Brief introduction:On June 26, 2021, the theme event “Edge X Kubernetes” co-sponsored by Ali Cloud, VMware and Intel was held in Hangzhou. Focusing on “Cloud Native Edge Computing”, a collaboration between the OpenYurt community and the Edgex Foundry community, 12 contributors from a variety of industries and scenarios, together with key members of two open source projects, will discuss the practice and application of cloud native on the edge.



On June 26, 2021, the theme event “Edge X Kubernetes” co-sponsored by Ali Cloud, VMware and Intel was held in Hangzhou. Focusing on “Cloud Native Edge Computing”, a collaboration between the OpenYurt community and the Edgex Foundry community, 12 contributors from a variety of industries and scenarios, together with key members of two open source projects, will discuss the practice and application of cloud native on the edge.



Before the event officially kicked off, Jim White, CTO of IOT Systems and co-founder of Edgex Foundry TSC Chair &, addressed the event in a video message. We also look forward to more developers joining the Edgex Foundry and OpenYurt communities to help drive the native rollout of the Edge Cloud.



Gim White,OTech Systems CTO

EdgeX Foundry TSC Chair

Then, Lu Guang, co-founder and maintainer of Edgex Foundry China and CTO of VMware China, and Sun Min, Senior Business Development Manager of Intel IoT, gave a brief introduction to the current situation and development of Edgex China.



Lu Guang, Director of VMware China R&D Center,

Edgex Foundry China Co-founder & Core Maintainer



Min Sun, Senior Business Development Manager, Intel IoT Division

Edgex Foundry China Co-founder & Core Maintainer

Focusing on the fusion problem of cloud native and edge computing, in order to help developers solve the delivery, operation and maintenance, control and other challenges in large-scale application scenarios, four representatives from the OpenYurt community shared the landing challenges of edge cloud native under the wave of cloud native technology. And the OpenYurt community based on their observations, reflections and future plans during the practice.

highlights

Wong Yuki: Continue to deepen the open collaboration between OpenYurt and the Edgex Foundry and other communities to build the standard of edge cloud native facts

Yuqi Huang, the co-founder and core leader of OpenYurt community and senior technical expert of Aliyun, introduced the development of Aliyun in the original direction of Edge Cloud, as well as the background, current situation and planning of OpenYurt project and community.



Huang Yuqi is a senior technology expert at Alibaba

Co-founder & core leader of the OpenYurt community

Since AliCloud was formally established in 2009, Alibaba has accumulated, deposited and practiced in the field of cloud native for more than 10 years since the development of Feitian operating system and the beginning of the era of cloud. Now it has developed into one of the core technology strategies of AliCloud, and continues to explore how to extend the ability of cloud native to more industries and scenarios.

In 2019, Ali Cloud officially released the edge container service ACK@Edge, which is committed to helping enterprises achieve the integration and collaboration of “cloud edge” in the business under the Pan Edge scenario through non-intrusive enhancement. In order to promote the common development of edge computing and cloud native fields, in May 2020, Aliyun Cloud native team officially opened the ACK@Edge core framework OpenYurt, which has accumulated in dozens of industries, and donated it to CNCF. In September of the same year, it became a CNCF Sandbox project.





As a K8S-based edge computing cloud native intelligence platform project, OpenYurt was officially accepted as a Sandbox project by CNCF in September 2020, due to the concept of “non-intrusive enhancement” pioneered by the industry and more friendly support of cloud native ecological capacity, compared with the same type of open source project. At present, OpenYurt has cooperated with several edge computing industry, university and research institutions such as VMware, Intel, Shenconvinced, China MerchantBureau, Zhejiang University and so on to carry out the technology co-construction and application exploration based on OpenYurt, and is committed to developing it into the actual standard of edge computing cloud native direction.



In the future, OpenYurt will continue to work with communities like the Edgex Foundry, focusing on the following areas:

  • We will work together to create IoT device twinning and cloud native standard of IoT model to reach more developers in the Internet of Things industry and promote cloud native landing
  • Provides edge cloud resource interface capability standards
  • Continue to improve the hardware and software integration delivery model
  • Allied middleware vendors push traditional middleware capabilities to the sinking edge

He Linbo: Why is OpenYurt the preferred project for Edge Cloud native ecological compatibility scenarios



He Linbo is OpenYurt community leader and AliCloud technology expert

He Linbo, head of OpenYurt community and expert of AliCloud technology, first showed the case of a CDN user in the community who realized the transformation of cloud native architecture based on OpenYurt, and then expounded the impact on cloud service providers, cluster users and edge computing practitioners. What value does OpenYurt bring to users of different dimensions, such as cloud native technology enthusiasts?

In the face of the key challenges native K8S faces in the cloud edge integration scenario, what are OpenYurt’s thoughts on architecture design that are different from similar solutions in the industry, so as to truly realize the above values and help enterprises to build non-intrusive edge computing cloud native infrastructure:

Adhering to the design principle of non-intrusive native Kubernetes and “cloud-side collaborative architecture”, OpenYurt forms a three-layer cloud-side collaborative architecture based on K8S. Through core components such as Yurthub, Yurt-Tunnel, and Yurt-App-Manager, It solves the problems of edge autonomy oriented to weak network connection, cloud edge efficient collaborative network construction, and cross-regional large-scale application management, so as to realize the centralized management and control of cloud (central end) and moderate autonomy of edge.

At the end of the sharing, He extended his welcome to the new Members of the community and OpenYurt users. He also thanked the community contributors for their support of OpenYurt since it was opened to the public. He hopes that the partnership with Edgex Foundry will influence more developers to invest in both communities in the future. To promote the synergistic development of cloud native and edge computing to make a contribution.

Li Zhuhao, Xu Xiaohu. Application of OpenYurt in Urban Brain and Industrial Brain

At the end of the activity, Li Zhuhao and Xu Xiaohu, experts from Aliyun Intelligent Manufacturing Team, respectively introduced the application practice of OpenYurt in urban brain and industrial brain projects, as well as the existing pain points of using cloud native technology in the field of AI.



Li Zhuhao, Ali cloud technology expert

Li JuHao introduced that in the actual promotion of the city brain, it is obvious that although the container technology, especially Kubernetes, has been widely used and provides better resource utilization efficiency, the following problems still exist when K8S is used to solve problems in the AI field:

  • The native scheduling function of K8S cannot meet the training and reasoning requirements in the field of AI
  • In terms of data management, there is a lack of computing side data caching capability
  • Lack of time-sharing and low utilization rate in resource management
  • The ability of hardware isomerization is weak
  • GPU resource allocation and automatic scaling based on video analysis path number and QPS
  • Stateful routing, call forwarding, and custom load balancing

Xu Xiaohu also shared that in the process of the implementation of Industrial Brain products, “the pain of traditional control optimization scheme” is a common pain point for customers in this field, such as:

  • High service costs: high delivery and operation costs require a lot of on-site service and human experience
  • Limited technical ability: focusing on single point control ability, poor global optimization, unable to do automatic update operation and maintenance
  • The effect is difficult to last: after a period of time, the control performance is reduced due to the aging of the device and technical transformation



Xu Xiaohu, senior technical expert of Aliyun

Lee ZhuHao, cancel the xushuguan separately under each scenario is introduced based on an ACK @ Edge using OpenYurt build cloud collaboration architecture, by raising the technical advantage, make the Edge class business in the cloud, unified management and monitoring at the same time realize the Edge autonomy, make the Edge network even in the condition of without Internet and cloud, still support Edge business continues to run, When the node is restarted, the business container can still recover normally, ensuring continuous, stable and available business.



If you have any questions or would like to learn more about Openyurt, please join the community exchange group using the Scan QR Code:



Click https://developer.aliyun.com/live/246906, to go back to watch the event live.

Copyright Notice:The content of this article is contributed by Aliyun real-name registered users, and the copyright belongs to the original author. Aliyun developer community does not own the copyright and does not bear the corresponding legal liability. For specific rules, please refer to User Service Agreement of Alibaba Cloud Developer Community and Guidance on Intellectual Property Protection of Alibaba Cloud Developer Community. If you find any suspected plagiarism in the community, fill in the infringement complaint form to report, once verified, the community will immediately delete the suspected infringing content.