On September 26, Alibaba Cloud Cloud Native Open Day was successfully held during KubeCon. This event invited a number of Alibaba Cloud technical experts to work with open source community partners and enterprise users to comprehensively introduce new technology trends in the cloud native field, including K8s, microservices, Serverless, observables, message queues, etc. , new progress in projects, new implementations in practice, and new explorations in cutting-edge areas.
Scan the poster in the article to view the live broadcast replay.
Follow the public account and reply: 0926 information
You can get the PPT collection of lecturers from this open day.
Let’s review the exciting topics at this event.
Sharing topic: Practical sharing of container network troubleshooting in production environment under intelligent operation and maintenance
Guests: Wang Bingshen and Li Yutong
Topic introduction: Container intelligent operation and maintenance service is a series of inspection and diagnostic tool platforms for container clusters, which greatly saves the operation and maintenance energy of container clusters. A large number of tools and capabilities are also included for cluster network issues. This speech comprehensively introduces the capabilities of intelligent operation and maintenance service network operation and maintenance.
Sharing topic: Shangyou Network’s game cloud-native practice based on OpenKruise Game
Sharing guests: Hu Liangzhuang and Liu Qiuyang
Topic Introduction: Shangyou Network was founded in 2007. It has independently developed and operated a number of high-quality games, and completed the cloud-native transformation of games in its early years. In this speech, Shangyou Network combined its own game cloud-native experience to bring the application of its microservice architecture in games and the practice of OpenKruiseGame when the game server was implemented on Kubernetes, providing new ideas for the cloud-native implementation of the game server .
Sharing topic: Technical analysis and optimization of accelerated AIGC model inference service based on Fluid
Sharing guest: Xu Zhihao
Topic Introduction: As a data orchestration framework in cloud-native AI scenarios, the Fluid project has done a lot of optimization and exploration work in the engineering implementation of AIGC model reasoning. On the one hand, Fluid helps users simplify cache management and operation and maintenance of cloud-native AI scenarios through reasonable K8s abstraction; on the other hand, Fluid uses elastic distributed cache to flexibly adapt to AIGC business scenarios and reduce resource costs. In addition, Fluid additionally optimizes the efficiency of reading model data by the inference service and accelerates the model loading process. It has been proven that Fluid can increase the startup speed of the inference service for an LLM model of nearly 80GB by nearly 8 times.
Sharing topic: Fei Rui Interactive uses Higress as Kubernetes Ingress to improve stability and performance
Sharing guest: Zhang Tianyi
Introduction to the topic: Fei Rui Interactive built a zero-trust architecture through MSE Higress. It can migrate smoothly without reconfiguring Nginx Ingress rules. The performance is improved by 90%, the response time is reduced by 50%, and the stability and security of the business entrance are greatly improved, and efficient support is provided. 100 million+ fan interactions every day, 40,000+ offline stores, and 30 million+ mobile payment needs every month.
Sharing topic: Building the next generation of intelligent observable systems based on eBPF
Sharing guest: Dong Shandong and Liu Kai
Topic Introduction: Firstly, the advantages of K8s and the challenges brought to the observable field are introduced. Secondly, the data collection scheme in the K8s scenario is introduced. Finally, the practice of troubleshooting network faults in the K8s monitoring scenario is introduced. More importantly, we also summarized the systematic construction thinking of intelligent root cause analysis in K8s scenario: dimensional attribution, anomaly delimitation and fault tree analysis. By building these three core capabilities, we can quickly locate abnormal nodes with problems in the K8s scenario and provide fault classification and root causes.
Sharing topic: Minimalist Serverless R&D platform for small and medium-sized enterprises and developers
Sharing guest: Zhao Qingjie
Topic Introduction: The new version of Serverless Application Engine SAE is aimed at business-driven small and medium-sized enterprises and R&D engineers of startup companies. It provides a one-stop application creation, minimalist interaction, and highly automated integration Serverless application platform, breaking technical barriers and helping users to implement simple applications . Serverless transformation helps businesses reduce costs and increase efficiency.
Share topic: The road to serverlessization in Apache RocketMQ 5.0
Sharing guest: Jin Jixiang
Topic Introduction: RocketMQ 5.0 closely follows the new concept of Serverless cloud and implements the elastic Serverless architecture, fully demonstrating the advancement of the technical architecture and greatly improving the sense of security on the user side. The expected benefits of the implementation of RockeMQ Serverless architecture are: worry-free - extremely fast elasticity, no need for capacity planning; money-saving - pay for what you use, use as needed; labor-saving - less operation and maintenance, no need for operation and maintenance.
Direct access to the event site
