Evolution of Alibaba Operations: From Human/Script Ops to DataOps and AIOps

Alibaba senior operations expert Da Wu presents a retrospective of Alibaba's operations development, starting with the early Human/Script Ops era where engineers relied on manual actions and ad‑hoc scripts to deploy, troubleshoot, and respond to alerts.

As scale grew, repetitive scripts were consolidated into shared tools, marking the Tools Ops phase; this reduced inefficiencies, mitigated security risks, and laid the groundwork for early DevOps practices.

With massive business growth, the Platform Ops (DevOps) stage emerged, requiring global architecture planning, resource cost optimization, automated platform development, stability assurance, and massive data analysis to manage thousands of services and tens of thousands of machines.

The next transition is DataOps, where every operational object is digitized and managed through data pipelines, enabling data‑driven decision making, automated alert filtering, anomaly detection, and self‑healing mechanisms.

Building on DataOps, AIOps aims to integrate machine‑learning algorithms with big‑data‑based operation platforms to further automate monitoring, diagnosis, and remediation, though true full‑automation remains as challenging as autonomous driving.

To succeed, operations personnel must broaden their skill set, encompassing architecture design, software development, deep operational knowledge, fundamental engineering algorithms, and technical project management (TPM).

The shift from reactive firefighting to proactive, value‑driven, intelligent operations also demands cultural change, moving from experience‑based practices to engineering‑centric, data‑centric, and AI‑assisted workflows.