Customer Voice: Distributed Database Application Practice at a Small and Medium-Sized City Commercial Bank

In recent years, IT in the financial industry has developed rapidly, with technology architectures being updated at a fast pace. Distributed architecture has gradually become mainstream because it better aligns with the future business development needs of the financial sector. As a critical component of information systems, databases run through the entire front, middle, and back ends of systems and have become increasingly important. This article uses the exploration and implementation of a distributed database at a small and medium-sized city commercial bank as a case study to analyze and summarize the application experience of distributed databases in small and medium-sized banks' business systems.

Background

With the rapid development of banking internet services, the standalone model of centralized databases and their associated hardware resource configurations can no longer meet the demands of rapidly growing business volumes. Centralized architectures generally lack elastic scalability. As transaction volumes and data volumes increase, overall system throughput encounters hardware or technical bottlenecks. Especially when supporting internet-facing customer services, it cannot effectively handle sudden surges of high-concurrency transactions, constraining customer acquisition and large-scale business marketing.

Given this situation, the city commercial bank considered leveraging the technical unity and convergence of databases and cloud platforms to address the issue of architectural consistency. By sequentially deploying traditional IT infrastructure (steady state) and distributed cloud infrastructure (agile state), the data center's service capabilities now balance both security and agility needs. Using cloud-native and cloud migration approaches, traditional applications are transformed in batches and gradually moved to the cloud, while internet applications are deployed directly on the cloud, forming a complete cloud ecosystem.

Through a gradual transformation from centralized to distributed, a coordinated IT infrastructure combining centralized and distributed architectures has been essentially established. The bank fully leverages the security and stability value of the traditional centralized architecture while relying on the distributed architecture to aggregate internal and external resources to build core digital capabilities.

Application Case

Throughout the entire architecture upgrade process, the multi-modal distributed database GBase 8c provided foundational support and connection management, enabling architectural splitting of applications and re-adaptation of the data access layer. Meanwhile, application containers and deployment environments were upgraded using K8s+Docker, and data tables (partitioning and sharding) were re-planned and upgraded. In terms of scenario performance, the new architecture significantly improved application scalability, concurrent processing capabilities, and concurrent handling in complex scenarios, greatly enhancing batch processing efficiency.

In test applications, the GBase database demonstrated excellent TPS performance and easy business adaptation. The GBase 8c multi-modal database was introduced into core financial domains, exploring and innovating with distributed database technology in corporate lending services (replacing MySQL), retail/risk control services (replacing Oracle), and internet middle-platform services, gradually achieving domestic substitution across multiple business systems.

During implementation, the technical team overcame multiple technical challenges, including credit core business data planning under the distributed architecture, daily batch performance tuning, distributed deadlock detection and resolution in batch scenarios, and Kafka data integration, ensuring the smooth progress of the project.

Selection Experience

Planning First
Selection should align with future business development and digital transformation plans to improve resource utilization.

Implementation Cases
Assess the database's industry application track record, whether it has banking application cases and implementation experience, to reduce implementation risks.

Vendor Profile
Pay attention to the R&D team and the product's future evolution roadmap; prioritize vendors that are mainstream in the market and relatively stable.

Product Capabilities
Choose a database product with high availability, high performance, strong support, a mature ecosystem, and ready-to-use supporting tools.

Conclusion

After this practical exploration of distributed databases, especially the successful deployment of core systems, the technical feasibility of distributed databases has been fully verified, offering a viable solution for small and medium-sized banks to build a secure and reliable underlying database. It provides experience in the deployment of Xinchuang (domestic IT innovation)-compliant distributed databases and full-stack applications, and also offers a reference for the Xinchuang construction of critical systems in the financial industry.