Hebei Unicom Big Data Platform – Supporting Integrated Business Operations with High Efficiency and Low Cost

With the continuous development of mobile internet, the rapid popularization of smart terminals, and the explosive growth of mobile data traffic, traffic operation has become the focus of Hebei Unicom's strategic transformation, and the prerequisite for traffic operation is the sustainable operation of the business analysis system. Faced with massive network data volumes, traditional business analysis systems are clearly unable to meet Hebei Unicom's growing data processing requirements in terms of data storage, processing, and analysis; at the same time, traditional minicomputers have high scaling costs and limited expansion capabilities, which do not align with Hebei Unicom's future development plan for big data platforms.

Due to cost and performance constraints, the traditional minicomputer and Oracle-based BI system could not meet Hebei Unicom's data processing requirements. As more service platforms were to be connected, the provincial BI system's data scale would reach petabytes, exceeding the existing platform's capacity. To enable precise user profiling, both content data and Internet data needed to be integrated into a unified data platform. Driven by the principle of 'high efficiency at low cost' and the need for both massive data storage/management and horsepower for data processing/analysis, Hebei Unicom deployed GBase 8a MPP Cluster as the underlying data platform, building a big data platform that integrates business operations and delivers data services externally.

As shown in the architecture diagram, this project builds a core data warehouse with a 24+2 node GBase 8a MPP Cluster database, and uses GBase 8a fast loading tools to quickly load data into the data warehouse. The core data warehouse built with GBase 8a stores high-value structured data across the entire Hebei Unicom, and through associative analysis and computation on the full-domain data, it supplies highly aggregated result data to the platform application layer and external system data interfaces, while directly supporting report queries and ad-hoc queries on the front-end pages.

High efficiency, low cost: Build distributed clusters on low-cost x86 servers that deliver higher processing performance than minicomputers;

Flexible scaling: Free from data volume and hardware limitations, the cluster scales seamlessly from 8 nodes to 12 to 24 nodes on demand;

Unified data integration: Fully incorporates data from all provincial business support systems, extends data retention periods, and reaches a data volume of 300 TB;

Worry‑free migration: Our engineers provide end‑to‑end support for migrating data and applications from Oracle to GBase 8a, enabling zero‑code changes and direct operation on GBase 8a.