Shulou(Shulou.com)06/02 Report--

It took only a decade for cloud computing to grow into an unstoppable technology trend. Today, "cloud native" is regarded as the standard by enterprises and developers, and is regarded as the future of cloud computing.

Yi Li, head of Ali Cloud Container Technology, believes that cloud computing has three stages: cloud relocation, cloud readiness and cloud nativeness.

The first stage is to reduce the cost of moving to the cloud, migrating the applications running in the physical machine to the virtualized environment, and the application development and operation and maintenance methods are not very different.

In the second phase of cloud readiness, enterprises want to use cloud computing to improve overall efficiency, begin to experiment with micro-service architecture, and focus on metrics such as standardization, automation, scalability and high availability.

In the third stage, that is, at present, enterprises begin to fully embrace cloud computing, and many applications grow on the cloud from the very beginning, and agility, flexibility, fault tolerance, portability and other key features have also become key features. the entire life cycle of software architecture design, development, construction, delivery and so on has been reshaped.

At the first domestic KubeCon + CloudNativeCon 2018, cloud native application practice and landing have become the focus of the world's top technology enterprises and developers. In the keynote session of the conference, Aliyun researcher Ma Tao shared the current situation of cloud native ecosystem in China, and cloud native technology has been integrated into the public's food, clothing, housing and transportation.

Cdn.com/571d8987c33916c094cdaf8ccba9f3bcf9d2e706.png ">

Ma Tao, a researcher at Aliyun, delivered a keynote speech at the scene.

Talking about the Application practice of Cloud Native in China from Shuang 11

"Cloud native technology has been slowly integrated into the public's food, clothing, housing and transportation." Ma Tao said.

For example, during the period of 2018 Tmall double 11, Pouch and Aliyun container mirror service provided stable support for the majority of consumers to "buy and buy".

In general, before dealing with the traffic peak scenario, the traffic needs to be estimated in advance to prepare the calculated resources and application capacity, but only this step is not enough to support the traffic peak, and the ability of emergency capacity expansion is also extremely important. Facts have proved that the characteristics of container technology are very suitable for such scenarios and can stretch quickly and automatically when needed.

As early as 2011, Alibaba began to explore container technology. Double 11 in 2016 will containerize all core trading applications, with hundreds of thousands of containers supporting a peak of 175000 orders per second for double 11 transactions. Last year, Alibaba Group achieved full containment of online services and reached a million-level container deployment scale internally. Among them, double 11 transactions peak to 325000 / s, and can achieve rapid deployment of 10,000 servers within 10 minutes.

Up to now, the container image service supports the hosting of 100000 images of Alibaba Group and has accumulated 200 million image downloads.

Native cloud technology has been applied in various industries: ele.me uses TiKV to carry takeout orders that may soar at any time; Silan has built a floor-sweeping robot scheme based on Aliyun, Kubernetes and Harbor; and Weibo has built a deep learning and training framework to manage heterogeneous resources and AI task scheduling with Kubernetes.

Ali Yun experts and overseas experts

The methodology of Ali Yun Yun's original practice

The concept of cloud nativeness has a long history. It was first put forward by Matt Stine of Pivotal in 2015, but its definition has been vague for a long time in the industry.

This year, the CNCF Cloud Native Foundation gave a specific definition of cloud native: cloud native technology enables organizations to build and run resilient and scalable applications in new dynamic environments such as public, private, and hybrid clouds. Representative technologies of cloud natives include containers, service grids, microservices, immutable infrastructure, and declarative API. These technologies can build loosely coupled systems with good fault tolerance, easy to manage and easy to observe. Combined with reliable automation, cloud native technology enables engineers to easily make frequent and predictable major changes to the system.

If you build the top-level business abstraction at the containerized infrastructure and service governance layer, it requires an entire cloud native operating system. So cloud native is more of a set of technical system and methodology, and Aliyun container service is based on this concept.

Ali Cloud CCS Kubernetes (ACK) supports both public cloud, private cloud and hybrid cloud models, optimizing and integrating Ali Cloud's basic capabilities, making it easy for container applications to take advantage of the underlying powerful computing, storage, network and other capabilities.

For example, for deep learning and high-performance computing scenarios in pursuit of extreme efficiency:

You can use DPCA elastic bare metal servers or heterogeneous computing capabilities such as GPU and FPGA instances.

Using elastic network card with Aliyun Terway network driver can achieve 9Gb network bandwidth almost losslessly, or use RDMA protocol (RoCE) technology of converged Ethernet to 25Gb network.

It can also improve processing efficiency based on parallel file systems such as CPFS, providing up to 100 million IOPS and 1TBps throughput.

As for the service grid, it can be based on Aliyun Kubernetes as the infrastructure, combined with Istio service grid, supports multi-cloud and hybrid cloud, provides effective governance capabilities for micro-services, and provides different support for specific application loads based on Kubernetes and Istio, such as traffic management for Kubeflow services, routing management capabilities for Knative load, and so on.

Container has become a technology platform, which needs to meet the needs of various application scenarios, such as IoT, deep learning, block chain and so on. Yu Shan, a senior technical expert at Aliyun Block chain, believes that "there is no unified standard for blockchain business applications, which will involve various types of applications. Containers provide standard software packaging and distribution capabilities, achieve consistency of the running environment and decoupling from the underlying layer, and can give full play to the advantages of Kubernetes, such as resource scheduling, operation and maintenance management, security isolation, and micro service architecture."

In September 2017, Aliyun launched a container-based blockchain solution, which allows users to quickly build an enterprise-level blockchain environment through Kubernetes and Helm/Chart templates. In August this year, blockchain as a Service (Blockchain as a Service) was released, which supports one-click rapid deployment of blockchain environment in all aspects of system development, operation and maintenance, and realizes cross-enterprise and cross-regional blockchain applications.

Original text link https://yq.aliyun.com/articles/670758

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.