TXSQL: The Cloud‑Era Database Kernel – Overview, Design, and Future Directions

Background: On May 23‑24, the Tencent "Cloud + Future" summit was held in Guangzhou under the theme "Huanqi". Government leaders, academic experts, industry veterans, and technical elites gathered to discuss cloud computing and digital industry innovation.

At the developer session, Tencent MySQL kernel R&D expert Zhang Qinglin presented the topic "TXSQL: The Database Kernel Bomb for the Cloud Computing Era". The talk introduced TXSQL from five perspectives.

Part 1 – Concepts of Tencent Cloud and TXSQL – Tencent Cloud, a leading domestic cloud provider, operates more than 500 data centers serving millions of developers. The CDB product is Tencent Cloud’s TXSQL offering, providing a complete solution for cloud migration, backup, recovery, and upgrades, simplifying MySQL service deployment. CDB serves diverse industries such as banking, securities, logistics, and traditional enterprises. Users receive host, port, username, and password to connect directly to the backend database. TXSQL is a MySQL branch maintained by the database kernel team and is delivered to customers via CDB.

Part 2 – Why TXSQL Was Created – Rapid growth from 2016 to 2017 saw storage scale increase more than fourfold and instances exceed 100,000. Operating millions of servers introduces three major challenges: (1) stability – large‑scale CDB usage requires fast issue detection and resolution; (2) new business demands – sectors like finance need advanced security and functionality; (3) performance – high‑traffic events (e‑commerce, gaming) demand cost‑effective performance improvements. Consequently, Tencent Cloud needed an internal kernel team to quickly locate problems, enhance performance, and meet evolving requirements.

Part 3 – TXSQL Design – The B3M code‑design model improves usability and problem‑solving. Every modification to MySQL undergoes strict testing. The development flow includes requirement analysis (bug fixes, user requests, performance tuning), design, coding, and extensive quality controls such as code‑coverage and unit tests. New kernel versions are released every 2‑3 months after stability, performance, and crash‑recovery testing. Deployment follows a staged rollout: single instance, small cluster, then full‑network release, ensuring reliability.

Part 4 – User‑Facing Features and Optimizations – TXSQL adds performance‑tuned features not present in upstream MySQL: redo‑log optimization (reducing sync operations, multi‑buffering, concurrent writes), select‑offset‑limit push‑down, and thread‑pool integration that resolves global read‑lock deadlocks, dump‑thread interference, and provides monitoring tables in the information schema. Functional extensions include encryption, audit plugins (enterprise‑level audit delivered as a plugin), and parallel replication. Thread‑pool handling improves concurrency under heavy load.

Part 5 – Future Directions – TXSQL will continue to enhance stability, performance, and functionality. Planned initiatives include batch computation by pushing more logic to the engine, execution‑plan caching for further speed gains, and integrating RocksDB (TXRocks) to provide transactional support with lower storage costs.

Overall, TXSQL serves as a cloud‑native database kernel that combines rigorous development processes, performance engineering, and feature extensions to meet the demanding workloads of modern cloud applications.