How Alibaba’s ALPD Transforms Cloud‑Native DevOps: A 5‑Stage Journey

What is Cloud‑Native DevOps?

Cloud‑native DevOps fully leverages cloud‑native infrastructure, is based on microservice/serverless architectures and open standards, is language‑agnostic, and provides continuous delivery and intelligent self‑operation, delivering higher service quality and lower development‑ops cost while letting developers focus on rapid business iteration.

Analogy

A traditional DevOps scenario is likened to a single chef handling everything from ingredient procurement to cooking and sales, achieving high efficiency only when the chef is highly skilled, but scaling is difficult due to non‑standard processes.

The "Nanjing street stall" analogy shows a model where chefs can focus on improving dishes, experiment with new items, and quickly adapt to user demand, illustrating cloud‑native DevOps.

Core Principles and Capabilities

Cloud‑native DevOps follows two principles—open standards and language‑agnosticism—and rests on two foundations: microservice/serverless architecture and Serverless infrastructure (BaaS/FaaS). It delivers two capabilities: intelligent self‑operation and continuous delivery.

Open standards and language‑agnosticism give higher flexibility and ecosystem longevity.

Microservice and serverless foundations enable DevOps.

These principles and foundations enable continuous delivery and intelligent self‑operation.

Alibaba Cloud‑Native DevOps Upgrade Cases

1. Architecture Upgrade – Sidecar and Mesh

The service‑governance code is moved out of the application container into a sidecar, and a service mesh handles routing, logging, monitoring, and other governance functions, creating a “rich container” that can be upgraded independently of the application.

This decoupling lets business code focus on core functionality without being tied to governance concerns, and the migration can be performed gradually.

2. Decoupling Build, Release, and Ops

Three levels of decoupling are introduced: build, release, and operations. By containerizing each business component and using init containers, builds become independent. Services are classified by intimacy (process, IPC, RPC) and gradually split into independent services for separate release and ops.

Ultra‑close: same process, function calls.

Close: different containers in the same Pod, IPC.

Loose: separate network, RPC.

3. IaC & GitOps

Infrastructure as Code (IaC) repositories store both container images and configuration. Changes are pushed as code; a GitOps engine detects IaC changes, translates them to OAM‑compatible configurations, and applies them automatically, providing full traceability and repeatable releases.

4. Resource BaaS

Resources are described declaratively in IaC, enabling intelligent management, on‑demand usage, and migration‑friendly standard protocols, reducing operational overhead.

ALPD – Next‑Generation Lean Product Development Methodology

ALPD provides a systematic framework for cloud‑native DevOps adoption, integrating tools, processes, and best practices into a complete DevOps toolchain.

Roles and Responsibilities

Technical leaders/architects define overall R&D governance, architecture, cloud product usage, release strategies, and monitoring.

Engineers (dev, test, ops) use the Cloud‑Effect platform to code, build, integrate, test, and deploy automatically, achieving zero‑touch pipelines.

Five‑Stage Cloud‑Native DevOps Migration Path

Fully manual delivery and ops – no service‑oriented architecture, no cloud infrastructure, no CI/CD.

Tool‑assisted delivery and ops – introduce microservices, GitLab/Jenkins, basic CI for single modules.

Limited continuous delivery and automated ops – containerize infrastructure, adopt full toolchain, enable continuous deployment with some manual steps.

Continuous delivery with assisted self‑ops – adopt serverless, achieve near‑zero‑touch deployment, business‑level observability, partial self‑ops.

Full‑link continuous delivery and self‑ops – all services and infrastructure are serverless, fully automated release, rollback, and self‑operation.

While details are complex, leveraging Alibaba Cloud‑Effect and ALPD consulting helps enterprises avoid pitfalls and accelerate the journey to cloud‑native DevOps.