Keynote: Cloud Native at Alibaba Scale
summary
Alibaba runs Kubernetes clusters of tens of thousands of nodes and manages more than ten thousand heterogeneous applications, with stress-tested staging-to-production upgrade pipelines and workload-driven performance tuning. Engineering delivers 200+ upstream improvements focused on etcd performance and API scaling, validates a 50k-node containerd deployment, and open sources Dragonfly plus a virtual API-server multi-tenancy approach. Operational tooling adds broadcast jobs, advanced StatefulSet rollouts, and cluster automation so upgrades and peak traffic (shopping-festival bursts) succeed with measurable stability gains.
AI-written summary · about
chapters (15)
- 0:00Intro
- 0:21Why Cloud Native
- 3:39始于Kubernetes
- 4:52Scalability
- 7:29Integration
- 8:53Kubernetes Management
- 11:11Cloud Native Strategy
- 13:07Cloud Native For Everyone
- 14:24Container Ecosystem
- 15:00Kubernetes Multi-Tenancy
- 16:58OpenKruise
- 18:32Sidecar Set
- 19:21Advanced Stateful Set
- 20:54Sharing
- 21:43Cloud native needs everyone
related talks
22mKeynote: Cloud Native at Alibaba Scale
06mKeynote: The Cloud Native Ecosystem in China
44mSome Lessons We Learned from Moving E-business Giant to Cloud Native
06mSponsored Keynote: Kubernetes is Now, Cloud Native is Future – Ding Yu
08mKeynote: Cloud Native in the Enterprise: Production, Best Practice and Innovation
17mKeynote: Tencent: Kubernetes in the Billions
07mKeynote: Exploring the Frontiers of Cloud Native
18mKeynote: Mind the Gap: Bridging Cloud Native Innovation with Real-World Use