 Good morning, everyone. My name is Yin Xiong. I'm chief architect of the Cloud Platform at Huawei. Currently, I'm working on a project called PaaS, Platform Access Services, where I'm responsible for the architecture, the strategy, and the technology innovation of that project or platforms, where the container is a big part of it. So I'm honored to be here this morning. I want to thank the Linux Foundation to give us the opportunity to stand here to talk about, to share with you what we do with containers. Hold on. Sorry. What do we do with the container technology? How we think about the container ecosystems? Specifically today, I will talk a little bit about the container orchestration and managing platforms, how we can unleash the power of containers to our customers, to our partners. But first, for those who are new to Huawei, this is a brief introduction of the company. So we are a 29-year-old company founded in 1987 in Shenzhen, China. How many of you have been in Shenzhen? Good. Where is a great city? It has an economic called Silicon Valley of China. So I suggest you go if you have opportunities. So by the end of last year, year 2015, the company has 170,000 employees with sales revenue of more than 60 billion. What do we do? We provide the network and communication equipment. We build a network site, network tower for our telecom customers. We provide the storage and computer servers. We build the data centers, including the cloud data centers for all enterprise customers. We also provide the software solutions for our customers to manage the network, to manage the data center resources. We also have cell phones. Many of you may already have a Huawei cell phones. So we make the mobile device. We make the apps on top of the devices. Now Huawei has many, many R&D centers around the globe. Now one of the strategy we have is the investment to open source technology. We're really committed to open source communities. As you see, we are planting a member of NINUS Foundation, Open NFV, CNSafe. We are a gold member of OpenStack, Cloud Foundry, and many others. Now we do this not only because we believe the world is moving to the direction of open software solutions, but also because our customer demanded. And this is great news to us and great news to the communities. Now talk about the container. This is a container account, right? So we've got to talk about containers. Now I want to show you a few numbers regarding the container adoption in China. In the recent study, the 40% of a company in the survey say they are using the container in production. Well, this is not a big number. But when you combine this with the company that has container in depth or test and the company that planning to use the container technology within six months, that number is more than 80%. That's astonishing. What does it mean? To me, it means that we already passed a stage where people ask why, why container? I think the question is now when and how. And this is exciting to us and exciting to the open source communities. Another exciting number. The growth rate of the container adoption in production compared with last year, 2015, that number is up 250%. Increase, that number increased by 250%. That's an amazing number. Again, what it means to me, it means that the technology are mutual. The truth around the technology are mutual. And a big thanks to all of you who have worked hard and contributed to the community, contributed to the technology. You should be proud of yourself. Now our customer expect more from us, expect more from the community, and expect more from the open source communities and technology. Reason to use the container technology. There's no surprising here. The customer using the container for the benefits, this technology is designed for, such as agile development process, fast deployment, better resource utilization, better application portability, you name it. Let's look at another number. How companies manage the container in production. In this study, more than 50% of the company still manage the containers manually or write in simple scripts. Only 42% of the company use an orchestration platform, such as Kubernetes, Mezzos, or Docker, so on. You know we are in the generation of cloud, right? How come we still manage things in production manually or write in a simple script? This is not scalable. So we would like to see more company, more customer to use an orchestration platform. And we would like to see the number much bigger than 42%. So what is container technology or container orchestration? It turns out when you ask different people, they may give you different answers that focus on specific area of orchestration functions. I'm not trying to give an official definition of container orchestration, since we have a many experts here in this room. But simply that, given a connection of containers to deploy on the one side, and given a pool of resource, virtual or physical on the other side, the process to place these containers onto the servers, provisioning the storage, provisioning the network, keep them running at the desired stage, the whole process is called orchestration. So we have a term now, cluster management, scheduling, service discovery, provisioning. But those are not new terms. But when we apply to container, now we have innovation in the container ecosystems. Let's look at some of the orchestration function in detail and see how we innovate to bring the full power of container to the customers, to your customers, to the partners. Most likely your platform will run many different types of workload, right? Now with multi-scheduling framework, you can apply different scheduling upgrades to a different type of workload, such as non-running jobs, non-running applications, batch jobs, big data jobs. And that's the power of scheduling. Or simply you have one scheduler that target your resource utilization, because you care about your resource usage. Or you have one scheduler to talk about to target faster deployment, because you care about how quickly your application can start up. And that's the power of scheduling. You can do both. Reservation-based and SEO-based scheduling. Today, most of schedulers doing scheduling based on the CPU and the memory requirements. In some cases, however, in fact, in many cases, even the one of themself, don't know how much resource we need for our application. So we have divided with something called SEO-based scheduling. Where you tell the platform, you tell us what SEO you need for your application. The scheduler will figure out how much resource you need. And that's the power of scheduling. Another example in NFA scenarios, imagine that you container applications are actually virtualized network function. Where you place this container actually and you form a network topology. And that requires very different scheduling algorithms. In this scenario, in fact, you all define the network topology through the orchestration platform. And that's the power of scheduling. Cluster management, another key function of orchestration. Today, we have many research and development innovation in these areas, both industrial and academic. But just the simple fact that we can abstract a pool of resource, hide the differences of hardware, hide the differences of architecture, even operating systems, that make them transparent to our customer, transparent to your containers. And that's the power of cluster management. Additionally, you can share the same set of resource pool among different type of applications, but still isolate them without impacting each other. And that's the power of cluster management. If you want, you can steal some of the resource that already reserved for application, but has not been used much. You can steal this resource to temporary workload to increase your resource utilization. And that's our subscription. And that's the power of cluster management. Cluster federation. You hardly have one cluster in your production environment. With a project like Ublit for Kubernetes or Massos Federation, you can use the same API to deploy your containers to the multiple clusters, which may reside on different data centers, or even reside on different clouds, like AWS, public cloud, for OpenStack, private cloud. And that's the power of cluster federation, and that's the power of cluster management. Storage orchestrations. True, today we have many drivers or volume plugins to attach a container to different type of storage servers or storage services. But to unleash the full power of container, it would be nice that we can abstract storage resources, just like we abstract the CPU and the memory, so that we have a standard way to ask to provisioning the storage resources in a real time on demand. It would be nice that we have the ability to auto-scale the storage pools and auto-discover the storage services for your containers. If we can do that, that will be a power of storage orchestrations. Similarly for container networking, which is another key function for enterprise to use the container technology. And we have many network solutions there already. But again, to unleash the full power container, we think we need a common framework to plug in different type of network solutions on demand. We think we need a common way to configure your network policies in real time on demand to ensure the network security. We think we should have ability to schedule your containers based on network resource, because there are some applications that care about network bandwidth, network latency, then the CPU or memories. It would be nice if we have a schedule based on a network resource. Additionally, it would be nice that we can monitor the network resource consumption by your container and move your container around to achieve the application SOA or achieve your container SOA. And that would be great. And that's the power of the container networking. So put them all together, that forms our container strategies. This shows an end-to-end container technology stack. In addition to the container storage, the networking, the orchestration, we have a project to work in our container engine, container runtime, to make it secure, to change the corners, to make it secure containers. We have a project to work on container registry, like Docker Yard. We call it Docker Yard. We have a project working on the application model and deploying workflow, so that we can deploy the containers on different orchestration platforms, so we don't have a lock-in-one platform. We have a project working on the container depth of pipeline for agile development process. Again, we want to unleash the full power of container with the orchestration platform, with the orchestration and managing platforms. Lastly, but not least, I'm pleased to announce our first container services called Cloud Container Engine. CCE is based on the Kubernetes orchestration framework with advanced cluster management, advanced scheduling algorithm, enhanced securities, enhanced monitoring features, and much more. If you are interested in the CCEs, you can stop by at the Huawei's booth, and we'll be happy to talk to you for more information. That's it for me. Thank you very much for your time. I hope you enjoy the day and enjoy the rest of the conference. Thank you.