Unlock Alibaba’s Cloud‑Native Secrets: 15 Essential Lectures

Container, service‑mesh, micro‑service, and Serverless technologies represent a new way to build applications, expanding the boundaries of cloud computing toward multi‑cloud, hybrid‑cloud, and edge‑cloud collaboration. Under this trend, many enterprises are evolving their business and technology toward cloud‑native.

During this evolution, companies often face challenges such as migrating applications to a Kubernetes‑based system and achieving continuous delivery. Alibaba has been practicing cloud‑native technologies since 2011, gradually building a containerized infrastructure that rivals leading global tech firms and serves the entire Alibaba Group.

From June 24‑26, 2019, at KubeCon + CloudNativeCon + Open Source Summit Shanghai, Alibaba shared 26 talks on practice, trends, and technology. Fifteen representative talks have been compiled into the book Alibaba Cloud‑Native Practice: 15 Lectures , aiming to reveal Alibaba’s cloud‑native journey and provide practical references for developers.

坚持探索与落地开源，阿里巴巴云原生之路全景揭秘

Alibaba Cloud has successfully scaled cloud‑native practices. This talk shares concrete Alibaba cloud‑native experiences covering scale, reliability, development efficiency, migration strategies, and optimizations for large‑scale scenarios.

1-5-10：如何快速恢复大规模容器故障

Introducing the 1‑5‑10 theory (MTTD = 1 min, MTTI = 5 min, MTTR = 10 min) for rapid recovery of large‑scale container failures.

How to build an effective local proxy to detect issues within 1 minute.

How to use an expert knowledge base for intelligent container diagnosis.

How to automate fault‑driven container recovery.

阿里巴巴利用 K8S、Kata 容器和裸机云构建无服务器

Discusses using Kata Containers as a secure runtime to achieve strong multi‑tenant isolation and high performance in serverless scenarios.

CafeDeployment：为互联网金融关键任务场景扩展的 Kubernetes 资源

Introduces the custom CRD CafeDeployment to address IP immutability, high availability, and flexible deployment strategies for financial workloads.

Severless 市场观察和落地挑战

Explores Serverless’s fine‑grained resource management, event‑driven model, and how it replaces resource‑allocation‑centric approaches.

有效可靠地管理大规模 Kubernetes 集群

Shares Ant Financial’s design of a large‑scale Kubernetes management system and its core components.

云原生应用 Kubernetes 监控与弹性实践

Shows how to integrate monitoring and elasticity into cloud‑native applications on Kubernetes via Cloud Provider, CRD Controllers, and Operators.

了解 Kubernetes Master 的可扩展性和性能

Analyzes performance bottlenecks when scaling Kubernetes beyond 5 k nodes and presents benchmark‑driven optimizations achieving >100× performance gains.

云原生时代加速镜像分发的三种方法

Describes three methods to improve image distribution efficiency, including P2P‑based CNCF/Dragonfly, remote snapshotting, and dynamic loading strategies.

在 Web 级集群中动态调整 Pod 资源限制

Discusses scientific resource management for massive container fleets, covering current community practices, Alibaba’s challenges, diagnosis methods, and balancing efficiency with stability.

大规模 K8s 集群下的巡检

Details Alibaba’s automated inspection and analysis of massive Kubernetes clusters, focusing on metric collection, visualization, and alerting.

使用 Istio 管理跨地域多集群的服务

Shows how Istio enables multi‑cluster service mesh for a global e‑learning platform, addressing traffic management and cross‑region latency.

首个普惠社区的平民化方案：GPU共享调度

Introduces a native Kubernetes solution for GPU sharing, covering API definition, scheduler integration without kernel changes, and TensorFlow multi‑job support.

容器运行时管理引擎 Containerd

Explains containerd’s plugin architecture, integration with gVisor and Firecracker, and best practices for secure, isolated container runtimes.

基于 P2P 原理的高可用高性能大规模镜像分发系统：Dragonfly

Describes the Dragonfly project, a P2P‑based image and file distribution system designed for cloud‑native scenarios, emphasizing simplicity, efficiency, intelligence, and security.