基于时序库的分层存储技术设计与实现

Design and Implementation of Stratified Storage Technology Based on Time-sequential Databases

针对电力调度自动化系统实际运行过程中随着接入数据点数量的不断增加,逐渐出现历史数据查询速度慢或统计速度慢的问题,提出一种基于时序库的分层存储技术架构。基于原有的实时库结合关系库的存储和查询模式,建立时序库的缓存机制,在时序库中存储对象idkey信息和近期热点数据,实现分层存储和查询,加快查询和统计速度。具体实现过程中,为了对上层应用屏蔽不同时序库的使用差异,便于不同的工程投运现场选用不同类型的时序库,采用动态加载的开发方式提供一套通用的时序库调用接口,上层应用使用同一套程序可完成对不同类型时序库的读写,有效降低了上层应用的开发工作量。

In the actual operation process of the electric power automatic scheduling system,the continuous increase of accessed data points gradually results in sluggish historical data queries or statistics.In view of this,the present work studied a technological architecture of stratified storage based on time-sequential databases.It integrates the storage and query modes of the original real-time database with the establishment of a caching mechanism in the time series database to store object ID key information and recent hot data,thereby achieving hierarchical storage and queries,and accelerating query and statistical speeds.In the specific implementation process,in order to shield the differences in the use of different time-sequential databases for upper-layer applications and facilitate the selection of different types of time-sequential databases in different engineering operation sites,a dynamic loading development approach is adopted to provide a set of universal time-sequential database invocation interfaces.Upper-layer applications can use the same set of programs to complete the read and write operations for different types of time-sequential databases,effectively reducing the development workload of upper-layer applications.