摘要
站点数据是气象观测数据的主要组成部分,每日检索量大,且时效要求高。为了更好地管理与服务气象资料,国家气象信息中心从2013年开始致力于新存储技术在气象领域的探索。为了更好地兼容目前的存储与服务方式,选择了分布式关系型数据库作为技术方案,并开展了一系列仿真气象业务场景的测试。测试结果表明,分布式数据库无论在数据插入还是在检索方面都具有较好的性能,同时具有较好的可靠性、灵活性和扩展性。根据测试结论,搭建了分布式存储试验系统,导入了CIMISS地面、高空等建站以来的历史数据,并接入了站点数据的实时流程。实现了海量站点资料的完全在线,同时也满足了气象业务对资料高时效性的需求。
The station observation data is the main component of the meteorological observation data,which has high access and high timeliness needs. In order to manage and serve meteorological data better,the National Meteorological Information Center began to focus on new storage technology in the field of Meteorology from 2013. In order to be compatible with the current storage and service mode,the distributed relational database is selected as a technical scheme,and a series of simulation weather business scenarios testing were carried out. The results show that the distributed database has better performance in terms of data loader and retrieval,and has good reliability,flexibility and scalability. According to the results,the distributed storage system was built,and the historical data of the CIMISS were introduced,and the real- time process flow of the station data was accessed. This system have completed the goal of the stations information online,at the same time,it also meets the requirement of the meteorological data service for the high efficiency.
出处
《贵州气象》
2016年第5期61-68,共8页
Journal of Guizhou Meteorology
基金
中国气象局预报预测核心业务发展专项面上项目"分布式存储在统一气象数据环境中应用和推广"(CMAHX20160705)
关键词
气象站点数据
分布式存储
自动站
性能测试
data of meteorological stations
distributed storage
automatic meteorological station
performance testing
