How China’s Postal Savings Bank Built a Distributed Core System with OpenGauss and Kunpeng

Background and Motivation

Postal Savings Bank (PSBC), a state‑owned Chinese bank with nearly 40,000 branches and over 6 billion personal customers, faced performance bottlenecks in its traditional IOE‑based core system as transaction volumes and digital services grew.

To stay competitive in the digital economy, PSBC decided to replace the legacy core with a next‑generation, distributed architecture that could support high concurrency, rapid innovation, and long‑term scalability.

Goals of the New Core

The 2019 pre‑research defined five clear objectives: support business innovation, break through system capacity limits, achieve autonomous controllability, align with emerging distributed technologies, and sustain growth for the next 5‑10 years.

Technology Stack and Architectural Choices

PSBC adopted a cloud‑native approach built on a general‑purpose computing platform combined with open‑source software. Key components include:

Microservice architecture and containerization for modularity.

AI‑assisted operations and DevOps pipelines for automated deployment and monitoring.

Distributed data management using horizontal sharding, multi‑database deployment, and the open‑source relational database openGauss as the core data store.

Kunpeng processors and the openEuler operating system to provide a full‑stack, domestically sourced hardware‑software ecosystem.

Why openGauss?

During the selection process, PSBC evaluated technical leadership, ecosystem maturity, commercial viability, and talent availability. openGauss offered superior performance—over twice that of MySQL/PostgreSQL on both x86 and Kunpeng—thanks to its advanced kernel technology, extensive real‑world telecom usage within Huawei, and a rapidly growing open‑source community.

Implementation and Deployment

The new system follows a unit‑based deployment model: each unit contains identical hardware and software but stores data for a specific customer segment. Scaling is achieved by replicating units, enabling seamless horizontal expansion.

Key milestones:

May 2020 – Prototype validation on Kunpeng hardware and openGauss.

April 2021 – Full technology platform launched and integrated with production mirrors.

July 2021 – Distributed operation system went live, leveraging AI for massive node management.

Performance Outcomes

Since production, the distributed core has delivered:

~100% overall performance improvement.

~50% faster global routing queries.

~10% higher insert/update throughput.

The system is slated for full rollout by March 2022, targeting 20 billion daily transactions.

Industry Impact and Lessons Learned

PSBC’s experience demonstrates that a state‑owned bank can successfully adopt cloud‑native, open‑source technologies to achieve both stability and cutting‑edge performance. The project also contributed to the openGauss community, with PSBC acting as a council member and providing real‑world feedback.

Overall, the case provides a valuable reference for other financial institutions seeking to modernize legacy core systems through distributed, open‑source, and domestically sourced solutions.