Storage device performance prediction with CART models(16)

Storage device performance prediction is a key element of self-managed storage systems and application planning tasks, such as data assignment. This work explores the application of a machine learning tool, CART models, to storage device modeling. Our appr

References

[1]N.Allen.Don’twasteyourstoragedollars:Whatyouneedtoknow.Researchnote,GartnerGroup,

2001.[2]E.Anderson,M.Kallahalla,S.Spence,R.Swaminathan,andQ.Wang.Ergastulum:Quickly nding

near-optimalstoragesystemdesigns.TechnicalReportHPL-SSP-2001-05,HPLabs,2001.[3]EricAnderson.Simpletable-basedmodelingofstoragedevices.TechnicalReportHPL-SSP-2001-4,

HPLabs,2001.[4]L.Brieman,J.H.Friedman,R.A.Olshen,andC.J.Stone.Classi cationandRegressionTrees.

Wadsworth,1984.[5]JohnBucy,GregGanger,andcontributors.TheDiskSimsimulationenvironmentversion3.0reference

manual.TechnicalReportCMU-CS-03-102,CarnegieMellonUniversity,2003.[6]C.J.C.Burges.AtutorialonSupportVectorMachinesforpatternrecognition.KnowledgeDiscovery

andDataMining,2(2):121–167.[7]ShenzeChenandDonTowsley.AperformanceevaluationofRAIDarchitectures.IEEETransactions

onComputers,45(10):1116–1130,1996.[8]MarkE.CrovellaandAzerBestavros.Self-similarityinworldwidewebtraf cevidenceandpos-siblecauses.InProc.ofthe1996ACMSIGMETRICSIntl.Conf.onMeasurementandModelingofComputerSystems,May1996.[9]B.Dasarathy.Neareastneighborpatternclassi cationtechniques.IEEEComputerSocietyPress,

1991.[10]GregoryR.Ganger.Generatingrepresentativesyntheticworkloads:Anunsolvedproblem.InPro-ceedingsoftheComputerMeasurementGroup(CMG)Conference,pages1263–1269,1995.[11]GregoryR.Ganger,JohnD.Strunk,andAndrewJ.Klosterman.Self-*storage:Brick-basedstorage

withautomatedadministration.TechnicalReportCMU-CS-03-178,CarnegieMellonUniversity,2003.[12]GartnerGroup.Totalcostofstorageownership—Auser-orientedapproach.Researchnote,Gartner

Group,2000.[13]Mar´iaE.G´omezandVicenteSantonja.Analysisofself-similarityinI/Oworkloadusingstructural

modeling.In7thInternationalSymposiumonModeling,AnalysisandSimulationofComputerandTelecommunicationSystems,pages234–243,1999.[14]Mar´iaE.G´omezandVicenteSantonja.Anewapproachinthemodelingandgenerationofsynthetic

diskworkload.InProceedingsof9thInternationalSymposiumonModeling,AnalysisandSimulationofComputerandTelecommunicationSystems,pages199–206,2000.[15]J.Gray.AconversationwithJimGray.ACMQueue,1(4),2003.

[16]TrevorHastie,RobertTibshirani,andJeromeFriedman.TheElementsofStatisticalLearning:Data

Mining,Inference,andPrediction.Springer,2001.[17]InternationalBusinessMachinesCorp.Autonomiccomputing:IBM’sperspectiveonthestateof

informationtechnology./autonomic/manifesto/.