Panzhihua City Commercial Bank Netezza Appliance Localization Project

Bank of Panzhihua: Netezza Appliance Replacement Project with a Homegrown Database

Project Background

Bank of Panzhihua had accumulated massive business data, with data volumes growing exponentially and exhibiting vast, rapidly expanding characteristics. The original Netezza data warehouse system had encountered performance bottlenecks and could no longer meet the bank’s increasing data output demands. Recognizing that the Chinese banking industry has widely adopted MPP architecture databases on open x86 platforms to build structured data processing and analytics applications, the bank chose GBase 8a to replace the legacy Netezza data warehouse appliance.

Solution

Bank of Panzhihua’s data warehouse leverages GBase 8a MPP Cluster to achieve unified storage, management, information sharing, and data resource services for massive data volumes. It serves as the foundation for application systems, creating specialized data marts for different business lines and establishing a comprehensive architecture for data collection, loading, storage, analysis, and presentation.

Overall System Architecture

Data Warehouse Architecture Description:

  • Data Source Layer: Consists of the bank’s various existing business systems.

  • Extraction and Loading Layer: Uses ETL tools to extract, load, and transform massive data from source systems.

  • Storage Management Layer: Built with the GBase 8a MPP cluster. After data cleansing by loading servers, data is distributed to cluster nodes according to defined rules, forming the primary data warehouse and data marts, with the mart size tailored to business requirements.

  • Analysis and Presentation Layer: The bank uses third-party analysis and mining tools to extract data from the warehouse or marts, conduct further analysis, and feed results into corresponding business modules.

  • Application Portal Layer: Through middleware, the data required by each module is organized and presented to internal or external systems via the bank’s portal.

Results

  • Dynamic Scalability:

The system offers strong scalability, supports dynamic cluster expansion, and delivers linear performance improvement as nodes are added.

  • Data Migration:

A comprehensive solution enables risk-free migration from third-party databases to GBase 8a MPP, with a standardized process that simplifies operations and minimizes risk.

  • High Availability:

An active-active synchronization mechanism ensures complete data consistency between the primary and standby clusters after daily cluster-level batch synchronization. This guarantees high availability of data and services. Even if the primary cluster fails and cannot be quickly recovered, cluster switching can be rapidly performed, using the standby cluster to continue providing data and services, fully safeguarding system availability.

  • Low Investment, High Efficiency:

GBase 8a MPP Cluster runs on cost-effective x86 PC servers, delivering high performance with significantly lower hardware costs.

  • Visual Management and Easy Maintenance:

A user-friendly visual cluster management tool with comprehensive functions simplifies administration and maintenance, greatly improving operational efficiency.