物联网海量数据的分布式存储算法  

Mass Data Distributed Storage Algorithms in the Internet of Things

作　　者：薛建生[1] 于忠臣[1] 黄磊[1] 赵巍[1] 

机构地区：[1]辽宁大学信息学院,沈阳110036

出　　处：《小型微型计算机系统》2013年第9期2081-2084,共4页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(60703068)资助;辽宁大学"211"工程三期子项目资助

摘　　要：随着物联网技术的进一步发展以及应用规模的扩大,网络中的数据量呈几何级的疯狂增长,众多传感器终端的大量信源势必呈海量趋势.面对不断增长的数据,海量数据的存储不均匀的问题将成为限制物联网充分发展的"瓶颈".本文结合分布式云存储系统的架构性优势,在其基础上针对物联网异类数据实时存取特点,在多个数据中心对海量数据进行边缘化存储处理.为此本文提出了一种点对点的分布式存储方案,在方案中物联网各数据中心间采用云架构分级式存储,通过存储信息的交互推动,使各数据中心协同工作,达到对物联网中原始数据流有序分散负载均衡.实验证明,本文提出的方案能够实现物联网中海量数据的边缘化存储和负载均衡.With the further development of the Things of Internet technology, and the application of the expansion, the amount of data in the network was crazy geometric growth, the source of many sensor terminal was bound mass. To face the rapid growth of data, mass data storage marginalization of things will be restricted the full development of the ＂bottleneck ＂ . This paper combined with the advantage of distributed cloud storage system architecture for the Internet of Things, compared to objects based on networked real-time heterogeneous data access features, to intelligently marginalized store huge amounts of data in multiple data centers. Propose a peer to peer distributed storage solutions, The program of things between the data center using hierarchical storage, Communicate with each other through message passing to store information exchange, so that the data center work cooperatively, Achieve things in the origi- nal data stream, distributed storage, and ultimately things marginalization of vast amounts of data stored in memory. Experiments show that：The proposed scheme of things in the marginalization of the vast amounts of data storage and load balancing can be a- chieved.

关 键 词：分布式文件系统 云存储 物联网 云计算 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]