摘要
随着遥感技术的发展,遥感卫星数量增加、种类日趋丰富、分辨率不断提高,导致遥感数据量急速增长,给数据的存储管理效率提出了更高的要求。传统方法通常采用规则的方式对数据价值进行评估并形成相应的存储迁移策略,具有较强的主观因素,并未完全从数据的角度出发分析产品热度。从数据的访问规律着手,将数据分为访问频率高的热数据,访问频率相对较低的温数据和冷数据,利用机器学习算法对数据热度进行建模,结合磁盘阵列与蓝光光盘库等硬件存储设施特点,设计并实现一套基于蓝光光盘库与SAN存储的海量遥感数据的分级存储系统,并通过优化数据的迁移存储策略,实现海量遥感数据的高效、安全、节能存储。在实验平台上就2018年以来的产品需求进行了实验验证,表明采用提出的存储迁移策略比现有策略下系统服务请求的数据在线率提高44.4%。
With the development of remote sensing technology,there are more and more kinds of remote sensing data.The resolution of remote sensing image improved fast and the volume of remote sensing data growth rapidly.It brings pressure to data storage and management,which makes it significant to study the storage and migration strategy for remote sensing data.The traditional data storage management models often make the storage and migration strategy by evaluating data in regular way.This method has strong subjective factors and do not completely think about the influence of data.Based on the data accessing,this paper sorts the data as hot data,secondary data and cold data.Using the machine learning algorithms to modeling the hot data and based on the model we studied the hierarchical storage and the migration strategy.A remote sensing data hierarchical storage system based on Blue light storage and SAN storage is designed and implemented.We implement high efficiency storage of massive remote sensing data by optimizing migration strategy for data storage.Experiments show that the storage migration strategy proposed in this paper improves the service efficiency by 44.4%compared with the existing strategies.
作者
赵泽亚
杨迪
梁小虎
王荣
金雪
ZHAO Zeya;YANG Di;LIANG Xiaohu;WANG Rong;JIN Xue(Beijing Institute of tracking and Telecommunications Technology, Beijing 100094, China)
出处
《信息工程大学学报》
2020年第1期115-119,共5页
Journal of Information Engineering University
关键词
遥感数据
数据热度
迁移策略
机器学习
remote sensing data
data access popularity
migration strategy
machine learning
