摘要
首先设计并开发了面向对象的内存数据库,采用了分区、分页存储技术,可用于大规模电网模型的高效缓存和检索[4-5]。该内存对象数据库集成进一个实时GIS平台,为实现了大规模并发的拓扑追踪服务提供模型数据。其次本文介绍了基于任务调度队列的拓扑追踪和开关操作系列的并行处理技术,解决了多客户请求的并行执行及其冲突问题。接着本文针对跨分区的操作请求,采用父子任务的形式加入任务队列。最后本文在某省上线系统中进行了压力测试,本文开发的系统在千万级的电网设备模型规模下,每台服务器1 000并发请求的处理时间小于0.1秒。
Firstly, the object oriented memory database is designed and developed, which can be used in the efficient caching and retrieval of large scale power grid model by using the partition and paging storage technology. The memory object database is integrated into a real-time GIS platform, which provides the model data for the large-scale concurrent topology tracking service. Secondly, this paper introduces the parallel processing technology of the topology tracking and switching operation series, which solves the problem of parallel execution of muhi-client requests. Then, in this paper, the task queue is added in the form of a father-son task to a cross partition operation request. Finally, in the province of on-line system of the pressure test, the system developed in the scale of tens of millions of grid equipment model, each server processing time of 1 000 concurrent requests in less than 0. 1 seconds.
出处
《贵州电力技术》
2016年第11期29-34,共6页
Guizhou Electric Power Technology
