Minsheng Bank’s Distributed Transformation and NewSQL Practice with SequoiaDB

Historically, Chinese financial institutions relied on centralized IT architectures that struggled to keep pace with exponential growth in system scale and digital finance demands, prompting regulators such as the People’s Bank of China to encourage distributed transformation.

Minsheng Bank, a leading retail bank with RMB 3.2 trillion in assets, has been at the forefront of both business and technology innovation, especially in artificial intelligence and big data, and now seeks a distributed architecture to meet new challenges.

Key distributed transformation needs include:

Elastic management of massive data systems to handle petabyte‑scale growth.

High‑performance data processing for real‑time, high‑concurrency operations.

Unified handling of structured, semi‑structured, and unstructured data to improve system efficiency.

Simplified development and operations to reduce costs while maintaining elasticity and compatibility.

Domestic, controllable products that provide source‑level control for core business systems.

In the database layer, Minsheng Bank has undertaken three major initiatives:

Sharding and partitioning across Oracle, IBM DB2, and MySQL, achieving over 15 logical databases and 500+ tables.

Cross‑center distribution and active‑active deployment for traditional relational databases, improving safety, RTO, RPO, and resource utilization.

Adoption of new distributed database products such as SequoiaDB for massive data queries, distributed imaging platforms, and archival management.

After evaluating several NewSQL solutions, the bank selected SequoiaDB 3.0 for its full MySQL protocol compatibility, proven reliability in the financial sector, and native distributed architecture.

SequoiaDB’s advantages highlighted are:

Distributed scalability with elastic storage expansion.

Micro‑service‑style architecture allowing flexible configuration of SQL instances and storage nodes.

Complete MySQL compatibility, enabling seamless migration and reducing operational costs.

Performance tests across representative banking scenarios demonstrated impressive throughput:

Complex channel query (4‑table join, billions of rows) – 3,916.45 transactions per minute with stable throughput.

High‑frequency transaction flow query – 1,886,184.03 transactions per minute, maintaining stability.

Counter‑service (mixed query/update) – 51,090.03 transactions per minute.

Billing operations (insert, update, query) – 9,861.57 transactions per minute with minimal variance.

In conclusion, SequoiaDB 3.0’s strong MySQL compatibility, scalability, and performance satisfy the bank’s critical transaction requirements, and the bank plans to migrate additional workloads that are difficult for traditional sharding solutions to the NewSQL platform while continuing to evaluate large‑scale distributed database adoption.