Bank of China Liaoning Branch Big Data Platform — Innovative Data Architecture for Efficient Reporting Data Delivery

Management Innovation: Transformed the data architecture by consolidating all business model data into a single database, delivering report data to each model via Smartbi.

High Performance: The business scenario has undergone comprehensive testing, achieving batch processing performance 7 to 10 times faster than Oracle and tens of times faster for precise queries on large datasets — delivering high performance gains at a low investment.

High Scalability: The cluster supports online scaling with multiple nodes, ready to meet future business performance and data storage requirements.

High Availability: Ensures 24/7 continuous operation while eliminating downtime caused by single points of failure or system maintenance.

This project deploys a 2×3-node cluster, utilizing the GBase 8a MPP Cluster, a large-scale distributed parallel database system, to build the core data management layer of the data platform, forming a unified foundational data platform for all data. Currently, the data volume stands at 10 TB, with a daily increment of 30 GB. There are over 1,000 tables, the largest of which exceeds 10 billion rows, and daily batch processing runs for 4 to 5 hours. Based on business requirements, the configuration includes 1 loader node and 6 compute nodes.

Feature-rich: Supports large-scale integration of data migration, data processing, data querying, and application connectivity;

High Performance: Enables rapid response for data loading, import/export, processing, splitting, and querying;

High Availability: 7x24 continuous operation, avoiding downtime from single points of failure or system maintenance;

Linear Scalability: Supports online capacity expansion across multiple nodes;

Driven by the big data landscape, the Bank of China Liaoning Branch proactively drives business innovation and has built a big data platform underpinned by an MPP relational database to support massive data storage, analysis, and statistics. The MPP database product is currently mainly deployed in the historical query system, and new business systems will be developed according to data scale in the future.