On The 9th of December, The KubeCon + CloudNativeCon + Open Source Summit China 2021, an annual technology feast for Chinese developers, will be held soon Will meet Chinese developers live online.

As the top event in the field of cloud native technology, KubeCon + CloudNativeCon + Open Source Summit China has brought together the most active Open Source cloud native community at home and abroad, the most advanced technology representatives and the best practices of the industry. Promote knowledge update and technology progress in cloud native computing. Ali Cloud, as the diamond sponsor of the conference, provided strong support for CNCF and was highly recognized by the organizers of the conference.

Since 2017, Alibaba has invested great strength in the field of cloud native technology, deeply participated in the development and maintenance of etCD, Kubernetes, ContainerD and other top open source projects, and completed the self-upgrade of the overall infrastructure system through the cloud native technology stack. By the end of 2020, Alibaba has entered CNCF with more than 10 projects such as KubeVela, OpenYurt, Fluid and OpenKruise. The amount of contribution to Kubernetes project is also among the top 10 in the world.

Can’t miss it! Cloud native innovation practice from 10+ Ali cloud technical experts

In this conference, first-line cloud native technology experts from Aliyun brought a wealth of speech topics, and more than 10 topics passed the strict selection of the organizers. The content covers cloud native application delivery, cloud native AI, Kuberentes cluster management, container runtime, CNI, fault monitoring, Serverless and other cloud native segmentation technology fields, and has excellent performance in terms of the number of topics selected and topic richness.

The following are the wonderful sharing topics brought by Aliyun at this conference:

###z Keynote | Cloud future, new possibilities


Li Yi (Mark), ali Cloud senior technical expert, Ali Cloud container service research and development director

Brief Introduction:

In the post-COVID-19 era, digital technology is fully integrated into our daily life and playing an indispensable role. Green development and industrial upgrading have become new potential for economic growth. Cloud native technology is the driving force of enterprise digital transformation, and is also becoming the engine of business innovation.

Session | Ten topics, wonderful sharing

Topic 1:

Build and Manage multi-cluster applications with Consistent Experience – Yong Feng & Jianbo Sun, Alibaba

Summary: The current user experience of developing applications in a local environment is significantly different from deploying and managing applications in a multi-cluster environment. Moving a well-tested application to a multi-cluster environment requires a lot of effort. In this lesson, we will demonstrate how to combine KubeVela and OCM (Open Cluster Management) to solve problems in Ali Cloud. Users need only define one application and another markup language (YAML) according to the KubeVela application specification, and KubeVela and open cluster management will take care of the rest. KubeVela will prepare a deployment workflow, including creating Kubernetes clusters as needed. Open cluster management helps to register clusters and allocate resources between clusters. As a result, the user experience for deploying and managing applications is consistent between local development and multi-cluster production deployments. We’ll discuss these use cases, challenges, and related work and experience.

Topic 2:

Deep Dive into TAG App Delivery – Hongchao Deng, Alibaba & Thomas Schuetz Dynatrace

Summary: TAG Application Delivery focuses on simplifying application delivery on Kubernetes and improving the developer experience. This discussion will showcase the latest developments around operators, chaos engineering, application delivery methods, and demo applications. Whether you are just starting to learn about cloud native application delivery or want to keep up to date, this discussion will provide you with information on ongoing activities, latest developments, and selected trends based on feedback from CNCF projects in the application delivery space.

Issue 3:

How We Discover and Locate Kubernetes Cluster Problems Before Users at Alibaba-peng Nanguang, Alibaba

Overview: The ability to quickly discover and locate problems is the cornerstone of rapid system recovery. Only by first identifying and locating the problem quickly can we discuss how to solve the problem and minimize user losses. So how do we find and locate problems in front of users in complex, large-scale scenarios? I will describe some of our experiences and practices in quickly finding and locating problems in managing large Kubernetes clusters — how we solved the problems we encountered by creating a common link detection + directional detection tool, KubeProbe, to address the stability challenges of large clusters.

Issue 4:

EROFS, what are we currently doing for containers? EROFS, What Are We Doing Now For Containers? – Xiang Gao & Xuyang Ge, Alibaba Cloud

Summary: The EROFS file system is a Linux read-only file system designed to save additional storage space while maintaining end-to-end performance. The file system has been upgraded since Linux 4.19. Previously, this file system was primarily used for embedded devices, but we found that there was a pressing need for read-only high-performance file systems for container use cases as well. This topic details the new features we are developing for this new scenario.

Issue 5:

How To Migrate Kubernetes Cluster With Zero Downtime – Jing Gu & Yaoyao Xie, Alibaba

Summary: Upgrading Kuberentes across multiple versions is riskier. Many customers choose to use cluster migration (that is, create a new higher-version cluster and then migrate the application from a lower-version cluster to a higher-version cluster) rather than upgrading the cluster. However, how to migrate clusters with zero downtime has become a major challenge. This sharing suggests a solution to the problem.

Issue 6:

Best Practice: Observability and Root Cause Diagnosis of DNS Failures in Kubernetes Clusters DNS Failure Observability and Diagnosis in Kubernetes – Yuning Xie, Alibaba

Description: In aliyun tens of thousands of Kubernetes cluster, DNS domain name resolution failure is one of the most common problems. DNS resolution failures occur in a variety of ways. Some are intermittent, some persistent, some affect all types of domain name queries, and some only a small number. The root causes vary, mostly due to container network problems and sometimes configuration errors. Yuning Xie will introduce the visibility and root cause diagnosis of DNS resolution failures in Kubernetes. This topic will cover the following topics:

  1. This section describes common DNS resolution faults in Kubernetes

  2. Introduce CoreDNS built-in observability plug-ins, such as the log/errors/trace/dump/metrics

  3. How to diagnose CoreDNS exceptions using DNSTAP instead of tcpdump and other traditional methods with high overhead

  4. How to Diagnose the root cause of DNS exceptions on the client based on BPF


Issue 7:

Fluid: Kubernetes native Distributed dataset coordinator and Accelerator Kubernetes Native Distributed Dataset Orchestrator and Accelerator – Yang Che, Alibaba & Yuandong Xie, Tencent

Summary: Using Kubernetes to run big data and artificial intelligence applications on the public cloud is a new trend. However, accessing data in cloud decomposition environments such as S3, Google Cloud Storage, and Hadoop Distributed File System (HDFS) greatly challenges training performance and limits the scalability of computing. Yang Che and Chris will talk about Fluid and how it can improve the performance of Kubernetes big data and ARTIFICIAL intelligence applications. Fluid is an open source, community-driven, highly collaborative project delivered by university engineers and researchers. It is hosted by the Cloud Native Computing Foundation (CNCF) as a sandbox project. In this course, participants will learn about project motivations, architecture, latest features, and user adoption. After that, Yang Che and Chris will present the project roadmap and how new contributors can get involved, and present a demonstration of accelerating serverless computing through Fluid.

Issue 8:

Extend the operations for Container Runtime in OpenKruise – Siyu Wang, Alibaba

Summary: In general, one can only use plain old data as the smallest unit of operation in Kubernetes. They can create a plain old data, but cannot control containers in plain old data, nor can they extract images on certain nodes through the Kubernetes application interface. This is because Kubernetes does not provide a runtime API for operations such as Containerd or Docker. As FAR as I know, some companies have hacked Kubelet’s code in their clusters so that they can do more with containers. However, extending operations for the runtime is the wrong approach because it is not conducive to open source and community collaboration. OpenKruise, one of the Cloud Native Computing Foundation sandbox projects, now offers advanced capabilities to operate the container runtime in each of the original Kubernetes clusters. It supports large-scale image pre-downloading and container restart via CRD, so users can do this simply by applying CR’s other markup language. In this talk, we’ll look at the use of the functionality in OpenKruise and how it works with Kubelet and CRI.

Topic 9:

Realize FaaS + Dapr + Kubernetes technology fusion, Achieving FaaS + Dapr + Kubernetes Technology Integration to build a new Alibaba Cloud serverless R&D system and build a new Alibaba Cloud Serverless R&D system – Zhao Qingjie, Alibaba

Summary: Kubernetes has become a cloud operating system as the cloud native ecosystem continues to evolve. At the same time, there is increasing talk that serverless is the next decade of cloud computing, so how will Kubernetes better support this serverless scenario? How to use the Kubernetes ecosystem to better integrate the two? Serverless solutions based on Kubernetes such as Fission and Kubernetes have emerged in the industry, but they are still unable to break through technical bottlenecks such as zero restore and high-density + high-frequency creation. The discussion focused on how to achieve technological breakthroughs within Alibaba and how to achieve mass migration of existing applications.

FaaS bottlenecks on Kubernetes. How to break the node bottleneck and realize the high-density deployment of 1200 instances on a single node. How to break the performance bottleneck and achieve the high-frequency creation and deletion of 1W +/5s. How to reuse the ecological capabilities of Kubernetes and extend the boundaries of FaaS. How to quickly and massively migrate existing serverless applications?

Issue 10:

Overview of CNI 1.0.0 and preview of CNI 2.0 – Bruce Ma, Ant Financial & Bingshen Wang, Alibaba

Summary: CNI 1.0.0 is the first stable release recently, which means that the specifications and features will remain somewhat unchanged for the time being and will be widely available. This topic will review in detail the key features and enhancements in the evolution of CNI 1.0.0 to help you use CNI in a more rational and disciplined manner. In the second part, we will delve into the future of CNI 2.0 and some of the problems and shortcomings of CNI 1.0, including security, real-time network state, plug-in binary execution, configuration management, and so on. In addition, we hope this topic will attract more CNI 2.0 contributors.

Together with Ali Cloud and CNCF, to build China cloud native ecology

For more than 10 years, as a pioneer in the field of cloud computing in China, Ali Cloud has always paid attention to its experience in large-scale cloud native application practice, through products, open source projects and technology output to contribute to the vast number of enterprises and developers, and effectively promote the education and training of cloud native talents.

In August 2021, at the global Open Source Technology Summit GOTC site, Ali Cloud and Linux Open Source Software School jointly announced that they would jointly upgrade the “Cloud native Talent Training Plan 2.0” with Mage Education, a well-known IT education institution in the industry, to gather the ecological force of open source. Help cloud native developers get the systematic growth from theoretical basis, open source practice to enterprise application implementation.

In the future, Ali Cloud will continue to work with CNCF to promote the development of cloud native technology in China, and welcome more enterprises, open source enthusiasts and cloud native developers to become partners of CNCF and Ali Cloud to build a beautiful open source community.

The agenda for the KubeCon + CloudNativeCon + Open Source Summit China 2021 is now available

For more information about cloud native, please scan the qr code below or search wechat (AlibabaCloud888) to add cloud native assistant! Enter the group to get primary cloud information ~