Fuxin Bank's Project to Replace Netezza Appliance with an Independently Developed Solution

Project Background

In July 2019, IBM officially announced the end of all technical support for its Netezza data warehouse appliance. This led to a series of major data warehouse crashes and business outages at Bank of Fuxin between July and December 2019. The project integrator and software vendor were unable to provide effective technical support or solutions, which became the main driver for the bank to replace its data warehouse product. In addition, the high expansion costs of the IBM appliance and increasing compute demands further motivated the customer to replace the foreign-built product.

As a result, Bank of Fuxin actively sought a comprehensive solution to replace the IBM Netezza product with a China-developed alternative for its data warehouse, exploring a domestically controlled technology roadmap and seeking a solution that delivers high performance with low hardware investment.

Solution

Bank of Fuxin’s big data platform was built on the GBase 8a MPP Cluster data warehouse, replacing the IBM Netezza appliance. It fully supports credit risk systems, fund audit systems, reporting services, daily batch processing, and other business applications. Serving over a dozen business systems across the bank’s data warehouse domain, it provides batch processing, reporting, in-depth data analysis, data alerts, and other data support services, delivering a new high-performance data processing engine for business growth.

The big data platform architecture consists of the data source layer, data processing layer, data mart layer, analytics layer, and application layer.

Data Processing Layer

The GBase 8a MPP Cluster replaced the original IBM Netezza appliance to handle front-end data collection, data aggregation, and other processing tasks, including data comparison, loading, and summarization. A new active-standby cluster synchronization feature was also added within the same city to achieve intra-city disaster recovery.

Data Mart Layer

The data mart layer includes marts for credit risk, fund audit, reporting applications, and data mining. The GBase 8a MPP Cluster replaces all functions of the original IBM Netezza appliance, handling basic data batch processing and data summarization for related business modules. Depending on the business application, corresponding data marts are built to support reporting tools and data access for various business systems.

Application Results

Full Platform Localization with China-Built Technologies:

Built on a fully domestic platform, GBase 8a MPP’s superior architecture and high-performance data processing capabilities completely replace the traditional all-in-one data warehouse appliance.

Superior Performance:

GBase 8a MPP’s advanced architecture and solutions enable high-speed data loading, exporting, and processing. With just three computing nodes, the cluster’s business batch processing performance already surpasses that of the legacy Netezza appliance.

High Availability:

Multi-level high-availability technologies fully safeguard the uptime of the bank’s data warehouse, reporting, and audit systems, ensuring stable 24/7 operation with no single point of failure risk.

Security and Control:

The core data warehouse system is built on a domestically developed distributed database with full intellectual property rights. Independently developed and secure, it enhances both data security and business security for the bank’s existing data assets.