How Tencent’s Database Heroes Powered WeChat Red Packets, Honor of Kings, and Video Streaming

Inspired by the movie Inside Out , the author likens a database to the memory crystals that store a child's experiences, emphasizing that every digital interaction relies on a robust database.

CKV (later TDB) was born in 2006 when QQ’s user base surged to 20 million, forcing the team to build a non‑relational database capable of handling massive concurrent writes for features like QQ Farm. By 2009 the first CKV version was released, offering a NoSQL alternative to MySQL for social workloads.

TcaplusDB emerged in 2012 to meet the explosive growth of Tencent games. Designed as a three‑node “triple‑backup” system (primary, hot‑backup, cold‑backup), it provided seamless failover and fine‑grained server‑level redundancy, enabling games like Night Club King , Daily Run , and especially Honor of Kings to support millions of concurrent users without downtime.

The team continuously optimized storage‑compute separation, compressing data pages from 16 KB to 1 KB, reducing hardware costs by two‑thirds, and introducing dynamic memory‑disk hybrid caching to handle spikes such as the WeChat Red Packet Spring Festival event.

For video streaming, the “扁鹊” monitoring system was built in 2009 to automatically detect slow queries and failures, dramatically improving Tencent Video’s stability. Later, AI‑driven tagging and recommendation pipelines added massive read‑write pressure, prompting further database enhancements.

In 2016 the team launched DBbrain , an AI‑powered database operation platform that predicts load, auto‑generates scaling strategies, and provides real‑time alerts. DBbrain proved critical during the Creator 101 voting surge and large‑scale e‑commerce promotions, automatically expanding capacity and mitigating hotspots.

These technologies have since been offered as cloud services—CynosDB for relational workloads, TcaplusDB for high‑throughput gaming, and DBbrain for intelligent operations—supporting both internal Tencent products and external customers worldwide.

Through relentless engineering, Tencent’s database teams transformed from internal support units into pioneers of scalable, AI‑enhanced data infrastructure, illustrating how modern databases underpin everything from social apps to global gaming and video platforms.