Bank of China Shanxi Branch Basic Customer Analysis and Expansion System — Precision Marketing, Efficient Service

Data Query Performance Improvement: Leveraging the intelligent indexing technology and columnar storage of the MPP database, the efficiency of equality, range, and fuzzy queries in application scenarios increased by 300% to 500%, effectively supporting OLAP business analysis needs for marketing management, financial product management, and decision support.

Business Model Batch Processing Boost: The legacy database faced processing bottlenecks in daily batch jobs, causing delays that impacted business timeliness and new business expansion. By using the MPP database's fast multi-table join queries, average batch runtimes have been shortened by 5 to 10 times.

Data Synchronization Timeliness Enhancement: With GBase data synchronization tools enabling efficient sync with Oracle databases, data sync time was reduced from 2 hours to just 3 minutes, greatly improving data analysis effectiveness.

This project builds a 5-node cluster using the GBase 8a MPP Cluster, a large-scale distributed parallel database cluster system, to construct the core data warehouse and data marts for the data platform. It forms a unified fundamental data platform centered on customer management, encompassing customer assets, customer relationships, operations and marketing, contracting channels, and all related data. The current data volume stands at 2 TB, with a daily increase of 10 GB, over 1,500 tables in total, and the largest table containing 2 billion rows. Daily batch processing takes 3.5 hours. Based on business requirements, one loading server and one monitoring server are deployed.

High query performance: supports fuzzy query, equality query, and range query, with up to 5x performance improvement compared to Oracle database;

Efficient data synchronization: loads data from Oracle to GBase with up to 10x performance improvement;

High cluster availability: ensures 24/7 continuous operation and prevents downtime from single point of failure or system maintenance;

Cluster monitoring and management: enables monitoring of performance, storage usage, CPU status, memory status, and SQL performance.

Bank of China Shanxi Branch Co., Ltd. faced performance issues with its customer analytics system built on a traditional Oracle database. It took the lead in adopting an efficient MPP (Massively Parallel Processing) relational database to support a big data platform for massive data storage, analysis, and statistics. The MPP database product was deployed in the customer analytics system, enabling precision marketing through big data analytics to efficiently serve customers and marketing channels, and to add value to data management operations.