计算机专业文献翻译面向数字图书馆的海量信息管理体系结构研究Word格式.docx

计算机专业文献翻译面向数字图书馆的海量信息管理体系结构研究Word格式.docx

《计算机专业文献翻译面向数字图书馆的海量信息管理体系结构研究Word格式.docx》由会员分享，可在线阅读，更多相关《计算机专业文献翻译面向数字图书馆的海量信息管理体系结构研究Word格式.docx（14页珍藏版）》请在冰豆网上搜索。

（DepartmentofComputerScienceandTechnology,TsinghuaUniversity,Beijing100084,China）

XINGChun-Xiao+,ZENGChun,LIChao,ZHOULi-Zhu

Abstract

Thispaperinvestigatesthechallengingissuesandtechnologiesinmanagingverylargedigitalcontentsandcollections,andgivesanoverviewoftheworksandenablingtechnologiesintherelatedareas.Basedontheanalysisandcomparisonoftherelatedwork,anovelarchitectureofmassiveinformationmanagementfordigitallibraryisdesigned.Thekeycomponentsandcoreservicesaredescribedindetail.Finally,acasestudyTHADL（TsinghuaUniversityarchitecturedigitallibrary）thatcomplieswiththearchitecturalframeworkispresented.

Keywords:

digitallibrary;

architecture;

massiveinformationmanagement;

interoperability;

metadata

1Introducn

Intherecordedhitiostoryofhumanbeing,theprintedmaterialsusedtoplayadominantroleinthepreservationandpervasionofhumaninformationandknowledge.However,withtherapiddevelopmentoftechnologiesincomputer,communication,multimediaandstorage,thisroleisgivingawaytothedigitalresourcesinthenewera.Theexplosivegrowthofinformationindigitalformshasposedchallengesnotonlytotraditionalarchivesandtheirinformationproviders,butalsotoorganizationsinthegovernment,commercialandnon-profitsectors.AccordingtothelatestreportbyLymanandVarian,theworld’stotalyearlyproductionofprint,film,optical,andmagneticcontentwouldrequireroughly1.5billiongigabytesofstoragewhichisroughly250megabytesforeverypersonontheearth.Printeddocumentsofallkindscompriseonly0.003%ofthetotal.Magneticstorageisbyfarthelargestmediumforstoringinformationandisthemostrapidlygrowingsection,withashippedharddrivecapacitydoublingeveryyear.Thetypesofdigitalresourcesarediverse.Theyincludedigitaltexts,documents,scientificdata,images,animation,video,audioetc.Theapplicationsofthedigitalresourcesarequitebroad,includingDL（digitallibrary）,movie/videocenter,otherpublicmedia（television,broadcast,newspaper,etc.）,museum,andnationalorcooperativeinformationcenter.Atthesametimetheinformationhighway,whichisrepresentedbyInternet,hasbeenanimportanttoolofthepervasionofdigitalresources.Thegovernments,companies,groups,researchinstitutes,non-governmentorganizations,educationinstitutesallovertheworldputmassiveinformationontheWeb.

Technologychallengesandkeyissues

Thesemassivedigitalresourcespresentmanychallengingissuesindatamanagementtechnologyarea.Thefollowingaresomeexamples.

（1）Datamodel.

Traditionaldatamodeltheoriesareonlyapplicabletostructureddata,butnotforthemassivedigitalresourcesofvarioustypesandtheyaremostlysemi-structuredorunstructured.Thus,newdatamodelsaredemanded.

（2）Systemarchitecture.

Traditionaldatabasemanagementsystemsaredesignedforbusinessdataprocessingfeaturedbyconcurrent,short,andupdatetransactions.Thereforetransactionmanagementandconcurrentcontrolremainsasthecenterofsystemarchitecture.Thearchitectureisnotsuitableforthemanagementofdigitalresourcesasclassicaltransactionconceptisbecominglessimportantintheseresources.Weneedtopursuenovelanduniversalframeworksformassivedigitalresourcesmanagement.

（3）Massiveinformationstorage.

Thevolumeofdigitaldataresourcesiscountedbyterabytesorpetabytes.TraditionalstoragedevicesusingSCSIcannotworkforefficientstorage,onlinemigrationandpersistentarchiveofsuchmassivedigitalresources.Sotheresearchofmulti-levelstoragesystems,SAN（StorageAreaNetworks）andothertechnologyareinevitable.

（4）Queryprocessing.

Intraditionaldatabasesystems,queriesareexpressedinquerylanguagesuchasSQL,butinthequeryandsearchofmassivedigitalresources,manynewmechanismsshouldbeused,suchaskeywordsearch,full-textsearch,similarityquery,andcontent-basedmultimediaretrieval.Howtointegratethequerymethods（includingSQL,OQL,anddifferentXMLquerylanguages,e.g.,XQL,XML-QL,XML-GL）efficientlytobuildanefficientandflexiblequeryprocessingmethodhasnotbeensatisfactorilysolvedyet.

Tosolvetheproblemsmentionedabovewillremainasamajorgoaltoresearchersinthenextfewyears.Tofulfillthisend,wepresentanovelarchitectureformassiveinformationmanagementofdigitalresourcesinthispaper.Thisarchitectureisintendedtomeettherequirementsofmanagingdigitalresourcescharacterizedbydistributed,dynamic,massiveandheterogeneousproperties.

2OverviewoftheRelatedWork

TheIEEESTD610.12[2]definesarchitectureasthestructureofcomponents,theirrelat