Dahshenlin is a leading listed pharmaceutical retail group in China and ranks among the top three in the pharmacy chain industry. It has consistently been recognized as:
“No. 2 in Sales Among Chinese Pharmacy Chains” “Top 500 Chinese Service Enterprises” “Top 100 Most Comprehensive Chinese Pharmacy Chains” “Top 100 Private Enterprises in Guangdong Province” “Outstanding Enterprise in Guangzhou”
The company’s core business strategy centers on expanding its pharmacy chain network. It holds a significant position in the pharmaceutical retail sector through aggressive store expansion and digital operations.
Dahshenlin currently operates over 30,000 stores, with direct operations spanning 21 provinces, autonomous regions, and municipalities.
As the enterprise continues to scale, store count and data volume grow rapidly. At the same time, demands for financial management granularity are increasing. The existing Oracle-based finance and operations system shows clear performance limitations in:
Financial data ingestion throughput Storage efficiency Report query and rendering speed A systematic upgrade of the data infrastructure has become urgent.
Single tables contain tens of millions of rows. Daily data growth exceeds 1.5 million rows. Monthly growth reaches 45 million rows. Annual新增 data is projected at over 540 million rows. This scale places high demands on multi-level data compression capabilities.
With 30,000+ stores, concurrent queries from multiple locations stress the backend database’s processing power and reporting efficiency.
During month-end peaks, 200+ voucher batches are written simultaneously—
including vouchers transferred from external systems (e.g., operations-finance integration, shared services) and those generated internally by the financial system. This creates significant write-load pressure on the database.
YMatrix uses a distributed HTAP database architecture that combines transactional and analytical processing in a single system. This supports real-time transaction workloads and complex analytics simultaneously, enabling rapid use of business data in scenarios such as real-time reporting and financial balance analysis.
At the data coordination layer, the Domino in-database streaming engine synchronizes data in real time between the TP (transaction processing) zone and the AP (analytical processing) zone. Coupled with the MARS3 storage engine’s adaptive compression, this achieves multi-level compression of ODS-layer data and reduces redundancy.
The deployment consists of a five-node physical cluster:
One master node manages client connections and coordinates segment operations. Four segment nodes handle data storage and processing. Segment nodes are logically divided by workload type:
TP zone segments process high-frequency transactions, such as voucher entry and asset accounting. AP zone segments support analytical workloads, including financial report generation, intelligent analytics, and month-end closing calculations.
This architecture unifies operations and finance data, providing a single platform for upper-layer modules—such as event-based accounting, financial consolidation, and intelligent analytics. It delivers efficient, stable transaction processing alongside real-time analytical capabilities, with strong support for high-throughput data ingestion and complex query processing.
YMatrix HTAP Transforms Month-End Closing for a 16,000-Store Pharma Chain
Smart Manufacturing at Scale with YMatrix HTAP: Real-Time Ingestion & Unified Analytics
How YMatrix Powers SVOLT’s Smart Factory Transformation
How YMatrix Domino Replaces Lambda, Kappa, Flink, and Spark with One Engine🚀
From Greenplum to YMatrix: Migrating Core Business Data for a Leading Power-Battery Manufacturer
