基于字典表的边缘云文件同步方法  

Edge Computing File Synchronization Method Based on Dictionary

作　　者：林慕云 吴国驭 李强[1] 

机构地区：[1]杭州电子科技大学智能与软件研究所,浙江杭州310018 [2]杭州义益钛迪信息技术有限公司,浙江杭州310018

出　　处：《杭州电子科技大学学报（自然科学版）》2017年第6期30-35,共6页Journal of Hangzhou Dianzi University：Natural Sciences

基　　金：风力发电系统国家重点实验室资助项目(2013002)

摘　　要：针对边缘云架构因用户频繁异地登录带来的文件同步压力问题,提出了采用在核心云和边缘云各增加一张字典表记录文件变更记录,并以此表为依据定位目标文件的同步方法,较好地解决了分布式存储中查找目标文件效率的问题;再将传输时延碎片化,使用最早时限优先EDF策略进行调度,降低时延,使之拥有更好的用户体验.给出了该同步方法的具体算法和基于EDF的调度流程.模拟实验结果表明:在保证用户文件一致性和较低数据冗余率前提下,与基于网络延迟的文件传输方法比较,提出的新方法大幅降低了网络传输量和传输时延.Aimed at the file synchronization pressure problem in edge computing caused by frequent remote login,presents a new file synchronization method to record changes in the log and locate the target file by add each dictionary in the core server and the edge server.This method can solve the efficiency problem of finding the file in distributed storage,and fragmentates the file transfer delay by using EDF scheduling strategy to reduce the delay and get the better user experience.It gives the algorithm of the synchronization method and the scheduling process based on EDF in this paper.The experimental results show that under the premise of ensuring the data consistency and low data redundancy,the proposed method reduces network data traffic and data redundancy greatly compared with the file transfer method based on network delay.

关 键 词：云计算 边缘云架构 分布式存储 文件同步 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]