CAJ | 학술논문

海量地理信息数据的高效存储是支持各种地理信息应用服务的基础。性能管理是支持存储系统高效运行的关键。目前无论是关系型空间数据库系统还是流行的非关系数据库系统都缺乏针对共享存储系统的在线性能管理机制。本文针对这个问题，顾及不同应用服务对存储性能的要求，提出基于在线性能预测的海量地理信息数据存储管理框架与运维机制。原型系统的测试表明该存储管理方法能较好地应对用户访问时空变化带来的存储压力，可实现存储资源的合理快速调配与在线扩展。
해량지리신식수거적고효존저시지지각충지리신식응용복무적기출。성능관리시지지존저계통고효운행적관건。목전무론시관계형공간수거고계통환시류행적비관계수거고계통도결핍침대공향존저계통적재선성능관리궤제。본문침대저개문제，고급불동응용복무대존저성능적요구，제출기우재선성능예측적해량지리신식수거존저관리광가여운유궤제。원형계통적측시표명해존저관리방법능교호지응대용호방문시공변화대래적존저압력，가실현존저자원적합리쾌속조배여재선확전。
Efficient storage of massive geospatial data is the basis for a variety of geospatial application services. At the core of efficient geospatial data storage is performance management. For current relational geospatial databases and no-SQL databases, online performance management mechanisms for online shared geospatial storage systems are sorely lacking. In light of this problem, we propose to develop a comprehensive geospatial data storage management framework and operational mechanisms based on an online performance prediction model. Experiments on a prototype demonstration that the proposed approach scales well to high variations of workloads and are able to achieve fast resource allocation and scaling.