Guangdong Unicom OCS Cloud List Query Project — Significantly reduce costs and improve utilization

lPerformance Boost: Accelerate query performance—sub-second response for single-month queries even with 100 concurrent front-end users;

lStorage Savings: Achieve 1x to 10x data compression to reduce storage footprint and enable retention of 6+1 months of historical transaction data;

lService Enhancement: Shorten wait times on the customer service platform, extend the time range for queryable logs, and improve the overall user experience with the OCS system.

This solution uses GBase 8a MPP Cluster to build a 10-node OCS historical log storage cluster, achieving 5:1 data compression through columnar storage and adaptive compression technology. Intelligent indexing significantly improves query efficiency, with response times under 3 seconds. The SafeGroup node backup mechanism ensures high system reliability.

China Unicom Guangdong urgently needed to resolve insufficient storage capacity and slow query performance for OCS historical log data. The revamped storage system required high-compression-ratio data compression to support 6+1 months of historical billing records. It also had to dramatically improve query efficiency over massive datasets, reducing query response times to under 3 seconds. The database product needed a comprehensive technical support system to ensure timely, expert server assurance in the event of technical failures.

China Unicom Guangdong currently serves 30 million subscribers. The log management platform for its Online Charging System (OCS), originally built on the Cassandra database, suffered from slow query response times and lacked a comprehensive, dependable technical support system in China. As OCS log data continues to accumulate, system performance and platform reliability have become top priorities. China Unicom Guangdong urgently needs a high-performance database product backed by robust technical services to address these challenges.