Method and processing system for improving the pricing system IO and concurrency performance of a web-based car-hailing service
By introducing a local cache queue and asynchronous processing mechanism into the ride-hailing pricing system, the problems of database IO latency and excessive load in high-concurrency scenarios were solved, achieving faster response speed and higher concurrency processing capabilities, and ensuring system stability and data consistency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- 广州宸祺出行科技有限公司
- Filing Date
- 2026-03-10
- Publication Date
- 2026-06-19
AI Technical Summary
In high-concurrency scenarios, the cumulative database IO write latency of the ride-hailing pricing system leads to slow interface response. High-frequency and scattered data writes cause excessive database concurrency load. The lack of effective local buffering and anti-loss mechanisms results in insufficient system stability under peak traffic, making it unable to meet the requirements of high concurrency and low latency.
A local memory cache queue combined with asynchronous processing is used to trigger batch data write to the database based on a threshold. The pricing result data is written to the local cache queue and returned immediately. Asynchronous threads are used to perform cleaning and deduplication processing, and data is recorded through write-ahead logs (WAL). Combined with dual-condition triggering of batch data write to the database and asynchronous status feedback, data consistency and disaster recovery are ensured.
It improved interface response speed, reduced database I/O request frequency and load, ensured stable system operation under peak traffic, reduced database write volume, prevented data loss, and improved the system's concurrent processing capabilities.
Smart Images

Figure CN122243590A_ABST