How Alibaba Handled Real‑Time Billions of Events During Double 11

Alibaba Large-Scale Data Computing and Processing Platform

Speaker: Lin Wei, Senior Technical Expert, Alibaba Cloud

Speaker Bio: Senior architect of Alibaba Cloud big data computing platform, responsible for overall architecture planning, design, and optimization, with over ten years of experience in large distributed systems.

Talk Overview: The Double 11 (Singles' Day) event posed two main technical challenges for Alibaba Cloud's big data platform: real‑time data processing and massive offline data processing. Real‑time challenges included ingesting, distributing, and computing log and transaction data to render live statistics on media screens with stringent latency, stability, and throughput requirements. In 2016, Alibaba Cloud's real‑time big data system achieved three world‑class milestones: sub‑second end‑to‑end latency from the first transaction to display, processing speeds up to tens of millions of records per second (performance increased multiple folds over the previous year), and full‑day high availability without degradation. Offline, the MaxCompute service handled daily peaks exceeding hundreds of petabytes and millions of scheduled jobs, testing scheduling, compute performance, and system stability.

For more videos and PPT materials, follow Alibaba Technology public account.