How Alibaba Evolved Its Application Operations: From Scripts to DevOps
Alibaba’s application operations journey, detailed by researcher Lin Hao, traces the shift from early script‑based practices through tool‑centric phases to a full DevOps transformation, highlighting challenges, automation efforts, and the emerging push toward intelligent, data‑driven operations.
On December 1‑2, 2016, at Velocity China 2016, Alibaba Platform Architecture researcher Lin Hao (nickname "Bi Xuan") delivered a talk titled "Evolution of Alibaba's Application Operations System". He described how Alibaba's operations have progressed through distinct stages as the business grew, scale expanded, and industry technologies evolved.
He identified three major eras: the script era (around 2008‑2009) where most operational tasks were performed manually via scripts; the tool era , during which dedicated tool teams were created to develop software systems that encapsulated complex operations; and the current DevOps era , where application operations teams have been merged into business‑unit development teams to fully practice DevOps principles.
During the script era, large‑scale releases required dozens of manual scripts, leading to difficulties in tracking progress across multiple data centers. The tool era introduced separate tool and operations teams, but quality issues, fragmented architectures, and lack of unified standards limited effectiveness. Alibaba responded by consolidating tool teams into a single, software‑engineered group, improving quality and consistency.
Despite these improvements, challenges remained: ensuring high success rates for massive releases (e.g., Double‑11 peak traffic), achieving stability and performance at scale, and standardizing environments across diverse services. The company also explored automation to reduce manual intervention, but high reliability requirements and architectural constraints made fully unattended deployments difficult.
Looking ahead, Alibaba is pursuing intelligent operations . Intelligent ops require extensive, accurate data collection and machine‑learning‑driven feature extraction. Until automation and data quality are sufficient, intelligent automation remains limited to narrow scenarios such as automatic failover in a single data center.
Overall, Lin Hao emphasized that DevOps is the strategic direction for Alibaba's operations, supported by organizational restructuring, containerization (e.g., Docker), and a focus on building robust, automated, and eventually intelligent operational platforms.
Speech Full Text
大家好,我是毕玄,我今天分享内容的主题叫阿里应用运维体系的演变……(full transcript omitted for brevity)
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
Alibaba Cloud Developer
Alibaba's official tech channel, featuring all of its technology innovations.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
