PBOC Clearing Center 2022 Financial Big Data Analysis and Service Platform Construction Project

This project is the People's Bank of China Clearing Center's "2022 Financial Big Data Analysis and Service Platform Expansion Project." The procurement covers the deployment of a massively parallel processing (MPP) analytical distributed database system to build the foundational environment for the big data platform. It aims to support the main data warehouse (basic data zone, data processing zone, and indicator zone), data marts, and parallel batch computing.

To meet evolving business needs, the project involves expanding the MPP cluster at the Beijing center and establishing an identical MPP cluster at the Shanghai center.

In the first phase, 90 nodes of the GBase 8a database from GBase (General Data Technology Co., Ltd.) were procured. With rising business requirements, a second-phase expansion in December 2023 added 12 more nodes.

The Beijing site originally housed 4 clusters with a total of 30 nodes. To accommodate increasing data demands, the clusters were expanded, resulting in 4 clusters with a combined 60 nodes.

Planned: 4 clusters at the Shanghai site.

Synchronization Plan

This project, a major IT initiative for the headquarters, required strict adherence to timelines. During implementation, some headquarters applications were still running on the existing MPP cluster resources. To ensure stable operation of current business systems and minimize impact on the project schedule, the technical solution was designed to adapt to the existing production environment (based on GBase 8a), including but not limited to production application compatibility, ops script compatibility, and management through a unified monitoring platform.

GBase 8a database from General Data Technology (GBASE) ensures compatibility with the existing network, computing, and storage devices and business software of the People's Bank of China Clearing Center, achieving seamless interconnection. The new data storage and backup model and data model remain consistent with the original system, ensuring a seamless transition and minimizing application impact, rework efforts, and the risk of project delays.