A Computing Base for Bank of Communications Data Warehouse - Industry Cases

Client Profile

Founded in 1908, Bank of Communications is one of China's six largest banks and a major provider of financial services in the country. As digital innovation continues to advance, the outdated data architecture is no longer adequate, and the increasing volume of new business data requires stronger computing power to support it.

Business Challenges

With the rapid development of Bank of Communications' mobile banking services, the data generated by production and management systems is growing exponentially. The traditional data warehouse analysis methods and siloed architecture can no longer meet business demands.

As the pace of digital society construction accelerates and a new round of financial market opening picks up speed, Bank of Communications has taken the lead in the financial industry by conducting test and verification work. This involves upgrading the infrastructure and decentralizing the business and management systems.

The operational costs of digital business are gradually increasing, necessitating more optimized methods and tools. This requires suppliers to provide stable and reliable products and to have the capability to assist customers in planning their business and systems to meet the regulatory requirements.

Bank of Communications has built a FusionInsight+DWS big data cluster with over 1,000 nodes, reconfiguring the One Lake, One Warehouse data architecture system. Through the real-time engine, the system achieves real-time updates of massive data volumes. Given the trend toward domestic production, high-performance, domestically-produced servers are needed to provide sufficient computing resources.

Solution

Assisted the client in building a FusionInsight+DWS big data cluster with over 1,000 nodes, integrating and reconstructing multiple existing data platforms. Leveraging the capabilities of the new product version, the data architecture system was re-planned under the “One Lake, One Warehouse” model. This fusion of lake and warehouse ensures data consistency, and the real-time engine facilitates the real-time update of massive data volumes.

The Bank of Communications data warehouse (DWS) utilizes KunTai R722 servers as the data foundation. These servers employ a Massive Parallel Processing (MPP) architecture, supporting both row-based and column-based storage, and providing the ability to process data volumes at the petabyte (PB) scale.

Solution Advantages

KunTai R722 servers will primarily be applied to the following business scenarios:

Detailed Record Queries: Handles PB-level data loads with in-memory analysis for real-time querying in large-scale scenarios.
Data Warehouse: Supports hundreds of TBs, efficiently processing multi-table joins with billions of rows for ODS, EDW, and DM operations.
Hybrid Workloads: Simultaneously handles OLTP and OLAP tasks with hybrid row-column storage and vast dataset capabilities.
Big Data Analysis: Supports PB-scale structured data analysis with a distributed parallel database cluster.

Results & Values

Unified Data Management for Bank of Communications

By leveraging data from various sources, management no longer has to make business decisions based on limited data. The data warehouse and intelligent BI (Business Intelligence) tools can be directly applied to business processes such as market segmentation, inventory management, financial management, and sales.

Increased Efficiency and Cost Savings

•The data warehouse enables the creation of a data model for Bank of Communications, which is crucial for sales, cost control, and revenue distribution. This greatly reduces the company's costs and improves economic efficiency. Additionally, the data warehouse can analyze the relationship between human resources and foundational data, providing feedback for analysis and ensuring the IT system is utilized to its fullest potential.

Enhanced Data Quality and Consistency

The implementation of the data warehouse involves converting data from various source systems into a common format, facilitating data processing and improving accuracy.