海量连续运行参考站网数据云存储模型  

Cloud Storage Model of Massive Continuous Operating Reference Station System Data

作　　者：李林阳[1] 吕志平[1] 陈正生[1] 樊黎晖 

机构地区：[1]信息工程大学地理空间信息学院,郑州450052 [2]成都测绘信息中心,成都610000

出　　处：《导航定位学报》2014年第3期64-70,86,共8页Journal of Navigation and Positioning

基　　金：国家自然科学基金(41274015);国家863计划项目(2013AA122501)

摘　　要：国家、区域、行业级连续运行参考站网的不断建成及连续观测,连续运行参考站网的数据量呈几何倍数增长,面对海量连续运行参考站网数据,集中式数据管理的可靠性、扩展性、抗灾性和容错性受到限制,云存储将网络中的大量存储设备协同起来工作,具有高抗灾容错性、高扩展性、高数据写入和读取性能,同时支持大吞吐量数据的并发访问,可实现连续运行参考站网数据的高效管理。基于分布式处理平台Hadoop,提出了连续运行参考站网数据云存储模型,可将已建成的连续运行参考站网站、数据中心纳入到云存储集群。实验搭建了Hadoop集群,利用中国大陆构造环境监测网络数据,建立了存储目录,执行了连续运行参考站网数据上传、下载、生成指定采样率文件等操作,测试了集群的可靠性、扩展性、抗灾容错性和下载速率。作为新的连续运行参考站网数据管理方式,云存储提高了连续运行参考站网数据的管理效率。With the construction of continuous operating reference station system （CORS） in the level of nation, region and industry and continuous observation, the quantity of CORS data is increasing exponentially. Centralized storage is restricted with the aspect of reliability, expansibility, disaster resistance and fault tolerance because of massive CORS data. Cloud storage is disaster-resistant, fault-tolerant, extremely easy to expand and very fast in data writing and reading, which makes Lots of storage device work together and supports concurrent access to a large scale of data, and manages CORS data effectively. Based on distributed data processing software--Hadoop, the CORS data cloud storage model is provided, those existed CORS stations, data centers can be accepted in the storage cluster. In the experiment, totally distributed cluster is set up, storage directory of CORS data is established, data obtained from crustal movement observation network of China received in 2013 is uploaded and downloaded., and special rate files are derived, the reliability, expansibility, disaster resistance tolerance, fault tolerance and download rate of cloud storage are tested. As a new CORS data management method, cloud storage increases the management efficiency of CORS data.

关 键 词：连续运行参考站网 数据管理 云存储 HADOOP 分布式文件系统 

分 类 号：P228[天文地球—大地测量学与测量工程]