Stream Management Systems流管理系统

Stream Management Systems流管理系统

ID:40844250

大小:209.67 KB

页数:10页

时间:2019-08-08

Stream Management Systems流管理系统_第1页
Stream Management Systems流管理系统_第2页
Stream Management Systems流管理系统_第3页
Stream Management Systems流管理系统_第4页
Stream Management Systems流管理系统_第5页
资源描述:

《Stream Management Systems流管理系统》由会员上传分享,免费在线阅读,更多相关内容在教育资源-天天文库

1、Control-BasedQualityAdaptationinDataStreamManagementSystemsYi-ChengTu1,MohamedHefeeda2,YuniXia1,SunilPrabhakar1,andSongLiu11PurdueUniversity,WestLafayette,IN47906,U.S.A.2SimonFraserUniversity,Surrey,BC,CanadaAbstract.UnlikeprocessingsnapshotqueriesinatraditionalDBMS,theprocessingofcon

2、tinuousqueriesinadatastreammanagementsystem(DSMS)needstosatisfyqualityrequirementssuchasprocessingdelay.Whenthesystemisoverloaded,qualitydegradessignificantlythusloadsheddingbecomesnecessary.Maintainingthequalityofqueriesisadif-ficultproblembecauseboththeprocessingcostanddataarrivalrate

3、arehighlyunpredictable.Weproposeaqualityadaptationframeworkthatadjuststheapplicationbehaviorbasedonthecurrentsystemsta-tus.Weleveragetechniquesfromtheareaofcontroltheoryindesigningthequalityadaptationframework.Oursimulationresultsdemonstratetheeffectivenessofthecontrol-basedqualityadap

4、tationstrategy.Com-paringtosolutionsproposedinpreviousworks,ourapproachachievessignificantlybetterqualitywithlesswasteofresources.1IntroductionThepastfewyearshavewitnessedtheemergenceofapplicationssuchasnetworkmonitoring,financialdataanalysis,location-basedservices,andsensornetworksinwh

5、ichinformationnaturallyoccursintheformofcontinuousdatastreams.Theseapplicationsdonotfitinthedatamodelandqueryingparadigmoftra-ditionalDatabaseManagementSystems(DBMSs).Datastreammanagementsystems(DSMSs)suchasAurora[1],STREAM[2],andTelegraphCQ[3]arebuiltfordatamanagementandqueryprocessin

6、ginthepresenceofmultiple,continuous,andtime-varyingdatastreams.Insteadofbeingstoredinadvanceondisks,streamingdataelementsarriveon-lineandstayonlyforalimitedtimeperiodinmemory.Consequently,aDSMShastohandlethedataelementsbeforethebufferisoverwrittenbynewincomingdataelements.Processingofd

7、atastreamsbringsgreatchallengestoDBMSdesignfortwomajorreasons.First,streamdataarecontinuouslygeneratedinlargevolumesbyexternalsourcessuchasasensornetwork.Second,thedataareoftencollectedfromremotesitesandarepronetolossanddelayinmostcases.Fromtheappli-cationspointofview,queriestodatastr

8、eamsa

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。