分布式数据库Greenplum在地震前兆数据存储中的应用*
陈晓琳,李盛乐,刘 坚,刘珠妹

(中国地震局地震研究所 地震大地测量重点实验室,湖北 武汉 430071)

大数据存储; 前兆数据; 分布式数据库; Greenplum数据库

Application of Distributed Database Greenplum in Earthquake Precursor Data Storage
CHEN Xiaolin,LI Shengle,LIU Jian,LIU Zhumei

(Key Laboratory of Earthquake Geodesy,Institute of Seismology,China Earthquake Administration,Wuhan 430071,Hubei,China)

备注

随着地震前兆观测台网的加密、采样率的提高,地震前兆观测数据量也在快速增加。在进行地震数据共享服务时,需要快速获得大量数据集,无疑对前兆共享数据库的数据处理能力提出了更高的要求。针对这一问题,提出基于Greenplum数据库的地震前兆数据存储设计方案。通过搭建Greenplum分布式数据库环境,实现了海量前兆数据的快速处理,并与传统Oracle数据库进行对比,结果表明:Greenplum数据库读取前兆数据耗时更低,对于大批量数据的读取操作,Greenplum数据库的优势更加明显; Greenplum数据库良好的可扩展性和对应用编程接口(JDBC、ODBC)的支持,使得其在前兆数据分析处理中的应用前景广阔。

With the encryption of precursor observation networks and the improvement of sampling rate,the quantity of earthquake precursor observation data increases rapidly.It is a higher requirement for the data processing ability of the precursor database in the seismic data sharing service that how to make users quickly obtain large data sets.Aiming at this problem,the Greenplum-based precursor data storage method is proposed.Through setting up the Greenplum distributed database environment,the rapid processing of massive precursor data is realized.Performance testing experiments of precursor data show that the Greenplum database takes less time to read the precursor data,and the advantages of Greenplum database are more obvious for large-volume data operations,compared with the traditional Oracle11g database.In addition,the good extensibility of Greenplum database and the support of application interface(JDBC,ODBC)make its promising application in precursor data analysis and processing.