We also needed integration with various platforms in our large ecosystem. Solving across all these dimensions collectively has proven to be a successful formula for making OpenStack fully operational in a large ecosystem like ours. We think that replicating this formula can make OpenStack the de facto standard for private clouds of any size. So finally and most importantly, our call to action. We invite everyone to collaborate within the community structure across all projects to expediently solve for large scale operator and service provider needs. This will help us achieve our common goal of making OpenStack the standard for private clouds. Thank you for your time and enjoy the summit. Thank you, Saurabh. Some of those stats that you put up at the beginning were really impressive. I mean, 150,000% increase in the video. I was talking about disruption earlier. I think that's an example of something that's pretty disruptive and we were talking about how this is something that your whole industry is facing yesterday and how do you guys approach that? I think Jonathan changes inevitable and we ironically enough are causing the reasons for the disruption in our industry, which is the success of all the devices and all the traffic that these devices generate into the network. So we realized early on that you have two choices. You can lead the disruption and be in the driver's seat or be the passenger and we absolutely choose to be the disruptor and be in the driver's seat. So I think your talk on disruption was right on. Well, thank you very much for telling your story and AT&T has a number of other sessions this week where you guys talk about different aspects of it. So thanks for joining us and be sure to look up there other sessions. Take care, Jonathan, thank you.