[MUSIC] Hello everyone, I'm Liishan I'm Lead Solution Architect Cloud Intelligence. Today I will introduce Alibaba Cloud the native solution for you. The content, including industry, trending and challenge container, so its portfolio deep tile for container service. Finally, I will show you some use cases. As we can see, the evolutionary path from organization to infrastructure as code to containers to solace. So now we are in containers and solar sparrow. By 2022, more than 75% of global organizations will be running containerized application in production. So we know in the further employers Cloud Native Applications will become the norm. Actually, there have some common challenges during Cloud native transformation. How to use elasticity to saving costs and increase resource utilization. How to use a serviceless cluster running applications when posting workloads. How to monitor communities cluster? Easy analysis application log. And observer cluster event. Yeah, centralized event center. Currently, how more and more enterprise using hybrid Cloud and market Cloud. How to easy management, hybrid Cloud or multi Cloud environment? Above challenging can be solving by Alibaba Cloud. Let's take a look the solution. First of all, let's see container service portfolio. We provide to measure product ACK and ASK the shop named Alibaba Cloud container service for communities and Alibaba Cloud Service community service. SDK and ASK symbolize integrated with Alibaba Cloud products. Make you easy to use a community's cluster or Alibaba Cloud. And the ACK and ASK they integrated miss computing network and storage. You could create ECS, EBM, ECI as communities worker node. You could create a community's cluster here VPC define IP around made me switch. Assign port IP from ENI, exposed service by SRB. All kinds of storage you're using as a community's persistent volume. The storage included EBS, mass, CPFS, and OSS. Upon ACK and ASK we provide security to make your communties cluster and application runtime ECQ, observed your communities cluster and monitoring your application. For elasticity, we make you easy skill up or scroll down your communities cluster. We also have multi Cloud and hybrid Cloud management solution. If you want to use distilled or kinetic, you could easily install at all from console, or when you create a ASK you could directly deploy key native serving an escaped without installing any component. Sandbox containers which is container running engine. It provide isolation container runtime. ACR big change container registry. It can start docker image and ham chat. ASM, which is manage the service match. It make you. Easy using service mash, read that install and config anything. Based on those ability, all kinds of application can be running on ACK and ASK such as Microservice Application, Stateful Middleware, even Big Data and AI, Blockchain and IoT. So you could using ACK and ASK directly from public cloud or they can deploy in private cloud or they can running or edge computing. This is the newest report from Gartner this year. They are competitive Alibaba, Data Blast, Google, IBM, Microsoft and Oracle. And then says from six criteria so we can see ACK Harris products technique layout. With sunrise, the ACK advantage with four points. The first one is Rich Products, ACK have managed and dedicate cluster of available in 19 regions. ASK which has a solar container running on elastic container instance. ASM which is still base to manage the service mash. ACR which is private container registry. It can automatically synchronize images, cross region worldwide. It includes image scanning with delivery chain. It may also have hybrid cloud on the market cloud the solution allows you to manage external communities cluster outside Alibaba Cloud. The second one is Large-Scale practice. ACK used by Alibaba groups large e-commerce platform double 11. This demonstrates the robustness of ACK Technologies. The third one is Extremely Elasticity. Application can be simply auto-scale in second skill hub 500 nodes take 2.5 minutes in standard mode, take 92nd inspect mode, just take 32nd on ECA. Supports its PA, CrowdSPA, VPA with multiple dimension external metrics, rather the only scaled by CPU and memory metrics. Supports much more instance type includes the ECS elastic by mental GPU instance, spot instance, elastic container instance. The last command is Extreme Performance, container around you are elastic by mental improve performance by 20% than physical machine. Double 11, is the world's largest online shopping shopping label festival from Alibaba Group. We have 2 Million Containers Running on elastic by mental. We also have cloud native database we called polo DB. It provides great performance. It comes to 100 TV data and so many link ups. For Applications Reliability will lariat extreme elasticity and to add stress testing flow control, graceful degrees, monitoring and not fighting to achieve peak order creation per second of 544,000. For DevOps Efficiency, may have accelerated image distribution over than 10,000 nodes to make pulling image be stable and faster, enlarge cluster. For Security part, it can protect it attacks from Internet. It's also can detect risk from inside container. They also have much more innovation from peak data, IPI and which is running on container as well to increase resource utilization. May I add a member in some open source community? May I initiative member of open application model? Application team member elected as a member of C&C app, Technical Oversight Committee. Next, let's go through the tile for container service. Then they create a cluster. How my paper plates become twice. The standard manage the cluster, or heterogeneous computing cluster. He left it by mental cluster at cluster. Windows cluster, or standard that came cluster? Educate heterogeneous cluster. Also less cluster. All confidential computing cluster. So, what's the difference of dedicate managed and Soulless Cluster? But that's Kate communities. The Masternode and the worker node a managed by user. The manage the communities cluster. The masternode is managed by Alibaba Cloud, and the worker node is managed by user. For so nice communities cluster, the Masternode. And them worker node are managed by Alibaba Cloud side. For managed and dedicate communities cluster, we can see the port running or ECS or elastic by mental, but it can simply scale up too easy I. The scenario for application keep persistent running. It's supposed spotty stench, GPU, FPGA distance. For some less qualities, we can see the port running an ECA elastic container instance. The scenario for the application, which is a temporary running for example, data processing, CI CD pipeline. Or keep, like right now. It also support spot instance and GPU instance. I see our Wichita Container Registry Service, it can manage both Docker image and Ham Chad. It's provide unto and security. I love you pulling image with me PC. Ain't no end point, and you can define to mean at the network ACL. It also type scale under ability during build image. So, it how accelerated image distribution based on Big P2P Technology? In large cluster, to make pulling image be stable, it also how image replication crossed different region in worldwide. For stories all kinds of applications stories can be using. Miss communities persistent volume, they each started feeding for different scenario. The Clock disk is good for stateful application that fish. The mass is fitting for log shared data. I said it's good for media, genetic data. CPFS is good for HPC deep learning. Local disk, it's feeding for did face and Hadoop, and all kinds of stories. His support and to an encryption with customer master key and bring your own key. Clap Disk and local disk support online resizing, and metrics can be monitored. For OSS, we support distribution, catch acceleration for machine learning. For network, we provide two tag of network, flannel, and to be the flannel is easy to use. Pour the leverage VPC to communication attorneys support network policy. And the performance is better than flannel. In the Terway network the port connection with ENI elastic network interface. Terway is employment by Alibaba Cloud, it also support exclusive ENI mode. In that mode, the network performance do not have any consumption. For security, we provide end to end cloud native security. In your 1st starter layer we support fine green algorithm for port. For Software layer, we provide image scanning and image signing, DevSecOps to make software be secure during supply chain. For runtime layer we provide, Network policy, KMS, runtime defense to protect application runtime be secure. For elasticity, we provide two level scaling, resource scaling and application scanning. For Resource scaling it can be scaled up or scaled down, ECS, EDM, GPU based Instance, Spot Instance or Virtual Node. For application scanning we support HPA, VPA and Cron HPA. For HPA part, it can be triggered by external metrics such as QPS, website latency, etc. We know observability have three pillars. Alibaba Cloud have some services mapping for logging metrics and tracing. All those services are similar integrated with ACK as well to make you easy observed your communities cluster and monitoring your application. For AI and big data scenario, we provide heterogeneous computing, make you simply build CPU, GPU, FPGA, RDMA and store it for machine learning and big data, fundamentally environment. Support major framework such as TensorFlow, MxNET, Cafee etc, based on Kubeflow. For increase resource utilization, we optimize the communities in scheduling strategy to make pods be tight on GPU Node. We also have shared GPU ability and we provide GPU isolated. They also provide Co-scheduling and capacity scheduling strategy and the Co-scheduling divide part by Alibaba Cloud and contributed to community. The GPU and shared GPU utilization can be monitored in Prometheus. On top of that, we provide our arena, which is a command line tools to make you easy to run machine learning job. Arena initial divide power by Alibaba Cloud and contributed to Kubeflow. Hybrid-Cloud and the Multi-Cloud stretching from our experience, the three type of solution are popular. The first one is Active/Active. The multi cluster running on Cloud and IDC respectively, and Larry's Federation to manage the multi cluster. CI/CD platform deploy application to multi cluster and using GTM to manage global traffic. The second one is auto scale to Alibaba Cloud, the manager application running on my PC went. Application needs scale, it CAP, auto scale into Alibaba Cloud, both services its registered to the same macro service registration and the discovery. The third one is ACK Multi Cloud management. Communities clustering, IDC or other The Cloud provider can be registered to ACK. ACK Multi Cloud Management be able to manage cluster as the same as ACK cluster. The Register cluster could be using Log Service or ARMS or AHAS Pretty good middle-wares to logging and monitoring the cluster or application. This is an example for Cloud Native HA Application Architect. The communities cluster diploid market, so using Polla DB as they face. Lab reason, lab sewings, prometheus and arms yo observe application and communities cluster. Next that I will show you some user case. This is a part of our customers around the world like cement Olympic panels, Dana or investing. This customer was build Alibaba Cloud service with everything as cold? The red color is created by powerful. The blue color is created by hand chat. So I everything as cold, reduced or prison time and avoid. I will happen by human carrots. This camera this customer is running micro service on Alibaba Cloud. So the macro service is running on a CK. And expose API the API gateway using math and he did us to protect website. Some power starting OSS also using lock service arms, Prometheus to observe communities cluster and monitoring application. This customer is from Mormonism to microservices. The macroservice divide a CK and some felt starting OSS using Redis. Micycle Paula Debbie and 90 DB to store data and analysis data. The service can be accessed by website and the Moon, Pow Nobel application. The customer side, by leveraging the solution proposed by Alibaba Cloud, we completed the react texture from Wallace to microservices. Shut hinder his time of app development and deployment. Reduced IP and resource expenditure and improve reliability and agility. This Katima build ALAIOT on Alibaba Cloud. From collector sensor data yo transfer data to date management to build the data middle layer exposing machine learning model through API gateway. Some services running on a CK. About content is about a Alibaba Cloud native solution we talk about and Alibaba Cloud Container Service, Port Portfolio cloud native stories cloud unit him network cloud init him security auto scaling strategy and to an optimal ability. Cloud, native AI and big data., hybrid cloud a multi cloud solution. So thank you for your time. Thank you and handy that this session.