论文部分内容阅读
为解决智能电网系统中能量管理系统(energy management system,EMS)数据指数级增长带来的海量数据存储和快速访问问题,提出一种基于分布式键值对存储EMS数据的方法,将结构化与非结构化数据统一按照键值对形式进行表达,并分布存储于集群中。以该方法为基础,设计并实现了一种新的EMS数据库平台,该平台是混合数据库平台,支持关系库与非关系型数据库一体化访问,同时针对EMS的查询特征进行了系统优化,采用混合索引策略提升多维与高维数据的查询性能。实验结果表明:与传统EMS数据库管理系统相比,新的EMS数据库平台能够满足EMS海量多业务数据的一体化存储需求,支持高效并发访问,具有良好的扩展性和容错能力。另外,该系统可以在商用计算机上运行,能够有效地降低系统成本。
In order to solve the problem of massive data storage and quick access caused by the exponential growth of energy management system (EMS) data in smart grid system, a method of storing EMS data based on distributed key-value pairs is proposed, The unstructured data are expressed in the form of key-value pairs and stored in clusters. Based on this method, a new EMS database platform is designed and implemented. The platform is a hybrid database platform that supports the integrated access of relational database and non-relational database. At the same time, it systematically optimizes the query characteristics of EMS, Indexing strategies improve the query performance of multidimensional and high dimensional data. The experimental results show that compared with the traditional EMS database management system, the new EMS database platform can meet the integrated storage requirements of EMS massive multi-service data, supports efficient concurrent access, and has good scalability and fault tolerance. In addition, the system can run on a commercial computer, can effectively reduce system costs.