基于Oracle的海量实时报文存储访问策略

Strategy of Mass and Real-Time Message Access and Storage Based on Oracle

海量实时数据的存储访问是实时报文存储处理系统存储性能的瓶颈。分析了实时报文的应用场景;构建了实时报文数据访问的时间耗费模型,探讨了系统的主要时间耗费因素;提出了分区存储、批量存取和分页查询三种数据存储访问策略;编程实现了报文数据访问引擎,运用线性回归方法拟合了数据的保存速率。验证表明,该引擎具有高效性、资源占有率低、伸缩性强和稳定可靠的特点。

Mass access and storage of real-time data has capability bottleneck for real-time telegram processing systems.This paper firstly analyzes the application contexts of a real-time message system,figures out a model of time consumption for data access in such systems,and discusses the main time-consuming aspects.Then,it puts forward three strategies for mass access and storage of real-time message： database partition,batch data access and query page by page.Based on the strategies,the paper designs and implements a message access engine.Results show that the engine has characteristics of higher efficiency,lower occupancy rate of resource,better scalability,stability and reliability.