Civil Aviation Administration of China (CAAC) Comprehensive Statistical Information System Database Upgrade Project

Project Background

The CAAC Comprehensive Statistical Information System is the current operational business system for civil aviation. It collects and stores vast amounts of data on airports, transport production, general aviation production, and more. To unlock data value more quickly and easily, the CAAC decided to develop a data analysis platform for the comprehensive statistical system, enabling more refined extraction, mining, and analysis for greater business value.

Business Requirements

The CAAC Comprehensive Statistical Data Analysis Platform uses the comprehensive statistical business system as its data source. Centered on modules such as production statistics, specialized statistics, other statistics, comprehensive statistical analysis, and the platform itself, it builds a data warehouse based on the GBase database. ETL data extraction is performed through Oracle Data Integrator (ODI), and then tools including IBM Cognos, ArcGIS, and ECharts are used to present warehouse data in multi-dimensional, multi-level visualizations, ultimately meeting the needs of civil aviation customers.

Implementation Requirements

To accommodate future business growth, the original system is upgraded with GBase 8a MPP Cluster Database version 9, which supports virtual cluster capabilities and multi-tenancy management. This enables data isolation, resource isolation, fault isolation, and independent scaling across different business lines.

Currently, the MPP cluster hosts the data warehouse for the comprehensive statistics system. In the near future, data will be loaded into the data warehouse of the shared exchange platform and the data warehouse for the specialized data analysis subsystem. The upgrade will ensure the system supports virtual clusters, enabling multiple services to run efficiently and in an orderly manner within virtual clusters.

This project upgrades the MPP data capability platform to meet the needs of a wider variety of data statistics and analysis services.

To support future business growth, the MPP data capability platform is being upgraded to version V9, which provides virtual cluster capabilities and multi-tenant management. This enables data isolation, resource isolation, fault isolation, and independent scaling across different services. In the future, additional servers will be added, and physical isolation will be implemented: each service’s data will be deployed on a virtual cluster consisting of several servers, ensuring hardware isolation of data nodes while sharing existing cluster management nodes. Hardware isolation of data nodes between tenants ensures that tenants do not impact each other.

System Architecture Diagram (Figure 11-1)

Deployment Overview

The platform is deployed across 6 nodes (6 data nodes and 3 management nodes), with a total data volume of approximately 2 TB and around 200 tables in the database.

Benefits and Value

This upgrade project enables the existing MPP cluster to support virtual cluster functionality, better organizing roles for the MPP big data platform and achieving logical isolation of different services. It also enhances the reliability and scalability of existing database application systems.