Bank of Quanzhou Data Middle Platform Project

Leveraging big data technologies, this project integrates the bank's existing internal data with external data to build a big data platform with powerful parallel processing capabilities and seamless scalability. The core focus is the construction of a data warehouse and management & control platform, completing the infrastructure planning and implementation of the bank's next-generation data platform.

The domestically developed GBase 8a MPP distributed database is adopted as the data warehouse platform, and its logical data warehouse capabilities are used for deep integration with the Hadoop-based big data platform.

The GBase 8a MPP data warehouse platform processes and computes high-value-density structured data and delivers online data services for recent data (within the last five years).

The MPP data warehouse batch-synchronizes processed data to the big data platform for historical storage.

The big data platform leverages its inherent capabilities to handle storage and computation of streaming data, semi-structured/unstructured data, and low-value-density historical data, providing historical data storage and applications. Archival data services (over five years) are also fulfilled by the big data platform.

Phase one deploys a multi-node GBase 8a MPP cluster to support high-performance computing and queries for upstream analytics applications, including unified regulatory reporting, anti-money laundering, reporting, and management dashboards.

It delivers financial-grade high availability with a two-site three-center architecture.

High Performance:

Powers high-performance computing and queries for multiple upstream analytics workloads, including unified regulatory reporting, anti-money laundering (AML), reports, management cockpit, FTP system, and related-party transaction system.

High Availability:

GBase 8a MPP Cluster provides a comprehensive set of high-availability mechanisms to ensure maximum stability for Bank of Quanzhou’s data platform. Multi-replica technology safeguards single-cluster HA, virtual cluster technology enables logical cluster HA, and active-active cluster technology delivers cross-data-center HA, creating a financial-grade "two-site-three-center" architecture for robust disaster recovery.

Complete Domestic Technology Stack:

The project leverages a fully domestic technology stack—domestically manufactured servers, operating systems, and databases. This end-to-end domestic solution ensures complete independent controllability and provides a high level of security for the data platform.