面向混合存储的时空大数据存储策略研究

Research on Spatio-temporal Big Data Storage Strategy for Hybrid Storage

混合存储能综合各类管理系统和设备的优点,成为时空大数据存储的理想选择,但不同管理系统下的时空大数据之间难以建立有效关联,且数据存储策略缺少对时空大数据的动态感知和弹性调度能力。分析了当前典型混合存储技术产品,根据时空大数据特性,提出了面向混合存储的时空大数据存储策略。通过建立数据存储模型,在数据层面实现一个完整的时空大数据存储生态链,设计数据迁移函数来量化数据价值,灵活动态地匹配存储设备。测试表明,混合存储策略能有效提升时空大数据的存储访问性能,更好地服务于时空大数据的管理应用。

The hybrid storage has become an ideal solution for spatio-temporal big data storage,which can fully take advantage of both hybrid storage management systems and hybrid storage devices.However,it is difficult to establish effective correlation between spatio-temporal big data under different management systems,and the data storage strategy can neither dynamically perceive spatio-temporal big data nor adjust the position of spatio-temporal big data in time.According to the characteristics of spatio-temporal big data,we proposed a storage strategy for hybrid storage by the analysis of current typical hybrid storage technology products.We realized a complete spatio-temporal big data storage ecological chain at the data level by building organization and storage model.Spatio-temporal big data could match storage devices flexibly and dynamically by designing data migration function to quantify data value.The test result shows that the hybrid storage strategy can effectively improve the storage access performance of spatio-temporal big data,and better serve the management and application of spatio-temporal big data.