GBase completes the go-live deployment of Bank of Quanzhou's fully China-built data platform project.
Recently, the Bank of Quanzhou's data platform project went live successfully. Its data warehouse is built on GBase 8a MPP Cluster, leveraging GBase 8a's data fusion capabilities to share information with the big data platform. Phase one deployed a tens-of-nodes MPP cluster to support the bank's big data analytics platform and facilitate the deployment of its all-domestic technology solution.
Project Overview
Bank of Quanzhou is entering a new phase of business development, with ever-emerging business models and growing data volumes. As business data becomes richer in content and form, higher demands are placed on overall data support and data service capabilities. Therefore, building a highly parallel and easily scalable big data platform based on big data technology, integrating existing internal data and bringing in external data, has become imperative.
Through the construction of a data middle platform, Bank of Quanzhou aims to build a one-stop, integrated development platform with the business goal of enhancing 'data asset monetization' capability, enabling centralized management of bank-wide data, unifying the data development platform, standardizing data, and improving data quality to help the bank achieve its digital transformation objectives.
Solution
Bank of Quanzhou Data System Overall Architecture Planning
1. Data Integration
Includes structured data, semi-structured/unstructured data from internal source business systems and external data sources, providing source data for the data platform.
2. Data Governance
The data governance platform includes metadata management, data standard management, data quality management, and data asset management. Data governance runs through the entire construction process of the data platform. It is necessary to plan the governance platform's architectural framework, analyze and sort out the management processes for data standardization, data quality, and metadata within the bank, and work with the governance platform to improve data management and data quality.
3. Data Development and Operations
Establish unified development standards to improve data development quality and efficiency and reduce data operation costs. This primarily includes unified data collection, unified data exchange, unified ETL processing, and a unified scheduling platform, spanning the entire development and operation process of the data platform.
4. Data Platform
The data platform consists of the foundational data platform and the big data platform, providing data support for data applications. The data platform comprises the big data platform and the data warehouse. In this project, the data warehouse is built using GBase 8a MPP Cluster, leveraging its data fusion capabilities to share information with the big data platform. Phase one deploys a tens-of-nodes MPP cluster to support the bank-wide big data analytics platform in the future.
5. Data Services
Based on the business requirements of application systems, the data service system provides data from the data platform and offers classified and graded data exchange services, including online query services, real-time computing services, external data services, and batch data services, to meet the interaction needs between the data platform and internal/external systems. Among these, we also provide users with self-service analysis and other data service functions through self-service analysis tools.
6. Data Applications
As the presentation and application window of the data platform, the data application system primarily achieves data presentation and analysis (fixed reports, self-service analysis, flexible query, graphical display, historical query) through a unified application platform, meeting urgent user needs for report management and data analysis, supporting business decision-making, and fully mining data value.
Results and Value
1. Lakehouse Integration
GBase 8a MPP Cluster's data fusion capabilities enable seamless data integration between MPP and Hadoop platforms: using a unified access interface, it realizes online, transparent cross-heterogeneous platform data interaction and data flow. With this capability, Bank of Quanzhou can build a lakehouse integrated platform, achieving lake batch processing and warehouse computing, data lifecycle management, and multi-model data fusion.
2. Full Sovereignty and Control
The entire platform adopts purely domestic technology, including servers and databases. The all-domestic solution significantly enhances the platform's high reliability.
3. High Availability
GBase 8a MPP Cluster provides multiple high-availability mechanisms: single-cluster high availability through multiple replicas, logical cluster high availability through virtual clustering technology, and cross-data-center high availability through data disaster recovery, thus meeting the high-availability guarantee of the two-site-three-center architecture required by the financial industry.