集群排队的铁道供电海量准实时数据异步并行查询

Railway Power Line Mass Cluster Near Real Time Data and Asynchronous Parallel Query

针对铁道供电的海量准实时数据查询效率不高造成信息处理效率低下的问题,提出一种基于集群队列模型的铁道供电监控海量信息快速异步并行查询新方法。基于实时Ajax数据引擎,综合运用MPP查询架构,建立供电监控信息的轮询队列模型,实现海量准实时数据在铁道供电监控系统中的快速查询处理。以四机集群和一台查询测试机为平台,以铁道10kV配电网远动调度监控千万级时序数据为算例,进行加载测试和集群查询性能测试。结果表明:MPP结合Ajax异步回调机制,集群磁盘I/O的读取速度约为存储速度的10倍,能够将查询服务器端的交互更新降至数百毫秒级,验证了新的集群监控方法优于常规双数据库服务器并列运行的模式。

Aiming at the problem of inefficient information processing caused by the low efficiency of the railway power supply massive near real-time data query,a new fast and asynchronous parallel query method was proposed based on cluster queue model for railway power supply monitoring massive information.Based on the real-time Ajax data engine,the MPP query framework was used to to establish the polling queue model of power supply monitoring and control information,to achieve rapid query processing of mass near real time data in railway power supply monitoring system.With four clusters and a query test rig as the platform,based on the tens of millions of time series data railway remote 10 kV distribution network operation monitoring system,the loading test and cluster query performance test were carried out.The results show that when MPP is combined with Ajax asynchronous callback mechanism,the read speed of cluster disk I/O is about 10 times the storage speed,which can reduce the interactive server query update to hundreds of MS,and which verifies the new cluster monitoring method is superior to the conventional double database server in parallel running mode.