學術(shù)英語論文

1、南昌大亨NANCHANGUNIVERSITY課程名稱：學術(shù)英語題目：AStudyofEnergyEfficientCloudComputingPoweredby_WirelessEnergyTransfer英語班級：理工1615班專業(yè)/年級：物聯(lián)網(wǎng)工程161班姓名/學號：(47L二零一八年六月AStudyofEnergyEfficientCloudComputingPoweredbyWirelessEnergyTransferAStudyofEnergyEf?cientMobileCloudComputingPoweredbyWirelessEnergyTransferAbstractAchi

2、evinglongbatterylivesorevenself-sustainabilityhasbeenalongstandingchallengefordesigningmobiledevices.Thisstudypresentsanovelsolutionthatseamlesslyintegratestwotechnologies,mobilecloudcomputingandmicrowavepowertransfer(MPT),toenablecomputationinpassivelow-complexitydevicessuchassensorsandwearablecomp

3、utingdevices.Specifically,consideringasingle-usersystem,abasestation(BS)eithertransferspowertooroffloadscomputationfromamobiletothecloud;themobileusesharvestedenergytocomputegivendataeitherlocallyorbyoffloading.AframeworkforenergyefficientcomputingisproposedthatcomprisesasetofpoliciesforcontrollingC

4、PUcyclesforthemodeoflocalcomputing,timedivisionbetweenMPTandoffloadingfortheothermodeofoffloading,andmodeselection.GiventheCPU-cyclestatisticsinformationandchannelstateinformation(CSI),thepoliciesaimatmaximizingtheprobabilityofsuccessfullycomputinggivendata,calledcomputingprobability,undertheenergyh

5、arvestinganddeadlineconstraints.Furthermore,thisstudyrevealsthatthetwosimplesolutionstoachievetheobjecttosupportcomputationloadallocationovermultiplechannelrealizations,whichfurtherincreasesthecomputingprobability.Last,thetwokindsofmodessuggestthatthefeasibilityofwirelesslypoweredmobilecloudcomputin

6、gandthegainofitsoptimalcontrol.Andthefutureaspecttostudyissimplytobeanswer.Keywords:wirelesspowertransfer;energyharvestingcommunications;mobilecloudcomputing;energyefficientcomputingAStudyofEnergyEfficientCloudComputingPoweredbyWirelessEnergyTransferIntroductionMobilecloudcomputing(MCC)asanemergingc

7、omputingparadigmintegratescloudcomputingandmobilecomputingtoenhancethecomputationperformanceofmobiledevices.TheobjectiveofMCCistoextendpowerfulcomputingcapabilityoftheresource-richcloudstotheresource-constrainedmobiledevices(e.g.,laptop,tabletandsmartphone)soastoreducecomputationtime,conservelocalre

8、sources,especiallybattery,andextendstoragecapacity.Toachievethisobjective,MCCneedstotransferresource-intensivecomputationsfrommobiledevicestoclouds,referredtoascomputationoffloading.Thecoreofcomputationoffloadingistodecideonwhichcomputationtasksshouldbeexecutedonthemobiledeviceoronthecloud,andhowtos

9、chedulelocalandcloudresourcetoimplementtaskoffloading.TheexplosivegrowthofInternetofThings(IOT)andmobilecommunicationisleadingtothedeploymentoftensofbillionsofcloud-basedmobilesensorsandwearablecomputingdevicesinnearfuture(Huang&Chae,2010).Prolongingtheirbatterylivesandenhancingtheircomputingcap

10、abilitiesaretwokeydesignchallenges.Theycanbetackledbytwopromisingtechnologies:microwavepowertransfer(MPT)forpoweringthemobilescomputation-intensivetasksfromthemobilestothecloudandmobilecomputationoffloading(MCO).Twotechnologiesareseamlesslyintegratedinthecurrentworktodevelopanoveldesignframeworkforr

11、ealizingwirelesslypoweredmobilecloudcomputingunderthecriterionofmaximizingtheprobabilityofsuccessfullycomputinggivendata,calledcomputingprobability.TheframeworkisfeasiblesinceMPThasbeenproveninvariousexperimentsforpoweringsmalldevicessuchassensorsorevensmall-scaleairplanesandhelicopters.Furthermore,

12、sensorsandwearablecomputingdevicestargetedintheframeworkareexpectedtobeconnectedbythecloud-basedIOTinthefuture,providingasuitableplatformforrealizingMCO.MaterialsMCOhasbeenanactiveresearchareaincomputersciencewhereresearchhasfocusedondesigningmobile-cloudsystemsandsoftwarearchitectures,virtualmachin

13、emigrationdesigninthecloudandcodepartitioningtechniquesinthemobilesforreducingtheenergyconsumptionandimprovingthecomputingperformanceofmobiles.Nevertheless,implementationofMCOrequiresdatatransmissionandmessagepassingoverwirelesschannels,incurringtransmissionpowerconsumption.Theexistenceofsuchatradeo

14、ffhasmotivatedcross-disciplinaryresearchonjointlydesigningMCOandadaptivetransmissionalgorithmstomaximizethemobileenergysavings.Astochasticcontrolalgorithmwasproposedforadaptingtheoffloadedcomponentsofanapplicationtoatime-varyingwirelesschannel.Furthermore,multiusercomputationoffloadinginamulti-cells

15、ystemwasexploredbyShinohara(2014),wheretheradioandcomputationalresourceswerejointlyallocatedformaximizingtheenergysavingsunderthelatencyconstraints.AccordingtoSwan(2012),thethreshold-basedoffloadingpolicywasderivedforthesystemwithintermittentconnectivitybetweenthemobileandcloud.Lastly,theCPU-cyclefr

16、equenciesarejointlycontrolledwithMCOgivenamoreskilledandincreasinglyappropriateAStudyofEnergyEfficientCloudComputingPoweredbyWirelessEnergyTransferwirelesschannel.TheframeworkisfurtherdevelopedinthecurrentworktoincludethenewfeatureofMPT(Kostaetal.,2012).Thisintroducesseveralnewdesignchallenges.Among

17、others,thealgorithmicdesignoflocalcomputingandoffloadingbecomesmorecomplexundertheenergyharvestingconstraintduetoMPT,whichpreventsenergyconsumptionfromexceedingtheamountofharvestedenergyateverytimeinstant.AnotherchallengeisthatMPTandoffloadingtimesharethemobileantennaandthetimedivisionhastobeoptimiz

18、ed.Nowthetechnologyisbeingfurtherdevelopedtopowerwirelesscommunications.Thishasresultedintheemergenceofanactivefieldcalledsimultaneouswirelessinformationandpowertransfer(SWIPT).TheMPTtechnologyhasbeendevelopedforpoint-to-pointhighpowertransmissioninthepastdecades(Brown,1984).Furthermore,existingwire

19、lessnetworkssuchascognitiveradioandcellularnetworkshavebeenredesignedtofeatureMPT.MostpriorworkonSWIPTaimsatoptimizingcommunicationtechniquestomaximizetheMPTefficiencyandsystemthroughput.Incontrast,thecurrentworkfocusesonoptimizingthelocalcomputingandoffloadingunderadifferentdesigncriterionofmaximum

20、computingprobability(Huang&Lau,2014).MethodsandResultsConsiderasingle-usersystemcomprisingonemulti-antennabasestation(BS)usingtransmit/receivebeamformingfortransferringpowertoasingle-antennamobileorrelayingoffloadeddatafromthemobiletothecloud.Tocomputeafixedamountofdata,themobileoperatesinoneoft

21、hetwoavailablemodes:Localcomputingandoffloading:inthemodeoflocalcomputing,MPToccurssimultaneouslyascomputingbasedonthecontrollableCPU-cyclefrequencies.Nevertheless,inthemodeofoffloading,thegivencomputationdurationisadaptivelypartitionedforseparateMPTandoffloadingsincetheysharethemobileantenna(Shinoh

22、ara,2014).AssumethatthemobilehastheknowledgeofstatisticsinformationofCPUcyclesandchannelstateinformation(CSI).Theindividualmodesaswellasmodeselectionareoptimizedformaximizingthecomputingprobabilityundertheenergyharvestinganddeadlineconstraints.Fortractability,themetricistransformedintoequivalentones

23、,namelyaveragemobileenergyconsumptionandmobileenergysavings,forthemodesoflocalcomputingandoffloading,respectively.Comparedwiththepriorwork,thecurrentworkintegratesMPTwiththemobilecloudcomputing,whichintroducesnewtheoreticalchallenges.Inparticular,theenergyharvestingconstraintarisingfromMPTmakestheop

24、timizationproblemforlocalcomputingnon-convex.Totacklethechallenge,theconvexrelaxationtechniqueisappliedwithoutcompromisingtheoptimalityofthesolution.ItisshowninthesequelthatthelocalcomputingpolicyisaspecialcaseofthecurrentworkwherethetransferredpowerissufficientlyhighbySwan(2012).Furthermore,thecase

25、ofdynamicchannelformobilecloudcomputingisexplored.Approximationmethodsareusedforderivingthesimpleandclose-to-optimalpolicies.Mobilemodeselection:Theaboveresultsarecombinedtoselectthemobilemodeformaximizingthecomputingprobability.Givenfeasiblecomputinginbothmodes,theonlyoneAStudyofEnergyEfficientClou

26、dComputingPoweredbyWirelessEnergyTransferyieldingthelargerenergysavingsispreferredandtheselectioncriterionisderivedintermsofthresholdsontheBStransmissionpoweraswellasthedeadlineforcomputing(Huangetal.,2012).Optimaldataallocationforadynamicchannel:Last,theaboveresultsareextendedtothecaseofadynamiccha

27、nnel,modeledasindependentandidenticallydistributed.blockfading,andnon-causalCSIatthemobile(acquiredfrome.g.,channelprediction).Theproblemofoptimizinganindividualmobilemode(localcomputingoroffloading)isformulatedbasedonthemaster-and-slavemodelusingthesamemetricasthefixed-channelcounterpart(Kumar&

28、Liu,2013).ConclusionWirelessandmobilecomputingtechnologiesprovidemorepossibilitiesforaccessingservicesconveniently.Mobiledeviceswillbeimprovedintermsofpower,CPU,andstorage.Mobilecloudcomputinghasemergedasanewparadigmandextensionofcloudcomputing.Bytwokindsofavailablemodes,wecanpurelyknowoftheEnergyEf

29、?cientMobileCloudComputing.ThroughmystudyfortheMobileCloudComputing,wearehereexposingtwosimplesolutionstosolvethisproblem.Althoughmyresearchisprettybasic,itstillbenefittheprocessofthedevelopmentformobilecloudcomputingandhowtomakeitenergyefficient.Webelievethatexploringotheralternatives,suchasintrodu

30、cingamiddlewarebasedarchitectureusinganoptimizingoffloadingalgorithm,couldhelpbettertheavailableframeworksandprovidemoreefficientandmoreflexiblesolutionstotheMCCusers.Weknowthatthekindoftechnologywillplayanincreasingimportantroleinourdailylifeinthefuture.Bythisstudy,webetterknowofthenewestdevelopmen

31、tinoursciencearea.Thisworkcanbeextendedtoseveralinterestingdirections:First,full-duplextransmissioncanbeimplementedinthepro-posedsystemtosupportsimultaneousMPTandcomputationoffloadingtoimprovethepowertransferefficiency.Second,thecurrentworkfocusingonasingle-computingtaskcanbegeneralizedtothescenario

32、ofcomputingamulti-taskprogram,whichinvolvesprogrampartitioningandsimultaneouslocalcomputingandoffloading.Last,itisinterestingtoextendthecurrentdesignforsingle-usermobilecloudcomputingsystemtothemultiusersystemthatrequiresjointdesignofradioandcomputationalresourceallocationformobilecloudcomputing.ReferencesBrown,W.(1984).Thehistoryo