资源描述：
《Hengtian Newsletter Volume 8_Data as a Service (DaaS).pdf》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。
1、April,2015DataasaService(DaaS)HengtianFOREWORDNewsletterBigdataanalyticshavejumpedtothetopoftheagendaofmanycorporatestrategies.RecenttechnologydevelopmentslikeHadoopandMapReducehaveimbuedtheparallelprocessingindataanalyticswithanunlimitedpowerofhorizontalexpan-s
2、ion.Infact,companiesthatjumpontothebigdatabandwagonearlierdon’tnecessarilydrawvaluefromit.Whilemostcompa-nieshaveawell-establishedenterprisedatamodel,manywillneedtore-thinkordevelopanewdatastrategyinordertokeepupwiththisnewtechnologicalinnovation.Developingadata
3、strategymeanscompaniesneedtoclearlyidentifythebusinessneeds,choosetherightdatasetsandbuildagovernancemodeltocollect,validateandminethedatatocreatesmartvalueforitsconsumers.Thiscanbeadauntingtaskformanyorganizations.Companieswhowillbesuccessfulinthefuturewillneed
4、toknowhowtoembracebigdataanalytics.ITtransformationisnotsimplyaboutadoptingrelevanttechnologies,buthavingtheabilitytoenableself-servicing,accessible,realtime/on-lineana-lyticstolargecomplexdatasets.HowtoestablishaDataasaService(DaaS)platformwillbethekeytothesucc
5、essofmostcorporatedatapolicies.AlbertMaChiefInnovationOfficerofHengtian1TheMythsofEnterpriseDataStrategyMyth1-EnterprisedatamodelisonlywhatweneedToday,mostcompanieshaveastandardenterprisedatamodelanddatagovernancestructure.Atradi-tionaldatawarehouse/datamartisus
6、edtorunheavydutyETL(extraction,transformation,load)jobseveryday.Thebasicmodelwasbuiltmanyyearsago,aroundthetimethatbusinessunitsspentatre-mendousamountoftimetocollecttherequirements.Withtime,somehavebecomeobsolete,requir-ingcontinuousefforttobeputintothemaintena
7、nceofthosecentraldatarepositories.Whatcoststhecompaniesmorearetheadditionaldatawarehouses/datamartsneededtobebuiltfornewdatasources,whichmuddytheoriginaldatamodelswithcomplexjointrelations.Mostcompaniesshouldtakeastepbackandassumeaholisticviewoftheirdatastrategy
8、,identifyingwhatdatathebusinessneedsanddesigningabettermethodologytocollect,maintainandusethem.Myth2-Itisimpossibleforbusinesstoidentifyalldatarequire-mentsThestateme
