Xuzhou Rural Commercial Bank Core Database Replacement and Renovation Project

The Xuzhou Rural Commercial Bank originally built its data warehouse on the DB2 database. As business grew and applications became more complex, data volumes surged, putting increasing pressure on data storage and processing. Over time, DB2 could no longer meet business requirements. It also faced challenges such as an inability to scale linearly and declining efficiency, falling short of the storage and computing demands driven by rapid data growth.

Xuzhou Rural Commercial Bank built a cluster with GBase 8a MPP Cluster to underpin its internal data warehouse, supporting various subsystems and data model computation services.

After winning the bid in the first half of 2023, Xuzhou Rural Commercial Bank began deploying the GBase 8a cluster and conducting joint testing with applications. The production cluster was deployed on three nodes.

Migrating from DB2 to GBase 8a involved database objects, data, and application migration—including SQL queries, stored procedures, and functions. To achieve better performance, the team tuned the GBase 8a cluster by optimizing database and OS parameters, data models, and business scheduling.

GBase provides the solid foundation for Xuzhou Rural Commercial Bank's Lakehouse project, encompassing data warehouse, data mart, and data modeling applications. It supports business data analytics while enabling higher-level, large-scale complex queries. By upgrading the existing data warehouse architecture, it reduces data batch processing time, improves data query efficiency and computational performance, and delivers end-to-end data technology support for the bank's digital transformation.

High Performance: Intelligent indexing, a fully parallel architecture, and transparent compression deliver lightning-fast query analysis. Fully optimized for high-performance query and analysis scenarios, it achieves a 10x performance improvement.

Dynamic Scalability: The system offers strong scalability, supports dynamic cluster expansion, and performance scales linearly as nodes are added.