![云計算技術及應用課件_第1頁](http://file4.renrendoc.com/view/bfbf5e63ede2b434d46d8bfa2b3a6554/bfbf5e63ede2b434d46d8bfa2b3a65541.gif)
![云計算技術及應用課件_第2頁](http://file4.renrendoc.com/view/bfbf5e63ede2b434d46d8bfa2b3a6554/bfbf5e63ede2b434d46d8bfa2b3a65542.gif)
![云計算技術及應用課件_第3頁](http://file4.renrendoc.com/view/bfbf5e63ede2b434d46d8bfa2b3a6554/bfbf5e63ede2b434d46d8bfa2b3a65543.gif)
![云計算技術及應用課件_第4頁](http://file4.renrendoc.com/view/bfbf5e63ede2b434d46d8bfa2b3a6554/bfbf5e63ede2b434d46d8bfa2b3a65544.gif)
![云計算技術及應用課件_第5頁](http://file4.renrendoc.com/view/bfbf5e63ede2b434d46d8bfa2b3a6554/bfbf5e63ede2b434d46d8bfa2b3a65545.gif)
版權說明:本文檔由用戶提供并上傳,收益歸屬內容提供方,若內容存在侵權,請進行舉報或認領
文檔簡介
云計算技術及應用大連理工大學計算機科學與技術學院2023年春季基本狀況申彥明B810助教:齊恒B812Officehour:Fri3:30-4:30PMCoursewebsite:教材內容Project論文教材內容分布式系統(tǒng)旳概況分布式與集群基本概念分布式數(shù)據(jù)庫分布式文獻系統(tǒng)GFS分布式編程MapReduce算法簡介搜索引擎與PageRank其他有關技術DataCenterBigTableAppEngineGradingHW:40%FinalProject:60%FinalprojectproposalProjectreports12teams,4-5studentsSyllabus(Subjecttochange)Week2Mar8:Lecture1:IntroductionMar10:Lecture2:Map/ReduceTheoryandImplementation,HadoopWeek3Mar15:Lecture3&4:GuestSpeaker(8:00AM-11:35AM研教樓102)Mar17:Lecture5:DistributedFileSystemandtheGoogleFileSystemWeek4Mar22:Lecture6&7:GuestSpeaker(8:00AM-11:35AM研教樓102)Mar24:Lecture8:DistributedGraphAlgorithmsandPageRankWeek5Mar29:Lecture9:IntroductiontoSomeProjectsMar31:Lecture10:DataCentersSyllabus(Subjecttochange)Week6Apr5:Lecture11:SomeGoogleTechnologiesApr7:Lecture12:VirtualizationWeek7Lecture13&14:ProjectPresentationWeek8:NoclassWeek9:Lecture15&16:ProjectPresentationGartnerReportTop10StrategicTechnologyAreas
for2023VirtualizationCloudComputingServers:BeyondBladesWeb-OrientedArchitecturesEnterpriseMashupsSpecializedSystemsSocialSoftwareandSocialNetworkingUnifiedCommunicationsBusinessIntelligenceGreenInformationTechnologyTop10StrategicTechnologyAreasfor2023CloudComputingAdvancedAnalyticsClientComputingITforGreenReshapingtheDataCenterSocialComputingSecurity–ActivityMonitoringFlashMemoryVirtualizationforAvailabilityMobileApplicationsFromDesktop/HPC/Gridsto
InternetCloudsin30YearsHPCmovingfromcentralizedsuperputerstogeographicallydistributeddesktops,clusters,andgridstocloudsoverlast30yearsR/DeffortsonHPC,clusters,Grids,P2P,andvirtualmachineshaslaidthefoundationofcloudputingthathasbeengreatlyadvocatedsince2023Locationofputinginfrastructureinareaswithlowercostsinhardware,software,datasets,space,andpowerrequirements–movingfromdesktopputingtodatacenter-basedcloudsWhatisCloudComputing?1.Web-scaleproblems2.Largedatacenters3.Differentmodelsofputing4.Highly-interactiveWebapplications1.“Web-Scale”Problems
Characteristics:Definitelydata-intensiveMayalsobeprocessingintensiveExamples:Crawling,indexing,searching,miningtheWebDatawarehousesSensornetworks“Post-genomics”lifesciencesresearchOtherscientificdata(physics,astronomy,etc.)Web2.0applications…Howmuchdata?Googleprocesses20PBaday(2023)“allwordseverspokenbyhumanbeings”~5EBCERN’sLHCwillgenerate10-15PBayear640K
oughttobeenoughforanybody.Whattodowithmoredata?AnsweringfactoidquestionsPatternmatchingontheWebWorksamazinglywellLearningrelationsStartwithseedinstancesSearchforpatternsontheWebUsingpatternstofindmoreinstancesHowdoImakemoney?Petabytesofvaluablecustomerdata…SittingidleinexistingdatawarehousesOverflowingoutofexistingdatawarehousesSimplybeingthrownawaySourceofdata:OLTPUserbehaviorlogsCall-centerlogsWebcrawls,publicdatasets…Structureddata(today)vs.unstructureddata(tomorrow)Howcananorganizationderivevaluefromallthisdata?2.LargeDataCentersWeb-scaleproblems?Throwmoremachinesatit!CentralizationofresourcesinlargedatacentersNecessaryingredients:fiber,juice,andlandWhatdoOregon,Iceland,andabandonedmineshaveinmon?ImportantIssues:EfficiencyRedundancyUtilizationSecurityManagementoverhead3.DifferentComputingModelsUtilityputingWhybuymachineswhenyoucanrentcycles?Examples:Amazon’sEC2PlatformasaService(PaaS)GivemeniceAPIandtakecareoftheimplementationExample:GoogleAppEngineSoftwareasaService(SaaS)Justrunitforme!Example:Gmail“Whydoityourselfifyoucanpaysomeonetodoitforyou?”4.WebApplicationsWhatisthenatureoffuturesoftwareapplications?FromthedesktoptothebrowserSaaS==Web-basedapplicationsExamples:GoogleMaps,FacebookHowdowedeliverhighly-interactiveWeb-basedapplications?AJAX(asynchronousJavaScriptandXML)Ahackontopofamistakebuiltonsand,allheldtogetherbyducttapeandchewinggum?SomeCloudDefinitionsIanFosteretaldefinedcloudputingasalarge-scaledistributedputingparadigm,thatisdrivenbyeconomicsofscale,inwhichapoolofabstractedvirtualized,dynamically-scalable,managedputingpower,storage,platforms,andservicesaredeliveredondemandtoexternalcustomersovertheinternet(云計算是一種商業(yè)計算模型。它將計算任務分布在大量計算機構成旳資源池上,使多種應用系統(tǒng)可以根據(jù)需要獲取計算力、存儲空間和多種軟件服務。)IBMexpertsconsidercloudsthatcan:Hostavarietyofdifferentworkloads,includingbatch-stylebackendinteractive,user-facingapplicationsAllowworkloadstobedeployedandscaled-outquicklythroughtherapidprovisioningofvirtualmachinesorphysicalmachinesSupportredundant,self-recovering,highlyscalableprogrammingmodelsthatallowworkloadstorecoverfromHW/SWfailuresMonitorresourceuseinrealtimetorebalanceallocationsondemandInternetCloudGoals
Sharingofpeak-loadcapacityamongalargepoolofusers,improvingoverallresourceutilizationSeparationofinfrastructuremaintenancedutiesfromdomain-specificapplicationdevelopmentMajorcloudapplicationsincludeupgradedwebservices,distributeddatastorage,rawsuperputing,andaccesstospecializedGrid,P2P,data-mining,andcontentnetworkingservicesThreeAspectsinHardwarethat
areNewinCloudComputingTheillusionofinfiniteputingresourcesavailableondemand,therebyeliminatingtheneedforclouduserstoplanfaraheadforprovisioningTheeliminationofanup-frontmitmentbycloudusers,therebyallowingpaniestostartsmallandincreasehardwareresourceswhenneededTheabilitytopayputingresourcesonashort-termbasisasneeded(e.g.,processorsbythehourandstoragebytheday)andreleasethemafterdoneandtherebyrewardingresourceconservationSomeInnovativeCloudServices
andApplicationOpportunitiesSmartandpervasivecloudapplicationsforindividuals,homes,munities,panies,andgovernments,etc.CoordinatedCalendar,Itinerary,jobmanagement,events,andconsumerrecordmanagement(CRM)servicesCoordinatedwordprocessing,on-linepresentations,web-baseddesktops,sharingon-linedocuments,datasets,photos,video,anddatabases,etcDeployconventionalcluster,grid,P2P,socialnetworkingapplicationsincloudenvironments,morecost-effectivelyEarthboundApplicationsthatDemandElasticityandParallelismratherdatamovementCostsOperationsinCloudComputingUsersinteractwiththecloudtorequestserviceProvisioningtoolcarvesoutthesystemsfromthecloud–configurationorreconfiguration,ordeprovisionTheserverscanbeeitherrealorvirtualmachinesSupportingresourcesincludedistributedstoragesystem,datacenters,securitydevices,etc.CloudComputingInstancesGoogleAmazonMicrosoftAzureIBMBlueCloudGoogleCloudInfrastructureSchedulerChubbyGFSmasterNodeNodeNode…UserApplicationSchedulerslaveGFSchunkserverLinuxNodeMapReduceJobBigTableServerGoogleCloudInfrastructureS3EBSEC2EBSEC2EBSEC2EBSEC2SimpleDBSQSUserDeveloperAmazonElasticComputingCloudSQS:SimpleQueueServiceEC2:RunningInstanceofVirtualMachinesEBS:ElasticBlockService,ProvidingtheBlockInterface,StoringVirtualMachineImagesS3:SimpleStorageService,SOAP,ObjectInterfaceSimpleDB:SimplifiedDatabase
Azure?ServicesPlatformMicrosoftAzurePlatformDeveloperMonitoringApplicationServerProvisioningManagerUserOpenSourceLinuxwithXenTivoliMonitoringAgent……IBMBlueCloudCostConsiderations:Power,Cooling,
PhysicalPlant,andOperationalCostsCosttechnologycostscostofsecurityetc.Benefitsavailabilityopportunityconsolidationetc.CostBreakdown+Storage($/MByte/year)+Computing($/CPUCycles)+Networking($/bit)ResearchChallengesServiceavailabilityS3outage:authenticationserviceoverloadleadingtounavailabilityAppEnginepartialoutageprogrammingerrorGmail:siteunavailableSolutions:ThemanagementofaCloudComputingservicebyasinglepanyresultsinasinglepointoffailure(SPF).IntheInternet,alargeISPusesmultiplenetworkproviderssothatfailurebyasinglepanywillnottakethemofftheair.Similarly,weneedmultipleCloudComputingproviderstosupporteachothertoeliminateSPF.ResearchChallengesDataSecurityCurrentcloudofferingsareessentiallypublicratherthanprivatenetworks,exposingthesystemtomoreattackssuchasDDoSattacks.Solutions:Therearemanywellunderstoodtechnologiessuchasencryptedstorage,virtuallocalareanetworks,andnetworkmiddleboxes.ResearchChallengesDataTransferBottlenecksApplicationscontinuetobeemoredata-intensive.Ifweassumeapplicationsmaybe“pulledapart”acrosstheboundariesofclouds,thismayplicatedataplacementandtransport.BothWANbandwidthandintra-cloudnetworkingtechnologyareperformancebottleneck.Industrialsolutions:Itisestimatedthat2/3ofthecostofWANbandwidthisconsumedbyhigh-endrouters,whereasonly1/3chargedbyfiberindustry.Wecanlowerthecostbyusingsimplerroutersbuiltfrommodityponentswithcentralizedcontrol,butresearchisheadingtowardsusinghigh-enddistributedrouters.ResearchChallengesSoftwareLicensingCurrentsoftwarelicensesmonlyrestricttheputersonwhichthesoftwarecanrun.Userspayforthesoftwareandthenpayanannualmaintenancefee.ManycloudputingprovidersoriginallyreliedonopensourcesoftwareinpartbecausethelicensingmodelformercialsoftwareisnotagoodmatchtoUtilityComputing.Someideas:WecanencouragesalesforcesofsoftwarepaniestosellproductsintoCloudComputing.Ortheycanimplementpay-per-usemodeltothesoftwaretoadapttoacloudenvironment.ResearchChallengesScalablestorageDifferencesbetweenmonstorageandcloudstorageThesystemisbuiltfrommanyinexpensivemodityponentsthatoftenfailThesystemstoresamodestnumberoflargefilesTheworkloadsprimarilyconsistbothlargestreamingreadsandsmallrandomreads.Theworkloadsmanylarge,sequentialwritesthatappenddatatofilesandoncewritten,filesareseldommodifiedagain.Thecloudstorage(file)systemneedstosharemanyofthesamegoalsaspreviousdistributedfilesystemssuchasperformance,scalability,reliability,andavailability.Inaddition,itsdesignneedstobedrivenbykeyobservationsofthespecificworkloadsandtechnologicalenvironment,bothcurrentandanticipated,thatreflectamarkeddeparturefromsomeearlierfilesystemdesignassumptions.GFSFilesaredividedintofixed-sizechunks,Chunksizeisoneofthekeydesignparameters.GFSchooses64MB,whichismuchlargerthantypicalfilesystemblocksizes.Themasterstoresthreemajortypesofmetadata:thefileandchunknamespaces,themappingfromfilestochunks,andthelocationsofeachchunk’sreplicas.GFSsupportstheusualoperationstocreate,delete,open,close,read,andwritefiles.ResearchChallengesTransparentProgrammingModelProgramswrittenforcloudimplementationneedtobeautomaticallyparallelizedandexecutedonalargeclusterofmoditymachines.Therun-timesystemshouldtakecareofthedetailsofpartitioningtheinputdata,schedulingtheprogram'sexecutionacrossasetofmachines,handlingmachinefailures,andmanagingtherequiredinter-machinemunication.Theprogrammingmodelshouldallowprogrammerswithoutmanyexperienceswithparallelanddistributedsystemstoeasilyutilizetheresourcesofalargedistributedsystem.MapReduceScalableDataProcessingonLargeClustersAwebprogrammingmodelimplementedforfastprocessinga
溫馨提示
- 1. 本站所有資源如無特殊說明,都需要本地電腦安裝OFFICE2007和PDF閱讀器。圖紙軟件為CAD,CAXA,PROE,UG,SolidWorks等.壓縮文件請下載最新的WinRAR軟件解壓。
- 2. 本站的文檔不包含任何第三方提供的附件圖紙等,如果需要附件,請聯(lián)系上傳者。文件的所有權益歸上傳用戶所有。
- 3. 本站RAR壓縮包中若帶圖紙,網(wǎng)頁內容里面會有圖紙預覽,若沒有圖紙預覽就沒有圖紙。
- 4. 未經(jīng)權益所有人同意不得將文件中的內容挪作商業(yè)或盈利用途。
- 5. 人人文庫網(wǎng)僅提供信息存儲空間,僅對用戶上傳內容的表現(xiàn)方式做保護處理,對用戶上傳分享的文檔內容本身不做任何修改或編輯,并不能對任何下載內容負責。
- 6. 下載文件中如有侵權或不適當內容,請與我們聯(lián)系,我們立即糾正。
- 7. 本站不保證下載資源的準確性、安全性和完整性, 同時也不承擔用戶因使用這些下載資源對自己和他人造成任何形式的傷害或損失。
最新文檔
- 物理科技在智能交通系統(tǒng)中的應用
- 現(xiàn)代藝術與設計趨勢創(chuàng)新與變革
- 現(xiàn)代營銷中的用戶體驗設計
- 環(huán)境科學與未來綠色發(fā)展的結合策略
- 國慶節(jié)紅色電影活動方案
- Unit7《Lesson 26 I Love My Family》(說課稿)-2024-2025學年北京版(2024)英語三年級上冊
- 2024-2025學年高中地理 第4章 旅游與區(qū)域的發(fā)展 章末分層突破說課稿 中圖版選修3
- Unit 7 Happy Birthday!(說課稿)-2024-2025學年譯林版(三起)(2024)英語三年級上冊
- 2024年屆九年級歷史上冊 第11課 開辟新時代的“宣言”說課稿2 北師大版001
- 《18 初始機器人》說課稿-2023-2024學年清華版(2012)信息技術一年級下冊
- 醫(yī)院消防安全培訓課件
- 質保管理制度
- 《00541語言學概論》自考復習題庫(含答案)
- 2025年機關工會個人工作計劃
- 人事測評理論與方法-課件
- 最新卷宗的整理、裝訂(全)課件
- 城市旅行珠海景色介紹珠海旅游攻略PPT圖文課件
- 小學 三年級 科學《觀測風》教學設計
- JJF1664-2017溫度顯示儀校準規(guī)范-(高清現(xiàn)行)
- 第二講共振理論、有機酸堿理論
- 高考英語聽力必備場景詞匯精選(必看)
評論
0/150
提交評論