支持大规模地震探测数据快速可视化的云端数据缓存技术  

Cloud Data Cache Technology Supporting Rapid Visualization of Large-Scale Seismic Exploration Data

作　　者：魏晓辉[1] 崔浩龙 李洪亮[1] 白鑫[1] WEI Xiaohui;CUI Haolong;LI Hongliang;BAI Xin(College of Computer Science and Technology,Jilin University,Chan gchun 130012,China)

机构地区：[1]吉林大学计算机科学与技术学院,长春130012

出　　处：《吉林大学学报（理学版）》2018年第5期1147-1155,共9页Journal of Jilin University:Science Edition

基　　金：国家自然科学基金(批准号:61602205;51627805;61170004);国家重点研发计划专项基金(批准号:2016YFB0201503;2016YFB0701101);教育部高等学校博士学科点专项科研基金(批准号:20130061110052);吉林省科技攻关计划重大科技招标专项基金(批准号:20160203008GX);吉林省科技攻关计划重点科技攻关项目(批准号:20140204013GX);吉林省科技发展计划项目(批准号:20170520066JH)

摘　　要：首先,基于云计算应用模式,提出一种能有效利用云存储架构的双层缓存技术.通过在客户端和服务器端建立分布式缓存,能有效避免用户频繁访问远端数据,为用户构建轻量级的客户端,解决了目前地学数据可视化软件大量占用用户本地存储容量的问题.同时服务器端也避免了多次访问云存储文件系统,减少了大量的数据检索与加载时间.其次,提出一种ARLS(association rule last successor)访问预测算法,根据用户的历史访问记录,利用关联规则挖掘用户的访问模式,对其访问行为进行预测,进而提前加载数据,提高缓存命中率,解决了用户在可视化过程中不断移动兴趣区域,频繁更换渲染数据的问题,能有效应对用户具有多种访问模式的情况,提高了预测准确率.实验结果表明,该云存储架构显著减少了本地资源消耗,访问预测算法的准确率在最差情形下可达47.59%,平均准确率达91.3%,分布式缓存的平均缓存命中率达95.61%,可有效支持云端大规模地震数据的快速可视化.Firstly,based on the cloud computing application model,we proposed a double-layer cache technology which could efficiently utilize the cloud storage architecture.By establishing the distributed cache between the client and the server,it could effectively avoid users frequent access to remote data and build lightweight clients for users,which solved the problem that current geoscience data visualizat ion software occupied a large number of user’s local storage capacity,and adap t to the rapid development of mobile devices.In the mean time,the server side also avoided multiple access to the cloud storage file system,reducing a lot of data r etrieval and loading time.Secondly,we proposed an association rule last successor a ccess prediction algorithm,according to user’s historical access records,the asso ciation rules were used to mine the user’s access mode,and predict their access behavior.Then the data was loaded in advance,the cache hit rate was improved,we solved the problem of constantly moving region of interest and changing the rendering data frequently in the process of visualization,our system could eff ectively deal with the user’s multiple access patterns case and improve the accuracy of the prediction.Experimental results show that the cloud storage architecture significantly red uces the local resource consumption.The accuracy rate of the access prediction algorithm is 47.59%in the worst case,the avera ge accuracy rate is 91.3%,and the average cache hit rate of distributed cache i s 95.61%,which can effectively support the rapid visualization of large-scale sei smic data in the cloud.

关 键 词：云存储架构 双层缓存 大数据索引 访问预测 快速可视化 网络通信 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]