Evolution of Ant Financial Database Architecture and the OceanBase Migration Service

On January 4, 2019, the OceanBase Migration Service was officially announced at the ATEC City Summit, where Ant Group senior technical experts shared its key features and business practices.

Three Generations of Ant Database Architecture

Over the past decade, Ant has upgraded its database architecture three times: the first generation was based on the IOE stack (IBM mainframes, Oracle commercial databases, and EMC shared storage), which suffered high operational costs and stability challenges; the second generation centered on Oracle and EMC combined with Ant's own distributed middleware, providing horizontal and vertical elasticity; the third generation, driven by the rise of mobile payments and the need for higher stability, speed, and elasticity, is built on OceanBase, a financial‑grade cloud database, and distributed middleware.

Challenges of Database Architecture Upgrade

Upgrading database architecture is costly in manpower and expenses, and poses challenges such as ensuring stability during migration, maintaining data quality, and handling compatibility and performance risks.

OceanBase Migration Service (OMS): A Direct Path to Distributed Architecture

Leveraging ten years of Ant's database upgrade experience, OMS offers a one‑stop migration solution that includes load replay verification, sub‑second data validation, minute‑level instant rollback, support for multiple source databases (Oracle, MySQL, OceanBase), and one‑click migration.

OMS Technical Architecture

OMS treats OceanBase as a backup for Oracle/MySQL, establishing a virtual primary‑secondary link that simplifies migration and rollback to a primary‑secondary switch.

OMS Migration Process

The migration consists of seven steps: 1) Assessment via load replay; 2) PoC deployment; 3) Pre‑migration to a test instance; 4) Validation with load replay and optimization suggestions; 5) Full migration with incremental sync; 6) Multi‑level data verification; 7) One‑click cut‑over and rollback.

Case Study: Ant Merchant Platform

The platform, which uses both MySQL and Oracle, faced rapid data growth and performance bottlenecks. Using OMS, assessment and adaptation time dropped from 1‑2 months to one week, migration latency reduced to milliseconds, and cut‑over time decreased from weeks to minutes, while rollback became instantaneous and loss‑less.

Overall, OMS dramatically improves migration efficiency, reduces human and time costs, and helps enterprises adopt distributed architectures for scalable, reliable services.