Evolution of Alibaba's Application Operations System

At Velocity China 2016, Alibaba researcher Lin Hao (aka Bi Xuan) presented the evolution of Alibaba's application operations system, describing how the company progressed through distinct phases as its business grew and technology changed.

The early 2008‑2009 "script era" relied heavily on manual scripts for deployment, scaling, and maintenance, but increasing complexity soon exposed the limits of pure scripting.

In the subsequent "tool era," Alibaba introduced dedicated tool teams alongside traditional ops teams, later consolidating them to improve tool quality, unify architectures, and address coordination problems between tool developers and operators.

Recognizing the need for a software‑engineering mindset, Alibaba embraced DevOps principles, eventually dissolving the centralized application ops team and embedding operational responsibilities within each business unit's development teams.

Key challenges encountered included tool quality, fragmented standards across services, difficulty achieving high success rates, stability, performance at massive scale (e.g., Double‑11 traffic), and the need for robust automation to reduce manual intervention.

The speaker highlighted the influence of Google’s SRE model, the role of Docker and Dockerfiles in enforcing reproducible environments, and the importance of allocating sufficient developer time to operational tooling.

Automation was presented as a critical milestone, while intelligent operations were described as a future goal that depends on extensive, accurate data collection and machine‑learning‑driven feature extraction.

Overall, the presentation underscored that successful operations transformation requires unified tooling, strong software engineering practices, clear organizational structures, and a gradual move toward automated and intelligent systems.