基于节点中心性和热度的数据布局方法  

Data Layout Method Based on Node Centrality and Hotness

作　　者：汪雨 韩锐[1,2] 党寿江 WANG Yu;HAN Rui;DANG Shoujiang(National Network New Media Engineering Research Center,Institute of Acoustics,Chinese Academy of Sciences,Beijing,100190,China;University of Chinese Academy of Sciences,Beijing,100049,China)

机构地区：[1]中国科学院声学研究所国家网络新媒体工程技术研究中心,北京100190 [2]中国科学院大学,北京100049

出　　处：《网络新媒体技术》2024年第5期34-41,共8页Network New Media Technology

基　　金：国家重点研发计划课题“面向东数西算的多模态网络应用示范”(编号:2023YFB2906404)。

摘　　要：随着5G和大数据的飞速发展,海量数据的快速、持久存储需求对存储设备和网络性能提出了极大挑战。信息中心网络以及边缘存储等技术的出现,使数据能够在网络边缘就近存储,减少了数据传输时延。但数据长期存储在网络边缘势必会引发边缘节点空间不足的问题,迫使数据传向更远方,降低存储效率。针对上述问题,提出一种基于节点中心性和热度的数据布局方法,利用节点空闲时间将高空间负载节点的数据迁移至低空间负载的节点,确保边缘高热度的节点空间富余。实验结果表明,相比较Random和一致性Hash等存储布局方案,在进行长期存储任务后,该方法能在边缘空间相对不足的情况下,减少存储数据时的传输开销,使数据写入时间减少30%以上。With the rapid development of 5G and big data,the demand for fast and persistent storage of massive data poses a great challenge to storage equipment and network performance.The emergence of information-center networks and technologies such as edge storage enables data to be stored near the edge of the network,which reduces the data transmission delay.However,long-term data storage at the edge of the network will inevitably lead to the problem of insufficient node space at the edge,forcing the data to be transmitted to a more distant place and reducing the storage efficiency.To address the above problems,a data layout method based on node centrality and hotness is proposed in the paper,which utilizes the idle time of nodes to migrate data from high space load nodes to low space load nodes,ensuring space surplus for nodes with high heat at the edges.The experimental results show that compared with the storage layout schemes such as random and consistent Hash,after a long-term storage task,the method reduces the transmission overhead when storing data in the presence of a relative lack of edge space,reducing the data writing time by more than 30%.

关 键 词：信息中心网络 数据存储 数据布局 负载均衡 中心性度量 

分 类 号：TP333[自动化与计算机技术—计算机系统结构]