DisasterPlanningandTrustworthyDigitalRepositories

RebeccaD.FrankApril20,2012

Frank

2

1.Introduction........................................................................................................................4

2.LiteratureReview...............................................................................................................52.1DigitalPreservation....................................................................................................................5

2.1.1LOCKSS.......................................................................................................................................62.1.2iRODS.........................................................................................................................................7

2.2DigitalCuration..........................................................................................................................72.3Trust..........................................................................................................................................8

2.3.1TRAC...........................................................................................................................................82.3.2DRAMBORA................................................................................................................................92.3.3DataSealofApproval...............................................................................................................10

2.4ThreatstoDigitalCollections....................................................................................................112.5PlanningforDisasters...............................................................................................................12

2.5.1DisasterResponseandRecovery.............................................................................................132.5.2RiskManagement....................................................................................................................142.5.3BusinessContinuityPlanning...................................................................................................14

3Methodology.....................................................................................................................153.1SelectionofSites......................................................................................................................153.2DiscussionofEightSites...........................................................................................................16

3.2.1Chronopolis..............................................................................................................................163.2.2HathiTrust................................................................................................................................173.2.3Inter-UniversityConsortiumforPoliticalandSocialResearch(ICPSR)....................................173.2.4MATRIX.....................................................................................................................................183.2.5NationalLibraryofAustralia....................................................................................................183.2.6Portico......................................................................................................................................183.2.7TheInternetArchive................................................................................................................193.2.8TheMetaArchiveCooperative.................................................................................................19

3.3DocumentAnalysis...................................................................................................................203.4Interviews................................................................................................................................223.5InterviewAnalysis....................................................................................................................23

4.Findings............................................................................................................................254.1IncentiveforCreation...............................................................................................................254.2Documentation........................................................................................................................284.3ProcessofCreation..................................................................................................................314.4Obstacles.................................................................................................................................334.5TestingthePlans......................................................................................................................354.6AccesstoDisasterPlanDocumentation....................................................................................36

5.Discussion.........................................................................................................................38

6.Conclusion........................................................................................................................40

7.Acknowledgements..........................................................................................................41

References............................................................................................................................42

AppendixA:ConsenttoParticipateinaResearchStudyInterview.......................................48

AppendixB:QuestionsforSemi-StructuredInterview..........................................................50

Frank

3

Table1:InitialListofRepositories................................................................................................16Table2:AvailableDisasterPlanningDocumentation...................................................................20Table3:ParticipantsInterviewed.................................................................................................22Table4:DescriptionofDocumentCodingScheme......................................................................24

Figure1:CertificationandDisasterPlanningDocumentation.....................................................27Figure2:TRACAuditResults........................................................................................................28Figure3:DisasterPlanningDocumentation.................................................................................29Figure4:Obstacles.......................................................................................................................33Figure5:ObstaclesandDocumentation......................................................................................35Figure6:Access............................................................................................................................38

Frank

4

1.IntroductionDisasterresponseandrecoveryplanningremainsoneofthemostimportantcomponentsofapreservationprogramindigitalrepositories,andalsooneoftheleastunderstood.TheadoptionofstandardsandmodelsforpreservationsuchastheAuditandCertificationofTrustworthyDigitalRepositoriesandtheOpenArchivalInformationSystem(OAIS)modelhavehelpedtoclarifyandilluminatebestpracticesinthedigitalpreservationcommunity.However,ourunderstandingofdisasterplanningfordigitalrepositoriesremainslimited.InanarticlewrittenforWiredMagazine,ChrisAndersonarguedthatwearecurrentlyinthe“PetabyteAge”(Anderson,2008).Thisageismarkedbyanexponentialincreaseindigitaldata.Thisproliferationincludesscholarlyresearchdataaswellasdigitalinformationcreatedforentertainmentandpersonaluse,“thedigitaluniverse—informationthatiseithercreated,captured,orreplicatedindigitalform—was281exabytesin2007.In2011,theamountofdigitalinformationproducedintheyearshouldequalnearly1,800exabytes,or10timesthatproducedin2006”(Gantzetal.,2011,3).Intermsofstorage,“2007markedthe‘crossover’yearinwhichmoredigitaldatawascreatedthanthereisdatastoragetohostit”(Berman,2008,52).Thistippingpoint,thepointatwhichdatacreatedoutpacedourcapacitytostoredata,issignificantforthedigitalpreservationcommunity.Itisatthispointwhendecisionmakingfordigitalpreservationmustfocusnotonlyonhowtopreservedata,butalsoonwhattopreserve.Thesedecisionsarebasedonanynumberofcriteria,buttheimportantfactortoconsiderfordigitalpreservationanddisasterplanningisthattheinformationselectedforpreservationindigitalrepositorieshasultimatelybeenselectedbecauseofitsvalue.“Whilethecostsofmaintainingdigitalpreservationcapacityarenotinsignificant,thecostsofthealternativeareoftengreater.Re-creatingresearchdatasetscanbeprohibitivelyexpensive;intheextreme,itmaybeimpossibletore-createlostdata”(Beagrie,Chruszcz,&Lavoie,2008,16).Theimportanceanduniquenessofdatasuchasthis,compoundedwiththedifficultyorimpossibilityofrecreatinglostdata,makesastrongcaseforpreservation.Becauseofthisneedtopreservethedatathatisheldindigitalrepositories,disasterplanningisaparticularlyimportantactivity.Thedigitalpreservationcommunityisdevelopinganawarenessandunderstandingoftheconceptofdisasterplanningaspartofadigitalpreservationprogram,butathoroughunderstandingofdisasterplanninginpracticehasnotyetbeenachieved.Thegoalofthisstudyistounderstandifdigitalrepositoriesthathaveapreservationmandateareengagingindisasterplanningactivities,particularlytofurthertheirpursuitoftrusteddigitalrepositorystatus.Incaseswheredigitalrepositoriesareengagingwithdisasterplanning,thestudyalsoexaminestheprocessofcreatingdisasterresponseandrecoveryplans,withafocusonhowtheseactivitiesareintegratedintothemanagementofthedigitalrepositories.

Frank

5

Thisstudyfocusesonthepracticesofdigitalrepositoriesthathaveeithersoughttrustedrepositorystatus,haveundergonesometypeofself-audit,orhaveexpressedacommitmenttopursuingthistypeofcertificationprocessinthefuture.Astheliteratureindicates,disasterplanningisgenerallyunderstoodtobepartoftherequirementsfortrustedrepositorystatus,butthedetailsofsuchplanningactivitiesarenotwelldocumentedorunderstood.

2.LiteratureReview

2.1DigitalPreservationInordertounderstanddisasterplanningfordigitalrepositories,itisimportanttofirstexaminedigitalpreservationandtherelationshipofpreservationtodisasterplanning.Disasterplanningfordigitalrepositorieshasthesameintellectualrootsasdigitalpreservation,“digitalpreservationcanencompassarangeofactivities,fromsimplereplicationandstoragetomorecomplextransformation,dependingontheassessedvalueandrisktothetargetcontent”(Hitchcock,Brody,Hey,&Carr,2007,1).FrancineBermanstatesthatpreservationactionsare,“actionsundertakentoensurethelong-termviabilityandavailabilityoftheauthoritativenatureofdigitalmaterial.Preservationactionsshouldensurethematerialremainsauthentic,reliable,andusablewhileitsintegrityismaintained;suchactionsincludevalidation,assigningpreservationmetadata,assigningrepresentationinformation,andensuringacceptabledatastructuresandfileformats”(Berman,2008,55).Inshort,digitalpreservationconsistsofthoseactionsthatensuretheviabilityandauthenticityofdigitalobjectsovertimeanddisasterplanningisoneofthoseactions.Disasterplanningorpreparednessinatraditionalsense“referstoastateorsituationofthelibrariesinwhichtheyarewellpreparedtopreventseverelibrarydamagefrompotentialdisasters”(Wong&Green,2006,72).Andmorespecifically,adisasterplanisadocumentthatdescribespoliciesandprocedureswhichhavebeencreatedtoprevent,preparefor,respondto,andrecoverfromadisaster(Muir&Shenton,2002).Analogoustoanalogcollections,disasterplanningisanessentialactivityfordigitalrepositories(Patkus&Motylewski,1993).Theconceptsunderlyingdisasterplanningforanalogmaterialscanbeappliedtodigitalrepositoriesinthatthelong-termpreservationofdigitalmaterialsdependsontheabilityofanorganizationtoprevent,preparefor,respondto,andrecoverfromdisasterevents.In2007,theCenterforResearchLibraries,TheDigitalCurationCenter,DigitalPreservationEurope,andNESTORmetandidentifiedalistoftencharacteristicsofdigitalpreservationrepositories(CenterforResearchLibraries[CRL],2007).Thislist“providesastructurethatinformstheprocessesandoutcomes”ofrepositoryauditandcertificationprocessessuchastheTrustedRepositoryAuditandCertification(TRAC),whichwillbediscussedingreaterdetailbelow(McHugh,2008,133).Thecharacteristicsare:

1. Therepositorycommitstocontinuingmaintenanceofdigitalobjectsforidentifiedcommunity/communities.

Frank

6

2. Demonstratesorganizationalfitness(includingfinancial,staffing,andprocesses)tofulfillitscommitment.

3. Acquiresandmaintainsrequisitecontractualandlegalrightsandfulfillsresponsibilities.4. Hasaneffectiveandefficientpolicyframework.5. Acquiresandingestsdigitalobjectsbaseduponstatedcriteriathatcorrespondtoits

commitmentsandcapabilities.6. Maintains/ensurestheintegrity,authenticityandusabilityofdigitalobjectsitholdsover

time.7. Createsandmaintainsrequisitemetadataaboutactionstakenondigitalobjectsduring

preservationaswellasabouttherelevantproduction,accesssupport,andusageprocesscontextsbeforepreservation.

8. Fulfillsrequisitedisseminationrequirements.9. Hasastrategicprogramforpreservationplanningandaction.10. Hastechnicalinfrastructureadequatetocontinuingmaintenanceandsecurityofits

digitalobjects.Thesecriteriarelatebothdirectlyandindirectlytodisasterplanning.Specifically,criteriaregardingmaintenance(1,3,6,and7)assumethattherepositorywillbeabletomaintaindigitalobjectsandtheirmetadataovertime,presumablyinspiteofanydisastersthatmayoccur.Criteriaregardingpreservationandsecurity(9and10)arealsosignificantfordisasterplanninginthatdisasterplanningeffortsaremeanttoensurelongtermpreservationandsecurityofdigitalobjects.Onekeyproblemfacingthefieldofdigitalpreservationisthesheervolumeofdata.Whilethismaynotbeaproblemattheindividualrepositorylevel,aseachrepositoryisabletoacceptfordepositonlythatdatawhichmeettheirspecifiedcriteria,itisaproblemforthecommunityasawhole,“thescaleofdigitalcreationisfaroutpacingthecapacitytostorethedata”(Bermanetal.,2010,9).Thisproblemofscalehasbeenwidelydocumented(e.g.Berman,2008;Hey,2003).Anditisfromthisproblemthatothersarise.Specifically,problemsconcerninghowtoensurethelong-termviabilityofsustainabledigitalrepositorieswhilecontinuingtogrow.Thisproblemalsoleadstodifferentproposalsfordisastermitigationsolutions,twoofwhicharedescribedbelow.

2.1.1LOCKSSSomeapproachestodigitalpreservationhaveimplicitdisasterplanningstrategiesbuiltin.OnesuchapproachtodigitalpreservationistheLOCKSS(LotsofCopiesKeepsStuffSafe)system.LOCKSSismodeledonthesystemusedbylibrariestopreservephysicalcontentthroughduplicationofresourcesacrossmultipledistributedorganizations,“thephrase‘distributeddigitalpreservationfederations’isbeingusedincreasinglytodescribecooperativesofgeographically-dispersedinstitutionswhoarebandingtogethertoformsolutionstothedigitalpreservationproblem”(McDonald&Walters,2010,1).Fordigitalpreservation,“acombinationofmassivereplication,ratelimitation,inherentintrusiondetectionandcostlyoperationscanproduceapeer-to-peersystemwithremarkableabilitytoresistattacksbysomeextraordinarilypowerfuladversariesoverdecades.Itslackofdependenceonlong-termsecretsandstable

Frank

7

identitiesblocksmanyofthepathsbywhichsystemsaretypicallyattacked”(Maniatisetal.,2005,42).Thisparticularsystemofpreservationisreliablebecauseitallowsmultiplerepositoriestoshareresponsibilityfordigitalobjects.WhileeachpartnerinaLOCKSSsystemisindeedresponsibleformaintainingtheircopyoftheitems,theyareabletorestoreanyand/oralloftheitemsintheirrepositoryfromanotherpartnerintheeventofdataloss.Whileitisoftennotdirectlystated,the‘lotsofcopies’partofaLOCKSSsystemis,ineffect,meanttopreservethedatathatmaysufferadisasteratonelocationbyprovidingduplicatesacrossseverallocations.ArticlessuchasthosebyManiatisetal.(2005)highlightthestrengthofaLOCKSSnetworktoresist“attack”and“randomstoragefaults,”bothofwhichcanbeconsidereddisasterevents(30).Inastudypublishedin2007,SchroederandGibson(2007)conductedasurveyof“field-gathereddiskreplacementdatafromanumberoflargeproductionsystems,includinghigh-performancecomputingsitesandinternetservicessites.About100,000disksarecoveredbythisdata,someforanentirelifetimeoffiveyears”(1).Thestudyfoundthat“inthefield,annualdiskreplacementratestypicallyexceed1%,with2-4%commonandupto13%observedonsomesystems,”afailureratethatwassignificantlyhigherthantheauthorsexpected(Schroeder&Gibson,2007,1).ThissuggeststhattherandomstoragefaultsdiscussedbyManiatisetal.areindeedlikelytooccur.However,thesearticlesdonotfocusspecificallyondisasterresponseandrecoveryplanning,rather,implyingthattheduplicationforlong-termpreservationwillallowthesystemtoovercomeanytypeofdisruptionorlossinservice.

2.1.2iRODSAnotherapproachtodigitalpreservationistheIntegratedRuleOrientedDataSystems(iRODS)softwarethathasbeendevelopedbytheDataIntensiveCyberEnvironmentsgroup(DICE).iRODSis“asecondgenerationdatagridsystemthatfacilitatesdatamanagementspanninglargegeographicareasandacrossadministrativedomains”(DataIntensiveCyberEnvironmentsGroup[DICE],2008,1).AsdescribedbyMoore,“theiRODSdatagridisagenericdatamanagementinfrastructurethatcanbetunedtosupportdatapreservation,datapublication,datasharing,ordataanalysisthroughspecificationofappropriatedatamanagementpolicies”(Moore,2008,73).TheiRODSsystemisnotspecificallyapreservationsystem,butitcanbeusedtofacilitateandsupportpreservationbymitigatingagainstrisk.“Whenauserororganizationstoresdatawithassociatedmetadatainadatagrid,theyapplypoliciestoensurethattheresultingcollectionwillmeettheirgoals.Suchpoliciesincludedisasterrecovery(syntacticreplication),[and]persistentpreservationforthelongterm(temporalreplication)”(DICE,2008,2).Inotherwords,iRODSisasystemthatallowsrepositoriestocreateandenforcerulesandpolicies,thereforeensuringconsistencywithintherepository(Rajasekar,2010).Theserulesandpoliciesincludethoserelatingtolong-termpreservation.

2.2DigitalCurationDigitalcurationisavalueproposition.AccordingtoWaltersandSkinner,“digitalcurationreferstotheactionspeopletaketomaintainandaddvaluetodigitalinformationoveritslifecycle,includingtheprocessesusedwhencreatingdigitalcontent”(Walters&Skinner,2011,5).Similarly,MaureenPennockattheDigitalCurationCentredescribesdigitalcurationinthe

Frank

8

followingway,“digitalcuration,broadlyinterpreted,isaboutmaintainingandaddingvaluetoatrustedbodyofdigitalinformationforbothcurrentandfutureuse:inotherwords,itistheactivemanagementandappraisalofdigitalinformationoveritsentirelifecycle”(Pennock,2007,1).InPennock’sview,curationisdifferentfrompreservationinthatpreservationhasamorenarrowfocusonmaintainingcontinuedaccesstodigitalmaterialsoveralongspanoftime.Proponentsofdigitalcurationarguethatdigitalcurationcantakeplacewithcollectionsthatarethesubjectofdigitalpreservationefforts,andthatitcanalsotakeplacewithcollectionsthatarenotmeantforlong-termpreservation.Othersinthefieldofdigitalpreservationdonotnecessarilyagree.Giventhesedifferences,thequestionofwhetherdisasterplanningfallsintothecategoryofdigitalcurationarises.Whiledisasterplanningisnotfeaturedprominentlyinthedigitalcurationliterature,itcouldbearguedthatdisasterplanningisindeedanimportantactivityfordigitalcuration.Whileanitemorcollectionisneeded,therepositorymustbepreparedforanydisasterthatthreatensthevalueoftheinformation.Despitethefactthatdigitalcurationdoesnotrequirelong-termpreservation,theprocessoffulllifecyclemanagementofdigitalresourcesmeansthatthoseresourcesmustbemanagedandpreservedforaslongastheyareneeded.Susceptibilitytodisastersisaproblemnotonlyifitinterruptsaccesstocollectionsbutalsoifitthreatenstheintegrityofthosecollections,whethertheyareneededforoneyearortwenty.

2.3TrustAnotherimportantelementofpreservationanddisasterpreparednessfordigitalrepositoriesexistsattherepositorylevel,andthatistheconceptoftrust.GarrettandWatersmaketheclaimthat,“forassuringthelongevityofinformation,perhapsthemostimportantroleintheoperationofadigitalarchivesismanagingtheidentity,integrityandqualityofthearchivesitselfasatrustedsourceoftheculturalrecord.Usersofarchivedinformationinelectronicformandofarchivalservicesrelatingtothatinformationneedtohaveassurancethatadigitalarchivesiswhatitsaysthatitisandthattheinformationstoredthereissafeforthelongterm”(Garrett&Waters,1996,23).Theimplicationhereisthatifarepositoryisnottrustedbyusers,thenthedatastoredinthatrepositoryisnotpreserved.Usersmustbeabletotrustthatthedatacontainedwithinadigitalrepositoryiswhatitpurportstobe,andoneofthewaysthatusersjudgeintegrityofdigitalobjectsisthroughtrustintherepository.Trustisalsoanimportantcomponentofdisasterplanninginthatonewayinwhichrepositoriesgaintrustisthroughdemonstrationofpreparedness.Repositoriesdemonstratetheirabilitytopreservetheircontentthroughdisastersbymakingdisasterplanningdocumentationavailabletothecommunity,byconductingself-auditsofbestpracticesandmakingtheresultsavailabletothecommunity,orbyundertakingaprocessofauditandcertificationasadministeredbyanexternalorganization.

2.3.1TRACTheconceptoftrusthasemergedasacommunitystandardfordigitalrepositories;specifically,theassignmentofTrustedRepositorystatusthroughcertification.Threeexamplesofwhichare

Frank

9

theDataSealofApproval(http://datasealofapproval.org/)whichoriginatedintheNetherlands,DRAMBORA(http://www.repositoryaudit.eu/)whichwasdevelopedjointlybytheDigitalCurationCentreandDigitalPreservationEurope,andTrustedRepositories:AuditandCertification(TRAC)(http://www.crl.edu/archiving-preservation/digital-archives/metrics-assessing-and-certifying-0)whichisadministeredbytheCenterforResearchLibraries(CRL)intheUnitedStates.TRACcertificationisbasedontheTrustworthyRepositoriesAudit&Certification:CriteriaandChecklist(ISO16363,2012).Eachofthesecertificationsrequirethattherepositoryseekingcertificationundergoanauditprocess,althoughtheprocessofTRACcertificationismorerigorousandtimeconsumingthanDataSealofApprovalcertification,andDRAMBORAwasdesignedtobeaself-auditprocess(e.g.McHugh,2008;Sesink,2010).DisasterplanningisacoreconstructoftheTRACauditandcertificationrequirements.Sections5.1and5.2aremostexplicit:

“5.2.4Therepositoryshallhavesuitablewrittendisasterpreparednessandrecoveryplan(s),includingatleastoneoff-sitebackupofallpreservedinformationtogetherwithanoffsitecopyoftherecoveryplan(s)”(ISO16363,2012,78).

DisasterplanningisanexplicitelementoftheTRACcertificationprocess,andisanimplied(butnotdirectlystated)elementoftheDataSealofApprovalcertificationprocess.WhiletheguidelinesforTRACcertificationdonotprovidedetailedinstructionsorrequirementsfordisasterplanning,thecertificationdoesrequirethattherepositorybeabletodemonstratedisasterpreparedness.Thisdisasterpreparednessisgenerallydemonstratedthroughthecreationofadisasterplanor,moreaccurately,asuiteofdisasterplanningdocuments.Thecheckliststatesthat,“therepositoryshallidentifyandmanagetheriskstoitspreservationoperationandgoalsassociatedwithsysteminfrastructure”(ISO16363,2012,65).OfthethreerepositoriesincludedinthisstudythatareTRACcertified,twocreatedtheirdisasterplanningdocumentationfortheaudit(PorticoandChronopolis),andthethird(HathiTrust)completedtheauditwithoutdisasterplanningdocumentationinplace.HathiTrusthascommittedtocompletingtheirdisasterplanningdocumentationbeforetheirnextaudit.

2.3.2DRAMBORATheDigitalRepositoryAuditMethodBasedonRiskAssessment(DRAMBORA)isanothermethodforassessmentofdigitalrepositories“developedjointlybytheDigitalCurationCentre(DCC)andDigitalPreservationEurope(DPE)”(McHugh,2008,131).DRAMBORAassessment“requiresrepositoriestoexposetheirorganization,policiesandinfrastructurestorigorousscrutinythroughaseriesofhighlystructuredexercises,enablingthemtobuildacomprehensiveregistryoftheirmostpertinentrisks,arrangedintoastructurethatfacilitateseffectivemanagement”(McHugh,2008,131).ThefocusonriskintheDRAMBORAassessmentcanarguablybeseenasanalogoustotheTRACrequirementfordisasterpreparedness.TheDRAMBORAassessment,infact,hasastronger

Frank

10

focusonrisk(disaster)managementandmitigationastheentireassessmentisbasedonarepository’sabilitytomanageandrespondtorisks.Inthisrespect,“DRAMBORArepresentsabottom-upapproachthattakesriskandriskmanagementasitsprinciplemeansfordeterminingdigitalrepositories'successandforchartingtheirimprovement”(Innocenti&Vullo,2009,139).DRAMBORAuseslanguagethatplacesaheavyemphasisonriskmanagement,inserviceofevaluatingthepreservationeffortsofrepositories,“riskisutilisedasaconvenientmeansforcomprehendingrepositorysuccess-thoserepositoriesmostcapableofdemonstratingtheadequacyoftheirriskmanagementarethosethatcanhave,andengender,greaterconfidenceintheadequacyoftheirefforts.Preservationisafterall,atitsveryheart,ariskmanagementprocess.Thefundamentaltemporalchallengesofpreservationarenaturallycomplicatedbyfutureuncertainties”(Innocenti&Vullo,2009,144).WhiletheTRACcertificationprocessdiscusseddisasterpreparednessinonlyonesectionoftheauditdocumentation,“theDRAMBORAprocessfocusesonrisks,andtheirclassificationandevaluationaccordingtoindividualrepositories'activities,assetsandcontextualconstraints”(Innocenti&Vullo,2009,141).Theresultofthisprocessis“adeterminationoftherepository'sabilitytocontainandavoidtherisksthatthreatenitsabilitytoreceive,curateandprovideaccesstoauthenticandcontextually,syntacticallyandsemanticallyunderstandabledigitalinformation”(Innocenti&Vullo,2009,141).UnlikeTRACandDSA,theresultsofDRAMBORAauditsarenotnecessarilymadepublic.Oftherepositoriesincludedinthisstudy,HathiTrusthasacknowledgedthecompletionofaDRAMBORAauditbuttheresultsofthatreportarenotpubliclyavailable.

2.3.3DataSealofApprovalDataSealofApproval(DSA)isanassessmentconsistingofsixteenguidelines,which“recognizethatresponsibilityforarchivalqualitydataissharedamongstthreegroups:producersforthequalityoftheresearchdatathemselves,therepositoryforthequalityofdatastorageandavailability,andconsumersforthequalityofdatause”(Ball,2010,31).Underlyingtheseguidelinesarethefollowingfivecriteria,whichdeterminewhetherdatacanbeconsideredsustainablyarchived(Sesinketal.,2010,1):

1. TheresearchdatacanbefoundontheInternet.2. Theresearchdataareaccessible,whiletakingintoaccountrelevantlegislationwith

regardtopersonalinformationandintellectualpropertyofthedata.3. Theresearchdataareavailableinausableformat.4. Theresearchdataarereliable.5. Theresearchdatacanbereferredto.

Theguidelinesthemselvesareorganizedintothreesections,focusingonthedataproducer,thedatarepository,andthedataconsumer.Guidelinesfourthroughthirteenfocusspecificallyon

Frank

11

thedatarepository,andwhiledisasterplanningandriskmanagementarenotexplicitlydiscussedthefocusondigitalarchiving,long-termpreservation,andlifecyclemanagementarerelevanttotheareaofdisasterplanningandriskmanagement.Oftherepositoriesincludedinthisstudy,ICPSRisDSAcertified.TheresultsofthecertificationauditareavailableviatheDSAwebsite,asareallofICPSR’sdisasterplanningdocumentation.OnekeydifferencebetweenTRAC,DRAMBORA,andDSA,despitethefactthatthegoalofeachassessmentistodeterminethefitnessofarepositorytocareforandcuratecollectionsaswellasprovidelong-termpreservationsolutions,isthatTRACandDSAprovidestrictguidelinesforperforminganauditwhileDRAMBORAprovidesaframeworkthatcanbeadaptedtofittheneedsoftherepository(e.g.Ball,2010;CRL,2007;Patel,2007;Sesink,2010).Despitethesedifferencesinphilosophyanddegreeofformality,TRAC,DRAMBORA,andDSAallspecificallyincluderequirementforrepositoriestohavedisasterplanningandriskmanagementdocumentation(e.g.McHugh,2008;Ross,2006;Sesink,2010).

2.4ThreatstoDigitalCollectionsDisasterplanning,disastermitigation,andriskmanagementactivitiesarisefromrealandimaginedthreatstocollections(e.g.Aikin,2007;Altmanetal.,2009;Anderson,2005;Cervone,2006;Maniatisetal.,2005).Thesethreatscanbedividedintofourbroadcategories:

There aremany threats to archived digital information.Physical threats resultfrom chance, natural events, or age, and include failures inmedia, hardware,storage facilities, and so forth. Technological threats include formatobsolescenceanddestructivesoftwareerrors.Humanthreatsincludecuratorialerror, and insider and outsider attacks. Institutional threats include missionchange,changeoflegalregime,oreconomicfailure.Manyofthesethreatsareamelioratedthroughreplicationofthematerialstobepreserved,combinedwithregularauditing(Altmanetal.,2009,181).

Aswithdisastersfortraditionalanalogcollections,digitaldisasterscanbecausedbyphysical,human,andinstitutionalthreats.Tothislistwecanalsoaddthecategoryoftechnologicalthreat.Whilemanyincidentsthatfallintothiscategorycouldalsofallintooneoftheotherthree,andinfactnearlyeverydisasteratadigitalrepositorywillinvolvesomesortoftechnologyfailure,thistypeofincidentisuniquetodigitalrepositoriesandcaninfacthappenindependentlyoftheotherthreedisastertypes.FrankCervoneidentifiesthreetypesofdisastersfordigitalrepositories:technicalthreats,naturalthreats,andhumanthreats(Cervone,2006,175).ThisissimilartoAltman’scategorization,althoughperhapslessspecific.Digitalrepositoriesfacethreatsasdiscussedabove,butalsothreatsthatarenewanduniquetodigitalresources,“atonetime,fireandwaterwerethetwogreatthreatstoalibrary'scollectionandrecords.Nowtheyhavebeenjoinedbyother,moreinsidious,butjustasdisastrousthreats:computerviruses,hackers,fileformatobsolescence,storagemedia

Frank

12

degradationorobsolescence,platformdependence,catastrophicsystemfailure,naturaldisasters,terroristattacks,andsimpleneglect”(Anderson,2005,9).Preservationfortraditionalphysicalcollectionsgenerallyinvolvesprotectingcollectionsfromactivedangers,butbarringadisastertheobjectsgenerallydonotrequireregularinterventionforongoingmaintenancetomitigateagainst“silentcorruption”(Constantinescuetal.,2008,p.108).Silentcorruptionoccurs“whenincorrectdataisprovidedtotheuser,e.g.,writtentothememoryorI/Osystem,andnoerroristriggered”(Constantinescuetal.,2008,108).Thisisnotthecasewithdigitalobjects.Withoutregularactiveinterventions,digitalobjectswillquicklybecomeobsolete,“adigitalpreservationplanshouldincludescheduledmigrationofmaterialstonewmedia,offsitebackup,adisasterrecoveryplanandscheduledregulartestingofmediaandbackups”(Anderson,2005,10).Thisthreatofobsolescenceisinfactadisasterasitposesamajorthreattotherepository(Anderson,2005).Otherdisastersthatthreatendigitalrepositoriesincludepowergridevents,serviceinterruptions,anddatacorruption(Constantinescuetal.,2008).

2.5PlanningforDisastersDisasterplanningfordigitalrepositoriesisinmanyrespectsmorecomplicatedthandisasterplanningfortraditionalcollections.Withtheadvancementoftechnology,andthemovetowarddigitalresources,includingbothborn-digitalitemsandthedigitizationofphysicalitems,disasterplanninghastakenonnew,technologically-drivenandfocusedaspects.Whilegeneralrecommendationsandinstructionsforhandlingdamagedmaterialswillbesuitableacrossnearlyalltraditionalcollections,thisisnotthecasefordigitalrepositories.Thedisasterplanwillnecessarilyreflectthepoliciesandproceduresoftheorganization,andthesepoliciesandprocedureswillbeareflectionofthepreservationactivitiesinwhichtherepositorychoosestoengage.Forexample,adisasterplanforarepositorythatchoosestoonlybackuptheirdatauponmagnetictapewilllookquitedifferentfromtheplanforanorganizationwithamirrorsiteataremotelocation.Whileeachsolutionismeanttoaddressthesamethreatsdescribedabove,theactionsrequiredtocarryouteachpreservationactivityarequitedifferent,andthewaythatdatawouldberestoredafteradisastereventarealsoverydifferent.Mylessuggeststhefollowingactivitiesthataregenerallyapplicableacrossmanytypesofrepositories,(1)Inventoryallcomputerhardwareandsoftware.Describewhatservicestheysupport;(2)Determinewhatservicesarethemostcriticaltoyourlibrary.Describetheproceduresforcontinuingtheseservicesinadisastersituationandhowthelibrarycanrecoverfromthedisaster;(3)Makesurethatcomputerdataisbackeduponaregularbasis.Mission-criticaldatashouldbecopiedandstoredoff-site;...(5)Reviewthelistofcontingencyprocedurestodeterminewaystoreducethelengthofservicedisruption(Myles,2000,49).Disasterplanningdocumentsfordigitalrepositoriestendtoassumethatdisasters,largeandsmall,willoccurandthattheorganizationwillhavetorecover.Whileacertainamountof

Frank

13

preventioncanbehelpful,“maintenanceisalwayscheaperthanrecoveryorre-creation,soitmakesgoodbusinesssensetoplanforandfundpreservation,”therearesometypesofdisastersthatareoutsideofthecontrolorinfluenceoftherepository(suchaspowerevents)(Anderson,2005,9).Forthesetypesofdisasters,repositoriesmustdowhattheycantomitigatedataloss,“datalossincomplexsystems,whetherthroughnaturaldisasterormorelikelythroughhumanerror,isinevitable.Recoveringfromthesephenomenaisanorganizationalchallengethatwillbecomeanever-increasingdilemmaforresearch,educational,andculturalorganizationsastheirartifactsbecomeborn-digitalinnature”(McDonald&Walters,2010,4).Inanticipationoftheneedtorecoverfromdataloss,repositoriesaremovingtowardthewidespreadadoptionofbestpracticesforpreservation,“currentbestpracticeismovingtowardasystematicapproachtodatareplication,whichincludesmaintainingconsistentuniqueidentifiersforeachresource;explicitmetadatadescribingtheresources,provenance,version,andassociatedrights;andamanagedsetofreplicationservices.Bestpracticeismovingtowardmoresystematicandexplicitreplicationpoliciesthatincludemultiplyreplicatingentirecollectionsoff-site,explicitversioning,andaprocessofregularlyrefreshingandverifyingreplicatedcontent”(Altmanetal.,2009,181-2).ThesebestpracticesalsocontributetothegrantingoftrustedrepositorystatusasdescribedabovewithTRAC,DRAMBORA,andDSAcertifications.Literaturediscussingdisasterplanningfordigitalrepositoriesissparse,andassuchdiscussionisnecessarilylimited.However,thegeneraltrendsdiscussedabove,andtherecognitionbythecommunitythatdisasterplanningisabeneficialandrecommendedactionfordigitalrepositories,ispromisingandsuggeststhatthisisanareathatwillcontinuetoexpand.

2.5.1DisasterResponseandRecoveryLiteraturediscussingdisasterresponseandrecoveryfordigitalcollectionsisalsosparse.Whilethereissomeliteraturediscussingbusinesscontinuityplanningforprivatesectorcompanies,suchasfinancialinstitutions,thisliteraturedoesnotaddresssomeoftheimportantandspecificpeculiaritiesofdigitalrepositories(e.g.Andrew,2008;"BestPracticesinDisasterRecoveryBusinessContinuityPlanning,"2008;Cousins,2007;Nollau,2009;Wheatmanetal.,2001).Forexample,budgetaryconsiderationsarecompletelydifferentforaprivatecompanythananonprofitdigitalrepositorythatislikelypartofaneducationalinstitutionorlibrary.Additionally,thetypeofdataheldineachrepositorymayalsobequitedifferent.RoyTennantargues,“Oncetheemergencyhaspassed,youshouldknowwhatstepsmustbetakentogeteverythingbackupandfunctioning.Specifically,youshouldknowinadvancehowtoinstallnewhardwareandsoftware,retrievedatafromabackupsystem,andgeteverythingbackonline”(Tennant,2001,para.14).Thisadviceistruewhetherdataisbacked-uponmagnetictapesoratamirrorsite,andechoestheneedfordisasterpreparednesstrainingforstaffofdigitalrepositories.Justasstaffattraditionalorganizationsruntabletopexercisestotestoutthedisasterplan,staffatdigitalrepositoriesshoulddothesame.Staffatdigitalrepositoriesshould,infact,gothroughtheentireprocessofrestoringtheirdatafrombackup

Frank

14

sothatanyproblemsintheprocesscanbeaddressedbeforesuchactionisnecessary.Inrespondingtoadisasterwheretheinformationcontentisdamaged,animportantanduniqueconsiderationfordigitalrepositoriesistheissueofmanagingdamagedequipment.Whiledamagedequipmentcertainlyshouldbereplacedifnecessary,itisalsoagoodideatokeepthedamageditemsuntilthesystemhasbeencompletelyrestored,“tapespreviouslythoughtofasunreadablelaterturnedouttohaveusefuldata.Defectiveharddrivestoomayhaverecoverabledata.Donotletanyonedisposeofanyequipmentordatasourcesuntiltheemergencyistrulyandcompletelyover”(Brennan&O'Hara,2002,72).Despitethisapparentlackofliteratureregardingdisasterresponseandrecoveryfordigitalrepositories,literaturerelevanttodisasterplanningcanbefoundinseveralotherareas.Riskmanagementandbusinesscontinuity(orcontinuityofservice)planningaretwosuchareas,andbothwillbediscussedingreaterdetailbelow.

2.5.2RiskManagementRiskmanagementisatermthatisfoundinboththeliteratureandincommondiscussionofdisasterplanningfordigitalrepositories.Ithasbeenarguedthat,“protectingdigitalobjectsagainstthreatsisequivalenttoreducingtheriskofthosethreats,whichisthemaingoalofthebroadareaofRiskManagement”(Barateiro,2010,5).Thecertificationandassessmentprogramsdiscussedabove,TRAC,DRAMBORA,andDSAarebasedontheconceptofriskmanagementfordigitalrepositories.Ineachcase,thetrustworthinessofadigitalrepositoryisevaluatedbasedonthatrepository’sabilitytomanageriskand/ormitigatetheeffectsofdisastereventsontherepository.Thephrase‘riskmanagement’isusedinsomecasestodescribedisasterplanningactivities,andariskmanagementapproachcanbeusedtoinformdisasterresponseandrecoveryplanningactivities.Itisalsotrue,however,thatriskmanagementliteraturedoesnotplaceastrongemphasisondisasterplanningoverdigitalpreservationingeneral.Rather,riskmanagementliteraturetendstotakeamorebroadviewofriskmanagementintermsofdigitalrepositoriesand,asmentionedabove,discussriskmanagementinrelationtolongtermdigitalpreservationactivitiesandstrategiesratherthandisasterplanning.

2.5.3BusinessContinuityPlanningAsmentionedabove,andsimilartoriskmanagement,someoftheliteratureregardingbusinesscontinuityplanning(BCP)canalsobeusedfordisasterresponseandrecoveryplanningfordigitalrepositories,“BCPisconcernedwiththerecoveryandresumptionofactivitiesacrosstheentireorganization”(Cervone,2006,174).Aswithnearlyalloftheliteratureidentifiedintheareaofdigitaldisasterplanninganddigitaldisasterresponseandrecovery,thearticlebyFrankCervoneprovidessolidandclearadvicebutdoesnotprovidediscussionoranalysisofthedisasterplanningeffortsofanyparticularorganization.Whileacademicarticlesprovideinterestinganecdotalcases,andbusinessmaterials(suchasthosepreparedbyGartnerresearch)providesoundadvice,noneprovideanalysisofcurrentpractices(e.g.Battersby,2005;Fletcher,2006;Heiser,2011;McKnight,2006;Wheatman,2001;Wheatman&Witty,2001).

Frank

15

3MethodologyThegoalofthisstudyistounderstandwhetherdigitalrepositoriesthathaveapreservationmandateareengagingindisasterplanningactivities,particularlyinrelationtotheirpursuitoftrusteddigitalrepositorystatus.Incaseswheredigitalrepositoriesareengagingwithdisasterplanning,thestudyalsoexaminestheprocessofcreatingdisasterresponseandrecoveryplans,withafocusonhowtheseactivitiesareintegratedintothemanagementofthedigitalrepositories.Toanswerthesequestions,themethodologyofthisstudyinvolvesamixedmethodsapproachconsistingofdocumentanalysisandsemi-structuredinterviewstoexaminethedisasterresponseandrecoveryplanningpracticesofdigitalrepositories.ThisstudywasreviewedbytheInstitutionalReviewBoardattheUniversityofMichiganandwasgranted“NotRegulated”status.

3.1SelectionofSitesThesamplepopulationforthisstudyconsistsofdigitalrepositoriesthathaveeithersoughttrustedrepositorystatus,haveconductedaTRAC,DRAMBORA,orDSAself-audit(andmadetheresultsofthisauditpubliclyavailable),orhaveexpressedacommitmenttopursuingthistypeofcertificationprocessinthefuture.Forthepurposesofthisstudy,trustedrepositorycertificationreferstotheTrustworthyRepositoriesAudit&Certification(TRAC)asadministeredbyCRL,orDataSealofApproval(DSA)certificationastheresultsofthecertificationauditsforTRACandDSAarepubliclyavailable,andtheoutcomeofasuccessfulauditisanofficialcertification.TheinitiallistofrepositorieswascreatedinMayof2011,basedoninformationavailableviatheirownwebsites,theCenterforResearchLibrarieswebsite,and/ortheDataSealofApprovalwebsiteatthattime.Asaresultoftheanalysisofpotentialsites,aninitiallistof19organizationswascompiled.Intheendeightwereselectedforinclusioninthefinalstudybasedontheiravailabilityatthetimeofthestudyandthewillingnessofindividualsatthoseorganizationstoparticipateintheinterviewportionofthisstudy.Thisinitiallistof19repositorieswasnarrowedtothefinalgroupofeightbasedontheavailabilityofrespondentstoparticipateintheone-hourinterviewportionofthestudy.AllwhowereabletocompleteaninterviewbytheendofJanuary2012wereincludedinthestudy.ThelistoforganizationsconsideredforthestudyappearsasTable1,organizationsthatwereincludedinthefinalgroupofeightareidentifiedwithitalics:

Frank

16

Table1:InitialListofRepositories

Repository URL1 ArchaeologyDataService http://archaeologydataservice.ac.uk/2 ArchivesNewZealand http://archives.govt.nz/3 Chronopolis(TheUniversityofCaliforniaat

SanDiego)https://chronopolis.sdsc.edu/

4 DSpace(attheMassachusettsInstituteofTechnology)

http://dspace.mit.edu/

5 ECommons(Cornell) http://ecommons.cornell.edu/6 HathiTrust http://www.hathitrust.org/7 TheInter-UniversityConsortiumforPolitical

andSocialResearchhttp://www.icpsr.umich.edu/

8 LibraryandArchivesCanada http://www.collectionscanada.gc.ca/9 TheLibraryofCongress http://www.loc.gov/index.html10 MATRIX(MichiganStateUniversity) http://www2.matrix.msu.edu/11 TheNationalArchivesandRecords

Administrationhttp://www.archives.gov/

12 TheNationalArchivesofAustralia http://www.naa.gov.au/13 NationalLibraryofAustralia http://www.nla.gov.au/14 Portico http://www.portico.org/digital-

preservation/15 StatisticsNewZealand http://www.stats.govt.nz/16 TheCaliforniaDigitalLibrary http://www.cdlib.org/17 TheInternetArchive http://archive.org/index.php18 TheMetaArchiveCooperative http://www.metaarchive.org/19 TheUKDataArchive http://www.data-archive.ac.uk/

3.2DiscussionofEightSites

3.2.1ChronopolisChronopolisisageographicallydistributedpreservationnetworkthatusesiRODSto“federatethepartnersitesandtoreplicatedataamongthem”(SanDiegoSupercomputerCenter[SDSC],2011b,para.3).“OriginallyfundedbytheLibraryofCongress,theChronopolisdigitalpreservationnetworkhasthecapacitytopreservehundredsofterabytesofdigitaldata—dataofanytypeorsize,withminimalrequirementsonthedataprovider.Chronopoliscomprisesseveralpartnerorganizationsthatprovideawiderangeofservices”(SDSC,2011a,para.1).ThepartnerorganizationsthatcompriseChronopolisare:SanDiegoSupercomputerCenter(SDSC),UCSanDiegoLibraries,NationalCenterforAtmosphericResearch(NCAR),andtheUniversityofMarylandInstituteforAdvancedComputerStudies(UMIAC).“AsofJuly,2009,Chronopolishousesfourdiversecollections:abackupofthecompletedigitalholdingsoftheInter-universityConsortiumforPoliticalandSocialResearch(ICPSR,basedattheUniversityofMichigan,‘Web-at-Risk’collectionsfromtheCaliforniaDigitalLibrary(CDL),geospatialdataresourcesfromthe

Frank

17

NorthCarolinaGeospatialDataArchivingProject,andseveraldecadesofdatafromresearchcruisesfromtheScrippsInstitutionofOceanography(SIO)atUCSanDiego”(Minor,2010,121).Chronopolisfocusesonprovidinglong-termpreservationofdigitalresources.“FormatobsolescenceisnotanimmediateconcernoftheChronopolissystem.Instead,thisisregardedastheresponsibilityofthedataproviders.Thesingle,overridingcommitmentoftheChronopolissystemistopreserveobjectsinsuchawaythattheycanbetransmittedbacktotheoriginaldataprovidersintheexactforminwhichtheyweresubmitted”(SDSC,2011c,para.3).UserswillbeabletoretrievefromChronopolisexactlywhattheydepositedwithnochangestoformatorcontent.ChronopolisreceivedTRACcertificationin2012,withafinalscoreofelevenoutofapossiblefifteenpoints(CRL,2012a,4).

3.2.2HathiTrust“HathiTrustisapartnershipofmajorresearchinstitutionsandlibrariesworkingtoensurethattheculturalrecordispreservedandaccessiblelongintothefuture.TherearemorethansixtypartnersinHathiTrust,andmembershipisopentoinstitutionsworldwide”(HathiTrust,2012a,para.1).ThefoundingmembersofHathiTrustincludethe12-univeristyconsortiumknownastheCommitteeonInstitutionalCooperation(CIC,#63)andtheelevenuniversitylibrariesoftheUniversityofCalifornia(UC)system(Rombouts&Princic,2010).HathiTrustisbasedattheUniversityofMichigan’sAnnArborcampus,andispartoftheUniversityLibrary,withamirrorsitelocatedinIndianapolis,Indiana.Therepositoryfocusesbothonpreservationofandaccesstodata.“HathiTrustDigitalLibraryisadigitalpreservationrepositoryandhighlyfunctionalaccessplatform.Itprovideslong-termpreservationandaccessservicesforpublicdomainandincopyrightcontentfromavarietyofsources,includingGoogle,theInternetArchive,Microsoft,andin-housepartnerinstitutioninitiatives”(HathiTrust,2012c,para.1).ThecontentofHathiTrustisprimarilycomprisedofdigitizedmonographsandserialsfromtheparticipatingmemberinstitutions.HathiTrusthascompletedbothTRACandDRAMBORAaudits.TherepositoryreceivedTRACcertificationin2011,withafinalscoreofnineoutofapossiblefifteenpoints(CRL,2011,2).TheresultsofHathiTrust’s2008DRAMBORAauditarenotpubliclyavailable.

3.2.3Inter-UniversityConsortiumforPoliticalandSocialResearch(ICPSR)ICPSRis“aninternationalconsortiumofabout700academicinstitutionsandresearchorganizations”(ICPSR,2011a,para.1).Therepository,whichwasfoundedin1962,“providesleadershipandtrainingindataaccess,curation,andmethodsofanalysisforthesocialscienceresearchcommunity”(ICPSR,2011a,para.1).LikeHathiTrust,ICPSRislocatedattheUniversityofMichiganinAnnArbor.ICPSR,however,ispartoftheInstituteforSocialResearchratherthantheuniversitylibrary.Pertheorganization’stimeline,ICPSR’sfirstmainframecomputerwaspurchasedin1967,thefirst

Frank

18

DigitalPreservationOfficerwashiredin2006,databackupsweremovedtospinningdiskin2008,andwarmbackupserversweredeployedinremotelocationsin2009(ICPSR,2011b).Asanorganization,ICPSRhasastrongreputationforbeingaleaderinthefieldofdigitalpreservationanddisasterplanning.Inadditiontomaintainingasetofpublicly-availabledocumentsregardingdigitalpreservationanddisasterplanning,ICPSRalsoadministerstrainingworkshopsinthisarea.ICPSRparticipatedinatestauditforTRAC,administeredbyCRL,in2006.TheresultsofthisauditarepubliclyavailableviatheCRLwebsite(CRL,2012b).ICPSRreceivedDataSealofApprovalcertificationin2010,theresultsofwhichareavailableviatheDSAwebsite(DataSealofApproval[DSA],2012;DataSealofApprovalBoard,2011).

3.2.4MATRIXMATRIX:TheCenterforHumane,Arts,LettersandSocialSciencesOnlineatMichiganStateUniversityisadigitalrepositorylocatedatMichiganStateUniversity(MSU)inLansing,Michigan.TherepositorywasfoundedaspartofH-Netin1994andtoday,“housesmajordigitallibraryrepositoriesincludingtotheAfricanOnlineDigitalLibrary(AODL),DetroitPublicTelevision’sAmericanBlackJournalvideoarchives,HistoricalVoices,andTheQuiltIndex.MATRIXalsohoststheinternationalscholarlynetworkingcommunity,H-Net”(MATRIX,2012,para.2).MATRIXisfundedbyavarietyofsources,includingMSUandvariousnationalgrantsandcontinuestomaintainafocusonthehumanities,arts,socialsciences,andeducation.Pertheorganization’swebsite,MATRIXwas,atthetimeofrepositoryselection,workingonwritingseveraldigitalpreservationpolicydocuments,includingadigitalpreservationpolicyframework,adigitalpreservationplan,andadisasterplanningfordigitalassetsdocument.AsofMarch2012thewebsitecontinuestoreflectthisintent.

3.2.5NationalLibraryofAustraliaTheNationalLibraryofAustralia“definesdigitalpreservationastheprocessesinvolvedinmaintainingtherequiredlevelofaccessibilityofdigitalobjectsovertime”(NationalLibraryofAustralia[NLA],2012a,para.2).AlargepartofthedigitalpreservationeffortattheNationalLibraryofAustraliaisPANDORA,Australia’swebarchive.TheNationalLibraryofAustralia’sdigitalpreservationwebsiteincludesadiscussionofcriticalelementssuchascontingencyplanningandemergencyresponsepreparedness(NLA,2012a).Aversionofthelibrary’sdigitalpreservationpolicyisavailableonlineaswell,includingastatementthatthelibrary“storesandmanagesourdigitalcollectionsinwaysthatwillensuretheirintegrity,includingadequateandsecurebackupanddisasterrecoverysafeguards”(NLA,2012b,sec.6).Inapolicystatementcoveringtheperiodof2008to2012,thelibrary’sdigitalpreservationpolicystatesthatby2012thegoalistobe“wellplacedtopreventorrespondtothreatstothedigitalcollections”(NLA,2012c,sec.2.1).

3.2.6Portico“Porticoisamongthelargestcommunity-supporteddigitalarchivesintheworld”(Portico,2012a,para.1).Therepositoryworkswithacademicinstitutions,nonprofitorganizations,and

Frank

19

for-profitorganizationssuchaspublishers.AsofMarch2012theorganizationhas739participatinglibrariesand142participatingpublishersPorticoisaserviceofthenonprofitorganizationITHAKA.WhilePorticodoesnotprovideanyspecificdisasterplanningdocumentationthroughitswebsite,thereareseveraldocumentsavailablethroughthePreservationPoliciessectionthatcouldbeconsideredelementsofadisasterplansuchas,asuccessionplan,replicationandbackuppolicy,andescalationpathforproblemresolution.Overall,Porticoappearstobequitecustomer-andprofit-focusedincomparisontomanyoftheotherorganizationsinthestudy.Therepositoryisnotapartofanyparticularacademicinstitutionornationallibrary,andthewebsiteprovidesinformationsuchas“HowPorticoSavesYouTimeandMoney”(Portico,2012b).TherepositoryreceivedTRACcertificationin2011,withafinalscoreofelevenoutofapossiblefifteenpoints(CRL,2010,2).

3.2.7TheInternetArchiveFoundedbyBrewsterKahle,theInternetArchiveisanonprofitorganizationthatwasfoundedwiththemissionofarchivingandpreservingtheinternet.Thecollectionincludeswebpages,text,audio,video,andsoftware,althoughtheorganizationisperhapsbestknownbythegeneralpublicfortheWaybackMachine.TheInternetArchiveisbasedinSanFrancisco,Californiawithabackupsitethatislocatedatan‘undisclosedlocation’thatisalsoonthewestcoast(InternetArchive,2012).WhiletheArchiveisnotaffiliatedwithanyparticularacademicinstitution,itappearstobelessreliantonbusinessfromcustomerormemberorganizationsthantheothernon-academicrepositoriesincludedinthisstudy.ThisislikelyaresultofthecontinuedleadershipandsupportofBrewsterKahleasthefounderoftheorganization(InternetArchive,2012).In2006,theInternetArchive’sArchive-Itprogramunderwentapilotassessment,whichwasadministeredbyCRL.TheresultoftheauditcanbefoundviatheCRLwebsite(CRL,2012b).

3.2.8TheMetaArchiveCooperativeTheMetaArchiveCooperativewasfoundedin2006asamembershipofsixacademiclibraries.TheLOCKSSnetworkhassinceexpandedtoinclude“libraries,archives,andotherdigitalmemoryorganizations”(EducopiaInstitute,2012,para.1).TheCooperativepromotesaphilosophyofencouraginginstitutionstopreservetheirowndataratherthanoutsourcingpreservationservicestoexternalvendors.ThisisaccomplishedbyhavingeachinstitutionintheCooperativemaintainaserverthatisconnectedtothenetwork(EducopiaInstitute,2012).Thegeographicallydispersedlocationsofthememberinstitutionshelptomakethepreservationmoresecure.

Frank

20

DocumentationavailableviatheMetaArchiveCooperativewebsiteareportoftheresultsfromaTRACassessmentthatwascarriedoutbyanindependentconsultant.Noscoreisavailableasthiswasnotanofficialaudit(Schultz,2010).

3.3DocumentAnalysisDocumentsforanalysiswerecollectedintwoways:Internetsearchingandaskingparticipantsduringtheirinterviews.TheInternet,andspecificallythewebsitesoftherepositoriesandtheirparentorganizations,wassearchedforreadilyavailabledocumentationregardingdisasterplanninganddigitalpreservation.Atsomeorganizations,suchasICPSR,thisinformationwaseasytofindandtointerpret.Atothers,suchasPortico,itwasmuchmoredifficult.Similarly,someintervieweeswerehappytosharedocumentsandinformationandotherswerereluctant,unwilling,orunabletodoso.Giventhedesireforopennessandtransparencyamongorganizationswhopurportthemselvestobetrustedrepositories,andorganizationsthathaveundergoneTRACcertificationandreview,itwassurprisingthatdisasterplanningdocumentationwassodifficulttofind,andinmanycasesnotavailableatall.Documentsselectedforanalysisincludeanyandalldocumentsthathavebeenidentifiedbyaparticularrepositoryasbeingrelevantfortheirdisasterresponseandrecoveryplanningefforts.Thisincludesdisasterresponseandrecoveryplans,contingencyplanningdocuments,businesscontinuityplanningdocuments,successionplanningdocuments,preservationplanningdocuments,andTRACandDSAauditreports,aswellasdocumentsforinternaluseonlysuchastrainingdocumentsandmemos.Theavailabilityofthisdocumentation,andthetypesofdocumentationselectedforconsiderationforthispurposesofthisstudy,helptoprovideinsighttohoworganizationsareengagingindisasterplanning.Theavailabilityofthisdocumentationalsohelpstoshowwhatinformationrepositoriesmakeavailableinrelationtotheirpursuitordemonstrationoftrustedrepositorystatus.Whichistosaythatsomerepositoriesseemtofeelthatmakingdisasterplanninginformationpubliclyavailablehelpstocultivatetrustfromthecommunityandothersseemtoplacelessimportanceonmakingthisinformationavailable.AcompletelistingofthedocumentsincludedappearsasTable2below:Table2:AvailableDisasterPlanningDocumentation

Repository DocumentsChronopolis TRACCertificationReport

DigitalPreservationProgramWebpage

HathiTrust TRACCertificationReport

“HathiTrustisaSolution”Report

Frank

21

“BuildingAFutureByPreservingOurPast:ThePreservationInfrastructureofHathiTrustDigitalLibrary”

InternalPlanningDocuments,BusinessImpactAnalysis

ICPSR DisasterPlan–Records/Finance

DisasterPlan–RecordsHR(June2007)

DisasterPlan–MemberServices

DisasterPlanningResources(2008)

DisasterPlanningICPSRUpdate(2007)

CrisisCommunicationsPlan

DisasterPlanningPolicyFramework:ModelDocument

DisasterPlanningRolesandResponsibilities:ModelDocument

Version2.0DisasterPlanningTraining:ModelDocument

DisasterPlanningShort-TermActionPlan

DisasterPlanning:CrisisCommunicationsPlan

DisasterPlanning:WebServicesContinuityPlan

DataSealofApprovalAssessmentReport

CRLTRACAuditReport

MATRIX InformationSecurityforDigitalAssetsatMATRIX

NationalLibraryofAustralia BuildingTrust:PilotPreservationAuditofNationalLibraryofAustraliaDigitalRepository

NationalLibraryofAustraliaRequestforTender:DigitalLibraryInfrastructureReplacement(RFT11103)

NationalLibraryofAustraliaCollectionDisasterPlan

NationalLibraryofAustralia,DigitalPreservationPolicy,3rdEdition

Portico CRLReportonPorticoAuditFindings

PorticoTRACSelf-Report

TheInternetArchive StorageandPreservationwebpage

PetaboxStorageSystemInformationwebpage

CRLArchive-ItReport

TheMetaArchiveCooperative MetaArchiveCooperativeCharter

Frank

22

MetaArchiveTechnicalSpecifications

MetaArchiveTRACAuditChecklist(self-audit)

3.4InterviewsInterviewsubjectswereidentifiedateachoftheinitial20organizationsandwereselectedforinclusioninthisstudybasedoninformationavailableontherepositories’websitesindicatingthattheyareresponsiblefor,orinvolvedin,disasterresponseandrecoveryplanningactivitiesordigitalpreservationactivities.Insomecasesmultipleindividualswereidentifiedforasingleorganization.InterviewswereconductedwithtenindividualsfromeightdifferentorganizationsbetweenOctober2011andJanuaryof2012.AlistingofparticipantsinterviewedisincludedinTable3below:Table3:ParticipantsInterviewed

Code Repository Title/Role FunctionSubjectA Chronopolis DigitalPreservation

Librarian/ProjectManager

DigitalPreservation

SubjectB Chronopolis ProjectManager DigitalPreservationSubjectC HathiTrust AssistantLibrarian ITSubjectD HathiTrust DigitalPreservation

LibrarianDigitalPreservation

SubjectE ICPSR DigitalPreservationOfficer

DigitalPreservation

SubjectF MATRIX ChiefTechnologyOfficer ITSubjectG NationalLibraryof

AustraliaManagerofDigitalPreservation

DigitalPreservation

SubjectH Portico ArchiveServiceProductManager

Administration

SubjectI TheInternetArchive Director,ArchivingServices

Administration

SubjectJ TheMetaArchiveCooperative

MetaArchiveProgramDirector

Administration

Ofthoseinterviewedforthestudy,threeholdadministrativeroles,fiveholddigitalpreservationroles,andtwoholdpositionsininformationtechnology(IT).Astheanalysiswillshow,theserolesaresignificantinthatthesubjectsholdvaryingamountsofresponsibilityandauthoritywithintheirorganizations.Eachalsoplaysadifferentroleindisasterplanningactivitieswithinhisorherrespectiveorganization.Onelimitationofthisparticularsubjectgroupisthatonlyintwocasesweremultiplepeopleatonerepositoryinterviewed.Specifically,ChronopolisandHathiTrust.Astudythatinterviewedmultiplepeoplefromdifferent

Frank

23

departmentsandfunctionswithineachorganizationmightbeabletoprovideamorecompleteviewofdisasterplanningatthoseorganizations.Forexample,individualsinITwereabletospeakaboutthetechnicalsideofpreservationandadministratorswereabletospeakaboutpolicy,butneitherwaswell-versedinboth.Subjectswerecontactedviaemail,withasecondfollow-upmessagesenttothosewhodidnotrespondtothefirstemail.Atotalof21responseswerereceived.Eightrespondersprovidedareferraltoanotherindividualwithintheorganization,thirteenagreedtoparticipate,andonedeclinedtobeinterviewedbutofferedtoanswerquestionsviaemail.Ofthetwelvewhoagreedtoparticipate,onewasunabletoscheduleaninterviewinthetimeframeallotted,anotherwaswillingtobeinterviewedbutdeclinedtoberecordedortoshareanydocuments,andthethirdhadmovedintoanewpositionaswasnolongerinvolvedindisasterplanning.Thisproducedalistoftenfinalistswhowereselectedtoparticipateintheinterviewphaseofthestudy.Interviewswerescheduledwiththosewhorespondedexpressinginterest.Participantsweresentconsentformstoreviewandsignpriortotheinterview(seeAppendixA).Interviewslastedapproximatelyonehoureachandwereconductedviatelephoneorinperson,dependingonthelocationandavailabilityofthesubject.Allinterviewswererecordedusingtwoseparatedevicesinordertoensureareliablecaptureoftheevent.Post-interviewnotesweretakenaswell.Theinterviewsfollowedasemi-structuredlistofquestions,whichallowedeachparticipanttodiscusstheirownpoliciesandpractices,elaboratingwhenappropriate.Asemi-structuredinterviewallowstheinterviewertofollowapredeterminedlistofquestions,butallowsformodificationtothewordingand/ororderofthosequestions.Theintervieweralsohastheabilitytofurtherprobeparticularareasinordertoelaborateorclarifythesubject’sresponse(e.g.Babbie,2010;Robson,1993;Wildemuth,2009).Questionsaskedcoveredtheareasof:organizationalattitudestowarddisasterresponseandrecoveryplanning,developmentofdisasterplanningdocumentation,accesstodisasterplanningdocumentation,useandmaintenanceofdisasterplanningdocumentation,andbudgetaryconsiderations.Questionswerebasedonareviewoftheliteratureandapreliminaryreviewofthewebsitesandavailabledocumentationateachofthe20initialorganizations.Thesequestionsfallintothreeareas:creation/development,access,anduseofdisasterplanningdocumentation.Foracompletelistingofquestions,pleaseseeAppendixB.

3.5InterviewAnalysisOncecompleted,theinterviewsweretranscribedandcodedusingNVivo.Thesystemforcoding(or‘nodes’)wasdevelopedbasedonareviewoftheliterature,apreliminaryreviewofthewebsitesandavailabledocumentation,andinitialimpressionsfromtheinterviewsthemselves(e.g.Holsti,1969;Wildemuth,2009).Nodesfallintogeneralcategoriesofcommunication,documentation,administration,andpreservationandaredescribedingreaterdetailinTable4below.

Frank

24

Table4:DescriptionofDocumentCodingScheme

Node DescriptionAccess Accesstothedisasterplanningdocumentationforbothmembers

oftheorganizationandthegeneralpublic.Backup Anymentionofdatabackupusedfordigitalpreservationor

disasterplanning.BackupSites Usedonlywhenaparticipantdiscussedbackupviareplicationof

dataatmorethanonelocation.BackupTape Usedonlywhenaparticipantdiscussedbackuptapes(usually

storedataseparatelocation).LOCKSS UsedonlywhenanorganizationisamemberofaLOCKSS

system/network.BarriersorDifficulties Barriersordifficultiestodisasterplanningactivities.Usually

barrierstocreationorimplementationofdisasterplanningdocumentation.

Budget Discussionofhowdisasterplanningactivitiesarefinanced,howtheyfitintothebudget.Alsoincludesdiscussionofhowdisasterresponsefitsinthebudget.

Certifications DiscussionofauditandcertificationprocessesfortrustedrepositoriessuchasTRACorDataSealofApproval.

Collaboration Collaborationwithexternalorganizationsinordertofurtherdisasterplanningefforts.

Data Discussionoftheactualdataheldwithintherepository.DataLoss Discussionofspecificeventsinvolvingdataloss,orofconditions

underwhichdatalossoccurs.DataRecovery Discussionofspecificeventsinvolvingdatarecovery,orof

conditionsunderwhichdatamayberecovered.DisasterEvents Discussionofspecificdisasterevents,realorpotential.DisasterPlanDocumentation

Discussionofcreation,development,andimplementationofdisasterplanningdocumentation.

InternalCommunication Communicationwithintheorganizationwithregardtodisasterplanning,responseandrecoveryactivities.

Preservation Discussionofdigitalpreservation,notnecessarilywithaspecificfocusondisasterplanning.

Whilethiscodingschemewaslargelysuccessfulinhighlightingthetopicsofdisasterplanningandauditandcertificationprocessesfortrusteddigitalrepositories,thismethodofdocumentcodingandanalysisdoeshavesomelimitations.Forexample,havingonlyoneresearchercodinglikelymakestheresultsmoresubjectivethantheywouldbeifmultipleresearcherswereindependentlycodingandcomparingresults.Inaddition,havingthesameresearcherwhoconductedtheinterviewsalsocarryoutthecodingintroducesanotherlayerofpotentialbias.

Frank

25

4.FindingsThecombinationofcodedinterviewdataanddocumentanalysisyieldfindingsinsixareas:

1. IncentiveforCreation2. Documentation3. ProcessofCreation4. Obstacles5. TestingthePlans6. AccesstoDisasterPlanDocumentation

Thesecategorieshavebeeninfluencedbythestructureoftheinterviews,informedbypatternsofanalysisfromNVivo,andareorganizedinanorderroughlyreflectingthechronologicalprocessofdisasterplancreationandimplementation.

4.1IncentiveforCreationManyofthesubjectsexpressedtheideathatthedevelopmentofdisasterplanningpoliciesandprocedureshappenedasaresultofgrowthanddevelopmentoftherepository.However,eventhosewhoinsistedthattheirrepositoryhaddisasterplanningpoliciesandproceduresinplacebeforegoingthroughanauditcertificationprocessindicatedthatitwasonlythroughtheprocessofrespondingtotheneedsoftheauditorsthattheyactuallycreatedtheirformaldisasterresponseandrecoveryplanningdocuments.SubjectBfromChronopolisstated,“everybodykindofknew,butyouhadtobepartoftheChronopolisteamtoknowwhatthedisasterplanwas.NowduringtheTRACaudit...Chronopolisitselfhashadtomakethatmorepublic.”SubjectAfromChronopolisechoedthisstatement,“it[theTRACaudit]reallydidpushustocreatealotofdocumentationandtobeveryexplicitaboutthingsthatwehadjustkindofassumedbeforeorthatwehadn’tputintoplaceorhadlanguagefor.”AndinadditiontobothacknowledgingthattheTRACauditprovidedtheincentivetoformalizeinformationthathadbeeninformallyortacitlyunderstoodwithintheorganization,theyeachalsostatedmoreexplicitlythat“themainreasonthatwewrotespecificdocumentswasaspartofanauditprocesswhichbeganaboutthemiddleoflastyear”andthat“weonlydocumentedallofthisbecauseofTRAC.”TheMetaArchiveCooperativeisanorganizationthathasgonethroughaninternalTRACaudit.MuchthesameasSubjectsAandBfromChronopolis,SubjectJstatedthatdisasterplanningpoliciesandprocedureshadbeeninplacepriortotheaudit,“Iwouldsayinsomewaysthedisasterplanningactionhasbeeninplacesince2004,sincewefirstbroughtupthenetwork.”Healsostatedthat,“thereisasecondsetofdocumentationthatwepreparedinresponsetoaTRACauditthatwedidin2008,”addingthat,“it's[TRAC]verygoodatcrystallizingandcondensingdownwhatthingsyoushouldbedocumentinganditgivesyouagoodbaseinmyexperiencefordefiningandmakingsurethatyourpracticesareassophisticatedastheyneedtobeinordertoguaranteethatyou'redoingdigitalpreservation...thedisasterrecoverypieceisaperfectexamplebecausethatdocumentandthesuccessionplanningdocumentthosehavecomeoutofthatTRACexperience,notbecausewehadn'talreadythoughtthroughthosethings

Frank

26

andhadthemdocumentedinotherways,wedidnothaveonedocumentthatsaid'thisfocusescompletelyonthattopic'andthat,theimportanceofthat,washighlightedintheTRACdocumentandIthinkrightfullyso.Ithelpedtomotivateus.”Againinthiscase,itwastheauditprocessthatprovidedtheincentivetocreatethedocumentsthatarecurrentlyconsideredtobetheorganization’sdisasterplanningdocumentation.SubjectHfromPorticostatedthat,“growingsizewasthebigimpetus...Idon'tthinkthattherewasasituationthatcausedustothink‘ohmygoodnessweneedasetofplans.’Itwasmore,‘wowwe'regettingreallybigandifweeverhavetorecoverwewantwrittendownthestepswe'regoingtodo,’soitwasjustenvironmentmorethananythingelse...Astheorganizationgrows,andthecontentgrows,thesethingsbecomemoreurgentinpartfortheverypracticalreasonthattherealityisthatifadisasterweretooccurastheorganizationhasgrownandgottenbigger,thetraumaofrecoveringgetsbiggerandsotheverypracticalneedstohaveaplaninhandbecomemoreurgent.”Evenmorethanthepreviouslymentionedrepositories,thisinterviewemphasizedthefactthatthedisasterresponseandrecoveryplanningpoliciesandprocedureswerecreatedoutofarecognizedneedtosupportthegrowthoftherepository.However,eveninthiscase,theintervieweewentontoexplainthat,“ourfinalpoliciesweredraftedinsuchawaytotryandanswerspecificquestionsinTRAC...wedidn’treallywriteanythingspecificallytomeetTRACsneeds,but...whenwesatdowntowritepoliciesitwasreallyjustwritingpoliciesthatmatchedthedecisionsweweremaking...forthepurposesofCRLandtheirTRAC-typeaudit,wespentquiteabitofeffortframingthepoliciesandallofourotherdocumentationinsuchawayastohelpitanswerTRACquestions...itwasanordeal,itwasnoteasy.”Thestaffofthisrepository,likeChronopolisandTheMetaArchiveCooperative,ultimatelycreateddisasterplanningdocumentationbasedonwhatwasneededfortheiraudit.WhiletheintervieweesfromHathiTrustacknowledgedduringtheinterviewthattherepositorydoesnothaveformalizeddisasterresponseandrecoverydocumentationinplaceyet,SubjectDstatedthat,“alotofwhereweareintheplanningisdocumentingwhatwe'realreadydoing,creatingformalpoliciesthatdescribewhatwe'redoing,andthiswasalargepartofourTRACcertificationalso.”Inthiscase,theirpositionseemsalsotobethatthepoliciesandproceduresexistandtheprocessofdocumentingthesepoliciesandproceduresissomethingthatisbeingundertakenspecificallyasaresultoftheaudit.Oftheintervieweesfromorganizationsthathaveundergoneanauditforcertification,SubjectEfromICPSRistheonlyonethatdidnotreportthedevelopmentofspecificdisasterplanningdocumentationfortheaudit.Muchliketheothers,SubjectEdevelopeddisasterresponseandrecoverypoliciesandproceduresasaresultoforganizationalgrowthanddevelopment,“Ithinkitwasjustageneralsenseofalignmentwithgoodpractice.Thereasonthatwestartedtalkingaboutitatallwasthattherewasasensethatweneedsomekindofdisasterplanninginplace.”Additionally,“wehadthreenear-misssituationsorthingsthatwouldbeconsideredatleastemergencysituations,noneofwhichwecouldhavedescribedaheadoftime.”However,theresultofthisdecisiontodevelopthesepoliciesandprocedureswasthecreationofformalizeddocumentation,includingtemplatesforusebyotherorganizations,thatthestaffofICPSRhas

Frank

27

madeavailableviatheDisasterPlanningsectionoftheirwebsite.Ratherthanfollowingaformatbasedontheneedsofanaudit,theyreliedon“theNISTmodel.”Stafffromtheremainingrepositories(MATRIX,theInternetArchive,andtheNationalLibraryofAustralia),alldiscussedtheexistenceofdisasterplanningpoliciesandprocedures,generallyasdocumentsorcheckliststhatresidewithintheITdepartment.Itseemsthattheserepositorieshavemetthefirststepasdescribedbytheotherorganizationsdiscussedinthissection,buthavenotyettakenthestepofformalizingthesepoliciesandproceduresasasetofdedicateddisasterplanningdocuments.Onecouldspeculatethatiftheserepositorieswentthroughanaudittheywouldlikelycompletetheprocessbycreatingthesedocuments.Thethemeofformalizeddisasterplanningactivitiesforthepurposesofauditorcertificationwasprevalentthroughoutthediscussionswiththoserepositoriesthathavebeenthroughsomeformofaudit.Accordingly,thedocumentationcodingschemewasabletocapturetheoverlapbetweenrespondentsdiscussingcertificationanddisasterplanningdocumentation.Ascanbeseeninthechartbelow,institutionsthathaveundergonecertificationaudithaveahigherlevelofoverlapbetweendiscussionofcertificationanddisasterplanningdocumentationaswellasahigherinstanceofdiscussingeachindependently.

Figure1:CertificationandDisasterPlanningDocumentation

WhileTRACisonlyoneofseveralpossibleauditsthattherepositoriesinthisstudyhaveundertaken,TRACcertificationisthemostcommonlyheldtypeofcertification.Amongmyrespondents,threerepositoriesareTRACcertified,onerepositoryisDSAcertified,andseveralothershavecompletedinformalself-audits.OfthethreerepositoriesthathavebeenTRAC

0

5

10

15

20

25

30

Certi,ication

DisasterPlanningDocumentation

Certi,icationandDocumentation

Frank

28

certified,thetwowiththehighestscoresreportthemostcompletedisasterplanningdocumentation.

Figure2:TRACAuditResults

4.2DocumentationWhilealloftheintervieweeswerequicktoprovideassurancethattheirorganizationdidindeedhavedisasterplanningdocumentation,andmanywerehappytoprovideevidenceofthatdocumentationintheformofanauditreport,veryfewwereableorwillingtodiscusstheseplansindetailortoprovidecopiesofthecompletedocumentation.

0

1

2

3

4

5

6

Chronopolis HathiTrust Portico OptimumRating

OrganizationalInfrastructure

DigitalObjectManagement

Technologies,TechnicalInfrastructure,Security

Frank

29

Figure3:DisasterPlanningDocumentation

BothintervieweesfromChronopolisexplainedthattheorganization’sdisasterplanningdocumentationwascreatedinresponsetotherecentTRACaudit.PerSubjectA,“ingeneraltermswecreatedaTRACreportwhichbasicallyfollowsthequestionandanswerschemaoftheTRACaudititself.”Butuponfurtherdiscussion,SubjectBrevealedthattheChronopolisdisasterplanprimarilyservestopointuserstootherdisasterplanningdocuments,“wedohaveadocumentthatisChronopolisdisasterplanning,butalltheinstructionsforthatdisasterplanninglinkouttootherplaces.”Specificallythat,“Chronopolisisaconsortiumofthreeinstitutions...andeachofthoseentitieshasaspecificdisasterplanforwhathappenstodataintheirdatacenters.AndsowerelyonthosedisasterplansinthosedatacenterstomakeupthewholedisasterplanforChronopolis.”SubjectBdiscoveredwhilereviewingdocumentationduringtheinterviewthat,“it'sjustastatement,wedon'tactuallylinkouttotheotherinstitutions.”Meaning,theinformationthatisavailabletothepublicreferencesotherdocumentsbutdoesnotprovidelinkstothosedocuments.InthewordsofSubjectB,“they'reactuallydifficulttofind.Soyes,they'reavailabletothepublic-butthey'reavailableifyoucanfindthem.”Theintervieweeexplainedthathewasactuallyunabletofindthedocumentswithoutassistancefromtheindividualresponsibleforthemattheparentorganization.SubjectFfromMATRIXstatedthat,“wehavepracticesandwehavesomedocumentationindifferentlocationsthatmoreorlessequatetothat[disasterplanning]butwedon'thaveadirectformalplanthatspeakstoexactlywhatwe'lldointheeventofadisaster.”Inaninterestingexchange,SubjectFexplainedthataformaldisasterplanisnotneededbecausethestepsrequiredtorecoverfromadisastereventaresoobviousandsimplethatanycompetentSystemAdministratorwouldunderstandhowtocarryoutthisaction.“Wehaveawikiandwe'vebeenputtingalotofourdocumentationonthat.Andwedohavealotofourdocumentsregardinghowtobringthesystembackup,andwhatourplansare,andwhatourproceduresare.They'renotinoneactualspotonthewikiyetbutwe'regettingtothatpoint,andreally

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

FullDocumentationAvailable

SomeDocumentaiton

Available

NoDocumentationAvailable

NumberofRepositories

Frank

30

partofthedecisiontomakewithusisdowefocusondocumentingmoreorlessaknownprocedure...mostSysAdminswouldunderstand‘okthere'satapebackup,takethetapebackupandrestoreit’andnowyou'regoodtogomoreorless.Imeanifatworstcasesomeonehopefullywouldknowtoputatapeinthedrive,right?It’scommonsense...atthatpointit'sreallyaquestionoftowhatlevelofdetaildoweget...butdocumentationwehaven'treallyfocusedonalotjustbecauseofthefactthatwe'renotatapointwherewe'recomplexenoughtorequireitinmyopinion.”Thisisanopinionthatwasnotexpressedbyanyothersubjectinthisstudy,andwhichmaybearesultofthefactthatthissubjectwasinarolewithanITfunctionratherthanarolewithapreservationfunction.Aswillbediscussedlater,manyintervieweesdiscussedhavingdifficultyingettingproperdocumentationfromtheITdepartmentswithintheirrespectiveorganizations.Thisdiscussionperhapsprovidessomeinsighttotheothersideofthatfrustration.SubjectJfromTheMetaArchiveCooperativeexplainedthat,“wehavedocumentedcontingencyplansthatlookatanumberofdifferentpointsontheaxisofproblemsthatcoulderuptandwhatwouldhappeninthosekindsofdisasterscenarios”andalsothat,“thereisacontingencypieceforeachoneofourmemberinstitutionsthatispartoftheirowndisasterplanningsotherearethesetwolayerstodisasterplanningasweseeitatMetaArchive.”Inthiscase,therearemultiplecomponentstothedisasterdocumentation.SubjectJgoesontoexplainthat“intermsofdocumentation,itstartedwithourmembershipagreementandourcharterandthosetwocoredocumentsarethelegalunderpinningsfortherelationshipsthatcomprisetheMetaArchivenetwork”aswellas“asecondsetofdocumentationthatwepreparedinresponsetoaTRACauditthatwedidin2008that[resultedin]aformalizedcontingencyplandocumentandsuccessionplan.”Inthiscase,theintervieweewasabletodiscussanddescribeseveraltypesofdocuments,butagainspecificdisasterplanningpoliciesandprocedureswerenotavailable.ThestaffofPorticohavecreatedseveraldifferentdocumentsthatcomprisetheirdisasterplanningdocumentation.PerSubjectH,“ourpoliciesareverytargeted,sowedon'thaveonebigoverarchingpolicyforPortico.Wehaveaseriesofsmallerpolicies...soatPorticowe'vegot13,16,21differentpolicies...Iwouldsaythatthereareprobablythreepoliciesthataredirectlyimpactingdisasterrecovery.”Thisinterviewalsorevealedthat,“wehavetwosets[ofdisasterplanningdocuments].ThereisthesetthatismaintainedbyourITgroup...theyhaveasetofdisasterrecoverpoliciesthattheyhavedevelopedthatinvolvesalotofthisinfrastructuretypestuff.Porticoproperhasasetofgrouppreservationpoliciesarounddisasterrecovery,whichspecifythenumberofbackupsweneedtohavethenumberofreplicaswherethey'regoingtobelocated,ourgeneralphilosophyaboutit.”MuchliketheintervieweefromMATRIX,SubjectHfromPorticoseemstobedescribingadisconnectbetweentheITandPreservationfunctionswithintheorganization.Ratherthanhavingonesetofcombineddisasterplanningdocuments,therearetwoseparatesetsofdocumentsthatdonotseemtobecombinedinanyformalorsignificantway.Infact,theintervieweewasabletodiscusstheITdocumentsinonlybroadstrokes.

Frank

31

ThestaffofHathiTrustweretotallyopenintermsofsharingtheirdocumentationandworkinprogress.AsSubjectCstated,“it’sinprogressrightnow.Wehaveafoundationaloutlinethatwe’reworkingfrom”and“it’snotafunctionalrecoveryplanbyanymeansbutthegoalistogettothat.”Thisdiscussionreinforcedtheideathatthepoliciesandproceduresneededforactualdisasterresponseandrecoveryareinplace,andthatthecreationofformaldisasterplanningdocumentationisaformality,“alotoftheproperthinkinghasbeendoneinverymanywaysandtheproperworkhasbeendonetoensurethatthingswilllikelyfunctionverysmoothlyintheeventofadisaster,buttheworkhasnotbeendonetofullyarticulatetheprocessesinwhichitwilltakeplace.”ThestaffofICPSRwerealsocompletelyopenaboutsharingtheirdisasterdocumentation.SubjectEstated,“wefollowtheNISTmodelforthetypesofdocuments.Soit'sasuiteofdocumentsit'snotasinglething.It'songoing,it'saplanningprocess,thefocusisonplanningasaverb,notplanasanoun.”Nearlyalloftheirdocumentationwasavailableviathedisasterplanningsectionontheorganization’swebsite,andtheindividualinterviewedwasabletosharetheremainingdocumentsviaemail.ThestaffoftheInternetArchivehave,“aninternalchecklistabsolutelywhichwereview”thatismaintainedbyanITdepartment.However,asdiscussedabove,thischecklistisnotconsideredtobeacompletedisasterresponseandrecoveryplan.Itisalsonotavailabletothepublicinanyform.SubjectIinthiscasewaseitherunwillingorunabletodiscussspecificsofthisplan.SubjectGfromtheNationalLibraryofAustraliaexpressedasimilarsituation,“wehaveadigitalpreservationsection,andwehaveaverylargeITsection,andtheITsectiondealswithalotofthethingslikebackups...andsoasmuchasIcouldsaytoyou'yeswedohaveabackupregime'Ican'tgiveyoutheexactdetailsofitbecausetheyrunthosekindofthings.”ThedisconnectbetweenthepreservationandITfunctionsatthisorganizationissogreatthatthedigitalpreservationsectionisactuallynotfamiliarwiththedisasterplanningdocumentationatall.

4.3ProcessofCreationOneoftheprincipalfindingsofthestudyisthattheprocessofseekingcertificationisextraordinarilytimeconsumingandrequiresamajorcommitmenttodocumentation.Thecreationofdisasterplanningdocumentationisreportedbymostrepositoriestohavebeenoneofthemosttimeconsumingaspectsofthisprocess.Discussionswithsubjectsfocusedmuchmoreontheinvestmentoftimeandpeopleintotheprocessofcreatingdisasterplanningdocumentationthanonthespecificsofdecisionsmadewithinthosedocuments.ForSubjectAatChronopolis,“itwastheauditthatwas100%ourguidingforceincreatingthesedocuments.”Andspecifically,they“tooktherequirementsthatthey[TRAC]hadlisted,puttheminabigexcelspreadsheetandthenusedthattodrilldownintospecificquestions.”SubjectAwasalsoabletodiscusssomeoftheindividualsinvolvedintheprocess,althoughnotingreatdetail,“internallywehadacoupledifferentindividualscontribute,theprimaryoneobviouslywouldhavebeentheDataCenterManager,whowasinchargeofmaintainingthe

Frank

32

equipment.”Hewasalsoabletospeaktothecreationofdisasterplanningdocumentationatthethreepartnerinstitutions,“andsimilarlythatwasthecaseatallthreeinstitutionsthatthosetypicallywerethepeopleinvolvedincreatingthedocumentsneededfortheaudit.”Thedocumentationtook“agoodthreetofourmonths”tocomplete,and“wasoneofthemoresignificantsectionsthatwehadtodoalotofnewworkfor...it'sprobablyoneofthelargersectionsforusintermsofhowmuchtimewasspentonit.”ThediscussionwithSubjectsCandDfromHathiTrustfocusedonbothhowtheyareproceeding,andonhowtheyexpecttocreatetheirdisasterplanningdocumentation.ThegeneralphilosophyoftheorganizationexpressedbySubjectDto“doverymuchwhat'spracticalandtrynottopredictthefuture”andtheinterviewwithsubjectsfromHathiTrustreinforcedthisidea.Specifically,SubjectDstatedthat,“wetakeourpracticesandarenotconstrainedbyapolicynecessarilyuntilweactuallydothatthing.”Meaningthatthegoaloftheirdisasterplanningdocumentcreationprocessistoarticulatecurrentpracticesratherthansettingrigidpoliciesandproceduresthatwillneedtobeimplementedandenforced.SubjectJfromtheMetaArchiveCooperativefocusedontheamountoftimespentcompletingthedisasterplanningdocumentation.“Fortheinitialinvestment,whenyoulookatallofthedifferentpeoplewhoareinvolvedandallthedifferentstagesofthatdrafting,Iwouldsayatleast80hoursofpeopletimewentintothedrafting.Nottheapprovalprocess,notthecontinuedrevisionsthatwe'restilldoing,butjustthebase-leveldraftingtoreallygetallofthisdoneandlinedup...atleast80hours.”SubjectJ,however,didnotdiscussthespecificsofhowthattimewasspent,orofwhosetimewasspent.SubjectHfromPorticoalsotalkedaboutlengthoftimetocompletethedisasterplanningdocumentation,althoughitwasdiscussedonlyaspartofalargerprojecttodocumentdigitalpreservationpolicies.“Idon'tknowspecificallyarounddisasterrecovery,Iwouldsaythatitwasprobablyasixmonthprocessforus...weprobablytookaboutsixmonthstoreallyformalizeandfinalizearelativelysubstantialsetofourpreservationpolicies,disasterrecoverybeingoneelementofthatthewholeprocess...itwasactuallyquiteachunkoftimewithparticipationfromthreeorfourpeople.Itwasnotaneasyprocess.”SimilartoHathiTrust,thestaffofPorticobasedtheirdisasterplanningdocumentationdecisionsonfactorsoutsideoftheTRACauditprocess,focusingonthreespecificelements,“environmentalreviewofwhatisrecommended,someverypracticalconsiderationsaboutwhatisphysicallypossible,andthensomebusinessconsiderationsaboutmarketingandoutreach.”AlsolikeHathiTrust,thestaffofPorticoexpressedtheorganizationalattitudeofcraftingpoliciesanddocumentsthatmeettheirneedsanddescribetheirpracticeswithoutbeingoverlyprescriptiveorrestrictive,“wehaveastandardtemplateforhowwewritepolicydocuments.Ourpolicydocumentstendtobe...relativelyhighlevelandstrategic,whichallowsustowriteandchangeimplementationsovertimeandhavedifferentimplementationdocumentationtosupportthehighlevelstrategy.”Inthiscase,SubjectHisagainconfirmingthepreviouslydiscussedfindingthatrepositoriesarereluctanttodiscussdisasterpreparednessdocumentationingreatdetail.

Frank

33

4.4ObstaclesThemajorityofintervieweesincludedinthisstudyreportedsignificantobstaclesorchallengesencounteredintheprocessofcreatingtheirdisasterresponseandrecoveryplanningdocumentation.Themostcommonthemeswerethedifficultyofgettingbuy-infromothermembersoftheorganization,difficultycollaboratingandcommunicatingwiththeITdepartment,andtheamountoftimerequiredforcompletionofthedocumentation.Theseobstaclesalignwiththepreviousfindingthatmostrepositoriesthathaveformalizeddisasterplanningdocumentationcreatedthatdocumentationastheresultofanaudit.Inotherwords,theywereunableorunwillingtocreatethedocumentationwithoutanorganizationalmandatetodoso.Thechartbelowindicatesthenumberoftimesthetopicofbarriersordifficultiestodisasterplanningwasmentionedineachinterview.Thischartshowsthattheintervieweefromtheorganizationwiththemostcompleteandpubliclyavailabledisasterplanningdocumentation(ICPSR)alsospentthemosttimediscussingobstaclestodisasterplanning.TherepositorywiththenexthighestincidenceofdiscussionofobstaclesisHathiTrust,arepositorywhosestaffhasyettocompletetheirdisasterplanningdocumentation.

Figure4:Obstacles

SubjectAfromChronopolisdescribedtheprocessofcreatingthedisasterplanningdocumentationas“herdingcats.”HewentontosaythatthemostsignificantbarriertocompletingthedisasterplanningdocumentationpriortotheTRACaudit“probablywouldhavebeennothavingabigenoughsticktoforcepeopletodoit...itwasn'tuntilwehadanauditorcomeinthatwesaid,'oklookhereguysthat'snotgoodenoughorifthatiswhatwe'regoingtoget,that'sdefinitelygoingtoreflectonourauditreport.'AndsoIhavetosayforus,andIdon'twanttooverstatethisbecausewedidn'thaveaproblembefore,butinordertogetdetaileddocumentsitreallydidtaketheaudittopullthosethingsout.”Chronopolisstaffwereunable

0

1

2

3

4

5

Obstacles

Frank

34

togetdetaileddisasterplanningdocumentationfromthethreepartnerorganizationsuntiltheauditreportimposedahigherdegreeofaccountability.TheintervieweefromICPSR,theonlyorganizationinthisstudywithstaffwhocreateddetaileddisasterplanningdocumentationindependentofanaudit,focusedontheproblemsoforganizationalcooperationanddifficultycoordinatingwiththeITdepartment.SubjectEbeganwithdiscussionofthehistoricalresistancetoformaldisasterplanningactivities,“inthepastIthinkthatitwasoftenlookedatasaluxury...it’sanaturalhumanthingtonotwanttotalkaboutadisasteruntilthedisasteristhereandthenbecaughtshortbecauseyoudon’thaveanyplanninginplace.”Inordertoovercomethisresistance,itwasnecessarytogetbuy-infromseniormembersoftheorganization,memberswhowereinitiallyunwillingtodevotetheirtimetotheprocess,“partofthedifficultyofengaginginrolesandresponsibilitiesisthattheyhavetoatleaststartatthehighestlevelsoftheorganization.Theyviewitascostlytheyviewitasadistraction,butyoucan'tworkatthebottomwhenyou'redealingwithdecisionmakingandactualauthority.”StaffofICPSRalsoexperienceddifficultyin“parsingouttheITpiece...becausewhenyouhaveITasanintegralpartofyourorganizationandyourorganizationiscommittedtolifecyclemanagement,thereisthis‘now’and‘future’andthepeoplewhoaredoingthesethingsdon'toftendistinguishbetweenthehatsthattheyhave.Itwashardtogetthemtofocusonthedifferentparts...wehaveareallygoodITgroup,butit'salsoachallengefordigitalpreservation.”SubjectJfromtheMetaArchiveCooperativefocusedmoreontheproblemoftime,“itisallabouttime,ittakessomuchtimetowritedocuments,figuringoutwhatgoesinthedocumentationandgettingitbythecommittee,andthenstartingthedraftandgettingthatpastthecommitteeandgettingitapproved...Iwouldsaythattimeisthegreatestchallengeofeverythingthatwe'redoingaroundpolicycreation,includingdisasterrecoveryplanning.”Thisalsoindirectlydiscussestheproblemoforganizationalbuy-inandcooperationasitistheprocessofgettingcommitteeapprovalthatseemstotakethemosttime.SubjectGfromtheNationalLibraryofAustraliaexpressedthegreatestdegreeoffrustrationincoordinatingwiththeITsectionoftheorganization,“wehavecertainpeopleinourITsection[who]I’llsayarepreservationdeniers.”Hethengoesontoexplainthat,“sometimesourITdepartment'knowbest'-theytakethehighmoralgroundandthenwecatchthemoutandsay'hangonaminute,you'renotdoingthis,orthis'...andthenITsays'we'reworkingonit'andwefoundsomefrustrationbecausetheyareaseparatedepartment.”Whileformaldisasterplanningdocumentationisnotcurrentlyinplaceatthisorganization,thisdifficultyincoordinatingwiththeITsectioninordertodocumentandcarryoutdigitalpreservationactivitiesextendstodisasterplanningaswell.IntervieweesfromHathiTrustweretheonlytoexplicitlystatethattherewerenoproblemsgettingorganizationalbuy-in.PerSubjectC,“thereisverylittle,andmaybesafetosayno,organizationalresistanceto...makingthedataassafeandasusefulaspossible...organizationalphilosophyistomakesurethatalloftheworkthathasgoneintothisdataisnotgoingtobefornaughtwhenitallgetswipedoutinsomesortofdisaster.”However,thisisan

Frank

35

organizationthathasgonethroughanauditbuthasnotyetproducedformalizeddisasterplanningdocumentation.AsSubjectCstated,theyarestrugglingwiththechallengeofproducingdocumentationthatisbothexcellentandusefulandareworkingtoembracethephilosophyofnot“let[ting]theperfectbetheenemyofthegood.”Thiscaseseemstohighlightthevalueoftheobstaclesthattheotherorganizationshavefaced.Thefigurebelowillustratesthediscussionofobstacles,ofdocumentation,andtheoverlapofobstaclesanddocumentation.IntervieweesfromthreeofthefourrepositoriesthatshowanoverlapindiscussionofobstaclesanddocumentationhavesuccessfullycompletedcertificationwitheitherTRACorDataSealofApprovalandreportthattheyhaveformalizeddisasterplanningdocumentationinplace.Thefourth,HathTrust,hasalsosuccessfullycompletedTRACcertification,butdoesnotyethaveaformalizeddisasterplan.

Figure5:ObstaclesandDocumentation

4.5TestingthePlansAnotherthemethatemergedfromtheinterviewswasthatoftestingthedisasterplanningdocumentationthroughexercisesor“firedrills”oncetherepositoryhadaformaldisasterplaninplace.Threeoftheeightrepositoriesstudieddiscussedsomeformofexercisetotesttheplans,andoneexplainedthattheychosenottorunthesedrillsbecausetheyencounterenoughissuesonaregularbasisthattheirpoliciesandproceduresareunderconstantscrutiny.IntervieweesfromChronopolis,ICPSR,andTheMetaArchiveCooperativeeachdiscussedtheireffortstoregularlytesttheirdisasterplanningdocumentation.ForSubjectAatChronopolis,thatinvolvestestingwherein“webasicallygoandunplugsystemsandthenforceourselvestorecover.”SubjectJatTheMetaArchiveCooperativealsodiscussed“disasterplanningexercises”thatareconducted“onatwoyearbasis.”However,therepresentativefromThe

0

5

10

15

20

25

30

Obstacles

DisasterPlanDocumentation

ObstaclesandDocumentation

Frank

36

MetaArchiveCooperativestatedthat,“wehaven'trunafullscaleagain...we’vedonepiecesofitatdifferenttimes.”BothoftheseintervieweesexplainedthattheyfirstranthetestingexercisesforthepurposesoftheTRACaudit.ThestaffoftheMetaArchiveCooperativehavenotrunanotherfulldrillsincethattimeandChronopolishasnotyethadtheopportunitytorunanotherdrillsincetheirauditwasconductedsorecently.Itwillbeinterestingtoseeiftheycontinuewitharegulartestingschedule.SubjectEfromICPSRdescribedanelaboratetabletopexercisescenarioinwhich,“somebodywhoisinwebservicesinourorganization[who]isreallyinvolvedingamingcreatedthiswholescenarioforusandhandedoutcardsofwhoareyouandwhat'syourpartofthescenario.”Fromthisinitialexercisetheywereabletoidentifyweaknessesintheirplans,“somebodyhadtogoawayandfillinsomegaps.”Asaresult,theorganizationhas“committedtodoingthatatleastonceayear.”SubjectFatMATRIX,whoisinanITrolewithintheorganization,statedthat“wehaveenoughissuesfromtimetotimethatwedon'trunfiredrillsanymorebecauseithappensandalotofithastodowithupgradingourinfrastructure...it'sjustreallywhenthere'sunderlyingcoresystemsthatgetupgradedthat'swhentheissuescouldoccur.”Inotherwords,theyspendenoughtimefixingproblemsthattheyknowhowtorecover.ThisresponseechoesotherfindingsregardingtherelationshipbetweenpreservationandIT.Namely,thatitisthoseindividualswhoserolesfallunderpreservationfunctionswhodrivethedevelopmentofformalizeddisasterresponseandrecoveryplanningdocumentation,andthatindividualswhoserolesfallunderITfunctionsaremoreresistanttothesetypesofactivities.

4.6AccesstoDisasterPlanDocumentationMostsubjectsinthestudyexpressedadesiretomaketheirorganization’sdisasterplanningdocumentationavailabletothepublicatsomepointinthefuture.Todate,ICPSRandHathiTrustaretheonlyrepositorieswhosestaffhavemadealloftheirdocumentationpubliclyavailableviatheirwebsiteorviaemailrequest.Intermsofinternalaccess,mostrepositorieshavedisasterplanningdocumentationthatisavailabletoallemployeesoftheorganizationviasomesortofstaffwikiorsharedstoragespace.SubjectEfromICPSRindicatedthatapproximately80%oftheirdisasterplanningdocumentationisavailableontheirwebsite,andthat“wetrytobereallytransparentandifpeopleaskusreallywewouldprovideanythingelsethatisn’tonline.”SubjectEdid,infact,shareadditionalinformationviaemailforthepurposesofthisstudy.ThestaffofHathiTrusthasplacedthefoundationdocumentsfortheirdisasterplanningeffortsonlineviatheorganization’swebsite,andoneoftheintervieweesstatedthattheyintendtocontinuemakingadditionaldocumentationavailable,includingthecompleteddisasterplanningdocumentsoncetheyarecompleted.Theirpositiononthematter,accordingtoSubjectD,isthat“it’sonlytoourbenefitIthinktomakeasmuchofthatavailabletopeoplesotheyknowwhatwe’redoing...ourorientationisopen.”Forthepurposesofthisreport,theywereabletoshareadditionaldocumentationandinformationviaemail.

Frank

37

SubjectAfromChronopolisexplainedthat,“wedointendtomakeitaspublicaspossible,”butalsostated,“we'regoingtohaveacouplepointswherewhateverwemakepublicwe'regoingtohavetodoitatsuchahighlevelthatitdoesn'trevealcompromisingdetails...andnotputanythingthere.”Theorganizationisstrugglingwithwantingtomaketheirinformationavailablebutneedingtorespecttheprivacyandsecurityneedsoftheirpartnerorganizations.ThesearethesamepartnerorganizationsthatwerereluctanttoevenprovidedisasterresponseandrecoveryinformationtoChronopolistobeginwith,asdiscussedabove.Theydid,however,indicatethattheentirestaffandallpartnerorganizationshaveaccesstothisinformationinternally.StaffatallmemberinstitutionsformingtheMetaArchiveCooperativehaveaccesstotheirdisasterplanningdocumentationforallstaff.PerSubjectJ,“that'sallonourwiki,whichispartofourcoreinfrastructurethatishousedinthecloudsoitisaccessibletoallofourmemberinstitutionsbothformaintenanceofthedocumentandthenalsoforviewingthedocument.Itisapasswordprotectedarea.”Accesstothisdocumentationisrestrictedto“membersorpeoplewhohaveformalizedrelationships”withtheorganization.ThestaffoftheMetaArchiveCooperativeisintheprocessoftryingtomakethisdocumentationavailableviatheirwebsite,andSubjectJstatedthat“forthedisasterrecoverythusfaritdoesnotlooklikewe'llneedtorestricttheinformationtherein,there'snothingtherethat...wouldcompromiseMetaArchive.”Portico,MATRIX,andtheInternetArchiveallhavedocumentationthatisavailableinternallytostaffbutdonotplantomakethatinformationavailabletothepublic.SubjectFstatedthatatMATRIX,“allstudentsandstaff,anyonethatbasicallyhasaMATRIXloginwhichiseveryonethatworkshere[hasaccesstothedocumentationonthestaffwiki].”SubjectHfromPorticomaintainsthat,“theITspecificdisasterrecoverydocuments,whichareprettyspecificandnotthetypeofthingyoucanmakepubliclyavailable,aremaintainedinternallyonly.”AndwhileSubjectIfromtheInternetArchivestatedthat,“everyone[intheorganization]hasaccesstoit[thechecklist]”andthat“oneofthefoundingprinciplesoftheInternetArchiveisuniversalaccesstoallknowledge...thattricklesdowntoallofourdocumentation,”healsostatedthathewouldnotbeabletosharetheirdisasterresponseandrecoverychecklistwithanyoneoutsideoftheorganization.SubjectGfromtheNationalLibraryofAustraliawastheonlypersonwhocommunicatedthatthedisasterplanningdocumentationwasnotavailablepublicly,norwasitavailableinternally.Rather,thedocumentsresidewiththeITdepartmentandhavenotbeensharedormadewidelyavailable,“Ihaven’t[hadcontactwiththeplan]...ournewsystemshopefullywillmakeitavailabletoeveryone.”ThisechoesthesentimentexpressedbySubjectGearlierintheinterview,anddiscussedelsewhereinthispaper,thatoneoftheprimaryobstaclestodisasterplanninginthisorganizationiscommunicationwiththeITdepartment.AsFigure6belowshows,theintervieweefromtherepositorywiththehighestlevelofaccesstodocumentation,ICPSR,alsodiscussedaccesstodocumentationmorefrequently.Staffatthe

Frank

38

repositorywiththelowestlevelofaccesstodocumentation,theNationalLibraryofAustralia,mentionedaccesstodocumentationtheleast.OtherorganizationswhoseintervieweesalsoreportlowlevelsofaccesstodocumentationsuchasMATRIXandtheInternetArchivealsoshowlowratesofdiscussionofaccesstodocumentation.Intervieweesfromrepositoriesthatprovideaccesstosome,butnotall,oftheirdocumentationgenerallyfallsomewhereinthemiddle.

Figure6:Access

5.DiscussionInthissection,Iwilldiscussthethreemainfindingsofthisstudyingreaterdetail.Thethreemajorfindingsofthisstudyarethat:

1. Formostorganizations,theprocessofgoingthroughanauditforcertificationasatrustedrepositoryprovidedtheimpetusforthecreationofformalizeddisasterplanningdocumentation.

2. Despitethedesireforopennessmostrepositoriesstrugglewithmakingtheirdisasterplanningdocumentationpubliclyavailable.

3. ThesinglegreatestobstacletodisasterplanningactivitiesatallstagesoftheprocessiscoordinationbetweentheITandpreservationfunctions.

Iwillalsoaddresspossiblelimitationsofthestudy,anddiscussdirectionsforfutureresearch.OnepatternIobservedthroughouttheinterviewprocesswasthatorganizationssuchasChronopolis,ICPSR,Portico,andTheMetaArchiveCooperative,thathavebeenthroughsomesortofauditprocess,weremorelikelytohavecompletedisasterplanningdocumentationinplace.Theseorganizationsdiscussedtherolethattheauditplayedinprovidingmotivationto

012345678

Access

Frank

39

completethisdocumentation,anddiscussedthechallengesthathadpreventedthemfromcompletingthisdocumentationpreviously.Centraltothiswastheideathatuntiltheorganizationwasprovidedwithasuitablyattractiveincentive(i.e.certification),itwasdifficultorimpossibletoconvinceotherdepartmentssuchasITandAdministrationtospendtimedocumentingpoliciesandproceduresthatwereeitherformallydocumentedelsewhereortacitlyunderstood.Repositoriesstrugglewiththedecisiontomakedisasterplanningdocumentationavailabletothegeneralpublic.Iexpectedrepositoriesthathadbeenthroughanauditforcertificationofsomesortwouldbewillingtomakeatleastpartsoftheirdisasterplanningdocumentationpubliclyavailable.AsoneofthevalueprinciplesofTRACistransparency,andthistransparencehelpspeopletotrustrepositories,itseemednaturalthattheywouldthenmakethatsameinformation,orsomeportionthereof,availabletothepublic.AsIdiscovered,thisisgenerallynotthecase.Theavailabilityofdocumentationregardingdisasterplanningactivitiesvarieswidelyamongrepositoriesandrunsthefullspectrumfromfullyavailabletocompletelyrestricted.Finally,Ifoundthatthesinglegreatestobstacledodisasterplanningactivitiesatallstagesoftheprocessiscoordination,orlackofcoordination,betweentheITandpreservationfunctionswithinanorganization.SubjectsinpreservationandadministrationrolesexpressedfrustrationwiththelackofcommunicationandcooperationfromtheITdepartmentsintheirorganizations.SubjectsinITfunctionsexpressedabeliefthatformaldisasterplanningactivitieswereunnecessaryandapooruseoftimeandresourcesfortheorganization.Thisisrelatedtothefirstfindinginthatitseemstobethecasethatorganizationsarebestabletoovercomethisobstaclearethosethatcandemonstrateaconcretebenefit,suchacertification,thatwillresultfromtheproductionofformaldisasterplanningdocumentation.Thesefindingssuggestthatoneoftheprimarybenefitsachievingtrusteddigitalrepositorystatus,inadditiontothecertificationitself,isthefactthatitprovidesanincentivefortheentireorganizationtocreateaccurate,up-to-date,thoroughdocumentationofpoliciesandprocedures.Fororganizationsthatalreadyhavedocumentationinplace,suchasICPSRandPortico,theauditprovidestheorganizationwithanopportunitytoimproveandupdatetheirdocumentation.ThesefindingsalsosuggestthatagreaterdegreeofcommunicationandcooperationisneededbetweenpreservationandITfunctionswithindigitalrepositories.AconsistentpatternintheinterviewswasthedifficultyinworkingwithIT,andtheresistanceofthatgrouptoparticipateinformaldisasterplanningdocumentationefforts.Conversely,thisproblemcanbeseenasashortcomingonthepartofdigitalpreservationpolicymakers.Perhapsanopportunityforeducationandbettercommunicationexistsbetweenthedifferentfunctions.WhiletheITfunctionseemstoalmostuniversallyhavebeenanobstacletodisasterplanningeffortsintherepositoriesinthisstudy,intervieweesalsostatedthatthisseemstobeacaseofindividualsintheITrolenothavingthesameunderstandingofandappreciationfordisasterplanning.Anopportunityexistsforthoseinthefieldofdigitalpreservationtofindwaysofcommunicating

Frank

40

withthoseinIT,inordertoimprovecollaborationandcoordinationthroughouttheorganization.Theinitialresearchquestionforthisstudyfocusedoninvestigatinghowrepositoriesareengagingindisasterplanningactivities.Afterexaminingthepracticesofseveralwell-respecteddigitalrepositories,ithasbecomeclearthatoneofthereasonsthatsofewstudieshavebeenconductedinthisareaisthatdigitalrepositories,untilrecently,didnottohavedocumentedtheirdisasterplanningeffortsatall.Ithasalsobecomeclearthatitisnotpossibletogainafullunderstandingofthedisasterplanningeffortsofanorganizationifthoseeffortsarenotcodifiedandmadeavailableforreview.Thefactthatonlytwooftheeightrepositorieswereableorwillingtomaketheirdisasterplanningdocumentationpubliclyavailablewasamajorlimitationforthisstudy.Additionally,thislackofmodelsmaybehamperingdisasterplanningeffortsinthecommunityThereareseveralotherfactorsthatcouldbeconsideredweaknessesorlimitationsofthisstudy.First,thesmallpopulationsizemakesitdifficulttodrawconclusionsthatcouldbegeneralizedtoalargerpopulation.Thislimitedscopeispartlyaresultofthesmallnumberofrepositoriesthatareengagingintrustedrepositoryauditsandpartlyaresultofthelimitedtimeframeinwhichthisstudywasconducted.Additionalstudiesinthisareamaywanttoconsiderincludingagreaternumberofrepositories.Second,speakingtoonlyoneortwoindividualsateachrepositorydoesnotprovideacompletepictureoftheentirelifecycleofthedisasterplanningprocess.Inordertogainafull,completeunderstandingoftheactivitiesatanygivenrepository,itwouldbeidealtointerviewseveralindividualsfromdifferentdepartmentsorfunctionswithinarepository.Forastudysuchasthis,interviewingindividualsindigitalpreservation,administration,andITwouldprovideawell-roundedviewofdisasterplanningactivities.Third,futureresearchwoulddowelltoeitherfocusononetypeofrepository,ortostudyabroadspectrumofrepositorytypes.Ofthefinaleightrepositoriesincludedinthisstudy,oneisanationallibrary,oneisaninstitutionalrepository,andtherestarenonprofitorganizationswithvaryingdegreesofaffiliationwithacademicinstitutions.Whilethisfinalselectionwasaresultofavailabilityandconvenience,theresultsmayhavebeenquitedifferentiftheresearchfocusedonlyonnationaldigitalrepositories,oroninstitutionalrepositories.Additionally,allbutoneoftherepositoriesinthisstudyarebasedintheUnitedStates.Withalargerandmoregeographicallydiversepopulation,thestudymighthavebeenabletoexamineregionalornationaltrendsinordertounderstandhownationalityand/orlocationaffectdisasterplanningactivities.

6.ConclusionThisstudyfoundthatwhilerepositoriesareengagingindisasterplanningactivities,theyaredoingsolargelyasameanstoobtaintrusteddigitalrepositorystatus.Furthermore,repositoriesarereluctantorunwillingtosharetheirdisasterplanningdocumentation.This

Frank

41

suggeststhatwhileoneofthekeyelementsofcertificationprogramsfordigitalrepositoriesisthecreationofformalizeddocumentationofpoliciesandprocedures,thesearenotbenefittingthecommunityasmuchastheycould.SincetransparencyisacoretenetofTRAC,auditorsshouldinsistthattrusteddigitalrepositoriessharedisasterplanningdocumentationandmakenon-sensitivepoliciesandproceduresavailabletothepublicinordertomeetthecriteriafortrustedrepositorystatus,ortoincludetherepository’sdocumentationinthefinalauditreportdemonstratingthattheyhavemetthecriteriaforcertification.Noneoftherepositoriesincludedinthisstudyhavehadtheopportunitytousetheirdisasterplanningdocumentation.Whileonehopesthattheseorganizationswillneverhavetheneedfortheiruse,anopportunityforfutureresearchexistsintheimplementationanduseofthesedocuments.Inherarticleondisasterpreparedness,Schmidtobservesthat,“givenenoughtime,thelikelihoodofamajordisasterataninstitutionbecomesanearcertainty”(Schmidt,2010).

7.AcknowledgementsIwouldliketothankDr.ElizabethYakel,Dr.PaulConway,andShannonZacharyfortheirsupportandguidancethroughoutthecourseofthisproject.

Frank

42

ReferencesAikin,J.(2007).PreparingforaNationalEmergency:TheCommitteeonConservationof

CulturalResources,1939-1944.TheLibraryQuarterly,77(3),257.

Altman,M.,Adams,M.,Crabtree,J.,Donakowski,D.,Maynard,M.,Pienta,A.,&Young,C.(2009).DigitalPreservationthroughArchivalCollaboration:TheDataPreservationAlliancefortheSocialSciences.TheAmericanArchivist,72(1),170-184.

Anderson,C.(2005).DigitalPreservation:WillYourFilesStandtheTestofTime?LibraryHiTechNews,22(6),9-10.

Anderson,C.(2008).TheEndofTheory:TheDataDelugeMakestheScientificMethodObsolete.WiredMagazine,16(7).RetrievedfromWiredMagazinewebsite:http://www.wired.com/science/discoveries/magazine/16-07/pb_theory

Andrew,G.(2008).BusinessContinuity:BestPractices.eWeek,25(33),32.

Babbie,E.R.(2010).ThePracticeofSocialResearch.Belmont,CA:WadsworthCengage.

Ball,A.(2010).ReviewoftheStateoftheArtoftheDigitalCurationofResearchData:UniversityofBath.

Barateiro,J.A.,Goncalo;Freitas,Filipe;Borbinha,Jose.(2010).DesigningDigitalPreservationSolutions:ARiskManagement-BasedApproach.TheInternationalJournalofDigitalCuration,5(1),4-17.

Battersby,R.(2005).Recoveringfromdisaster:thelossofEdinburgh'sAILibrary.Library+InformationUpdate,4(3),36-36-38.

Beagrie,N.,Chruszcz,J.,&Lavoie,B.(2008).KeepingResearchDataSafe:JISC.

Berman,F.(2008).Gotdata?:aguidetodatapreservationintheinformationage.CommunicationsoftheACM,51(12),50-56.

Berman,F.,Lavoie,B.,Ayris,P.,Choudhury,G.S.,Cohen,E.,Courant,P.N.,...VanCamp,A.(2010).SustainableEconomicsforaDigitalPlanet:EnsuringLong-TermAccesstoDigitalInformation;BlueRibbonTaskForceonSustainableDigitalPreservationandAccessFinalReport.

BestPracticesinDisasterRecoveryBusinessContinuityPlanning.(2008).Baseline.

Brennan,C.,&O'Hara,E.(2002).Murphywasalibrarian:acasestudyinhownottohandleasystemscrash.ComputersinLibraries,22(3),10-10-12.

Frank

43

CenterforResearchLibraries.(2007).TenPrinciples.RetrievedApril2012,fromhttp://www.crl.edu/archiving-preservation/digital-archives/metrics-assessing-and-certifying/core-re

CenterforResearchLibraries.(2010).CRLCertificationReportonPorticoAuditFindings.

CenterforResearchLibraries.(2011).CRLCertificationReportontheHathiTrustDigitalRepository.

CenterforResearchLibraries.(2012a).CRLCertificationReportonChronopolis.

CenterforResearchLibraries.(2012b).ReportsonDigitalArchivesandRepositories.RetrievedonApril18,2012fromhttp://www.crl.edu/archiving-preservation/digital-archives/digital-archive-reports

Cervone,H.F.(2006).Disasterrecoveryandcontinuityplanningfordigitallibrarysystems.OCLCSystems&Services:Internationaldigitallibraryperspectives,22(3),173.

Constantinescu,C.,Parulkar,I.,Harper,R.,&Michalak,S.(2008).SilentDataCorruption-Mythorreality?InternationalConferenceonDependableSystemsandNetworks.

Cousins,T.J.(2007).DevisingPost-DisasterContinuityPlansthatMeetActualRecoveryNeeds.IEEETechnologyandSocietyMagazine,26(3),13.

DataIntensiveCyberEnvironmentsGroup.(2008).iRODS:integratedRuleOrientedDataSystemWhitePaper:UniversityofNorthCarolinaatChapelHill,UniversityofCaliforniaatSanDiego.

DataSealofApproval.(2012).AboutDataSealofApproval.RetrievedApril18,2012,fromhttp://datasealofapproval.org/

EducopiaInstitute.(2012).TheMetaArchiveCooperative.RetrievedApril18,2012,fromhttp://www.metaarchive.org/

Fletcher,A.M.(2006).NoPointofReference:AHurricaneofMedicalInformationNeeds.JournalofHospitalLibrarianship,6(2),1-14.

Gantz,J.F.,Chute,C.,Manfrediz,A.,Minton,S.,Reinsel,D.,Schlichting,W.,&Toncheva,A.(2011).TheDiverseandExplodingDigitalUniverse:AnUpdatedForecastofWorldwideInformationGrowthThrough2011.Framingham,MA.

Garrett,J.,Waters,D.J.(1996).PreservingDigitalInformation.ReportoftheTaskForceonArchivingofDigitalInformation.[S.l.]:DistributedbyERICClearinghouse.

HathiTrust.(2012a).AboutHathiTrust.RetrievedApril18,2012,fromhttp://www.hathitrust.org/about

Frank

44

HathiTrust.(2012b).HathiTrustExectuiveCommittee.Retrieved2012,fromhttp://www.hathitrust.org/xcom

Heiser,J.(2011).BestPracticesforRecoveringCriticalDataFromDamagedHardDrivesandOtherPhysicalMedia:GartnerResearch.

Hey,T.T.,Anne.(2003).Thedatadeluge:ane-Scienceperspective.InF.H.Berman,A.;Fox,G.(Ed.),GridComputing-MakingtheGlobalInfrastructureaReality(pp.809-824):JohnWiley&Sons,Ltd.

Hitchcock,S.,Brody,T.,Hey,J.M.N.,&Carr,L.(2007).DigitalPreservationServiceProviderModelsforInstitutionalRepositories:TowardsDistributedServices.D-LibMagazine,13(5/6).

Holsti,O.R.(1969).Contentanalysisforthesocialsciencesandhumanities.Reading,Mass.:Addison-WesleyPub.Co.

IInstituteofRiskManagement,AssociationofInsuranceandRiskManagers&PublicRiskManagementAssociation.(2002).ARiskManagementStandard(p.14):InstituteofRiskManagement,AssociationofInsuranceandRiskManagers&PublicRiskManagementAssociation.

IInter-UniversityConsortiumforPoliticalandSocialResearch.(2011,2011).AboutICPSR.RetrievedApril18,2012,2012,fromhttp://www.icpsr.umich.edu/icpsrweb/ICPSR/org/index.jsp

Innocenti,P.,&Vullo,G.(2009).AssessingthePreservationofInstitutionalRepositorieswithDRAMBORA:CaseStudiesfromtheUniversityofGlasgow.BollettinoAIB,49(2),139-158.

Inter-UniversityConsortiumforPoliticalandSocialResearch.(2011).ICPSRTimeline.RetrievedApril18,2012,fromhttp://www.icpsr.umich.edu/icpsrweb/ICPSR/org/timeline.jsp

InternationalOrganizationforStandardization.(2012).Spacedataandinformationtransfersystems—Auditandcertificationoftrustworthydigitalrepositories(ISO16363).Switzerland:InternationalOrganizationforStandardization.

InternetArchive.(2012).AboutTheInternetArchive.RetrievedApril18,2012,fromhttp://archive.org/about/about.php

Maniatis,P.,Roussopoulos,M.,Giuli,T.J.,Rosenthal,D.S.H.,&Baker,M.(2005).TheLOCKSSpeer-to-peerdigitalpreservationsystem.ACMTransactionsonComputingSystems,23(1),2-50.

MATRIX.(2012).AboutMATRIX.RetrievedApril18,2012,fromhttp://www2.matrix.msu.edu/about/

Frank

45

McDonald,R.H.,&Walters,T.O.(2010).RestoringTrustRelationshipswithintheFrameworkofCollaborative.JournalofDigitalInformation,11(1).

McHugh,A.R.,Seamus;Innocenti,Perla;Ruusalepp,Raivo;Hofman,Hans.(2008).BringngSelf-assessmentHome:RepositoryProfilingandKeyLinesofEnquirywithinDRAMBORA.TheInternationalJournalofDigitalCuration,3(2),130-142.

McKnight,M.(2006).HealthSciencesLibrarians'ReferenceServicesDuringaDisaster.MedicalReferenceServicesQuarterly,25(3),1.

Minor,D.Sutton,D.;Kozbial,A.;Westbrook,B.;Burek,M.;Smorul,M..(2010).ChronopolisDigitalPreservationNetwork.TheInternationalJournalofDigitalCuration,5(1),119-133.

Moore,R.(2008).TowardsaTheoryofDigitalPreservation.TheInternationalJournalofDigitalCuration,3(1),63-75.

Muir,A.,&Shenton,S.(2002).IftheWorstHappens:Theuseandeffectivenessofdisasterplansinlibrariesandarchives.LibraryManagement,23(3),115-123.

Myles,B.(2000).Theimpactofalibraryfloodoncomputeroperations.ComputersinLibraries,20(1),44-44-46.

NationalLibraryofAustralia.(2012a).DigitalPreservation.RetrievedApril18,2012,fromhttp://www.nla.gov.au/preserve/digipres/

NationalLibraryofAustralia.(2012b).DigitalPreservationPolicy.RetrievedApril18,2012,fromhttp://www.nla.gov.au/policy-and-planning/digital-preservation-policy

NationalLibraryofAustralia.(2012c).DigitalPreservationDirectionsStatement2008to2012.RetrievedfromApril18,2012,fromhttp://www.nla.gov.au/digital-preservation-directions-statement-2008-to-2012

Nollau,B.(2009).DisasterRecoveryandBusinessContinuity.JournalofGXPCompliance,13(3),51.

Patel,M.C.,Simon.(2007).AStudyofCurationandPreservationIssuesintheeCrystalsDataReositoryandProposedFederation.eBank-UKPhase3:WP4,1-34.

Patkus,B.L.,&Motylewski,K.(1993).DisasterPlanning.PreservationLeaflets.RetrievedfromNortheastDocumentConservationCenterwebsite:http://www.nedcc.org/resources/leaflets/3Emergency_Management/03DisasterPlanning.php

Pennock,M.(2007).DigitalCuration:ALife-CycleApproachtoManagingandPreservingUsableDigitalInformation.Library&Archives(1),3.

Frank

46

Portico.(2012).AboutPortico.RetrievedApril18,2012,fromhttp://www.portico.org/digital-preservation/about-us

Rajasekar,A.(2010).iRODSprimerintegratedrule-orienteddatasystem.SanRafael,Calif.(1537FourthStreet,SanRafael,CA94901USA):Morgan&ClaypoolPublishers.

Robson,C.(1993).Realworldresearch:aresourceforsocialscientistsandpractitioner-researchers.Oxford,UK;Cambridge,Mass.,USA:Blackwell.

Rombouts,J.,&Princic,A.(2010).Buildinga'datarepository'forheterogenoustechnicalresearchcommunitiesthroughcollaborations.PaperpresentedattheInternationalAssociationofScientificandTechnologicalUniversityLibraries,31stAnnualConference.http://docs.lib.purdue.edu/iatul2010/conf/day2/10

Ross,S.M.,Andrew.(2006).PreservationPressurePoints:EvaluatingDiverseEvidenceforRiskManagement.PaperpresentedattheiPRES2006,NewYork,NY.

SanDiegoSupercomputerCenter.(2011b).Infrastructure.RetrievedApril18,2012,fromhttp://chronopolis.sdsc.edu/infrastructure/index.html

SanDiegoSupercomputerCenter.(2011a).AboutChronopolis.RetrievedApril18,2012,fromhttp://chronopolis.sdsc.edu/about/index.html

Schmidt,G.(2010).Web2.0forDisasterResponseandRecovery.JournalofWebLibrarianship,4(4),413-426.

Schroeder,B.G.,GarthA.(2007).DIskfailuresintherealworld:WhatdoesanMTTFof1,000,000hoursmeantoyou?PaperpresentedattheFAST'07:5thUSENIXConferenceonFileandStorageTechnologies,SanJose,CA.

Schultz,M.(2010).MetaArchiveCooperativeTRACAuditChecklist.Atlanta,GA:EducopiaInstitute.

Sesink,L.;vanHorik,R.;Harmsen,H.(2010).DataSealofApproval:QualityGuidelinesforDigitalResearchData(2ndEditioned.).TheHague:DataArchivingandNetworkedServices(DANS).

Skinner,K.W.,Tyler.(2011).NewRolesforNewTimes:DigitalCurationforPreservation,PublishedbyARL.Washington,D.C:AssociationofResearchLibraries.

Tennant,R.(2001).Copingwithdisasters.LibraryJournal,126(19),26.

TheDataSealofApprovalBoard.(2011).ImplementationoftheDataSealofApproval:Inter-UniversityConsortiumforPoliticalandSocialResearch:DataSealofApproval.

Frank

47

Wheatman,V.(2001).Aftermath:DisasterRecoveryAftermath.Stamford,CT:GartnerResearch.

Wheatman,V.S.,Donna;Witty,RobertaJ.(2001).Aftermath:BusinessContinuityPlanningAftermath.Stamford,CT:GartnerResearch.

Wildemuth,B.M.(2009).Applicationsofsocialresearchmethodstoquestionsininformationandlibraryscience.Westport,Conn.:LibrariesUnlimited.

Wong,Y.L.,&Green,R.(2006).DisasterPlanninginLibraries.JournalofAccessServices,4(3/4),71-82.

Frank

48

AppendixA:ConsenttoParticipateinaResearchStudyInterviewInvitationtoParticipateinaResearchStudy

You are invited to be a part of a study exploring how digital repositories are engaging in disaster planning activities.

Description of Subject Involvement.

Interview questions will concern the nature of your organization’s attitude toward disaster planning activities, the process by which your organization’s disaster plan was created, specific elements of that plan, and how your organization has implemented/used the plan. The interview will last approximately one hour.

Benefits.

Although you may not directly benefit from this study, others may benefit because of the comparison of digital disaster planning processes and documents.

Risks and Discomforts.

There are minimal risks associated with this study. Due to the fact that so few people are involved in digital disaster planning, the researchers are aware that you may not be able to remain completely anonymous.

Confidentiality.

We plan to publish the results of this study. If we quote from you, you will be given the opportunity to review and approve the use of any quotations that could be attributed to you. There are some reasons why people other than the researchers may need to see information you provided as part of the study. This includes organizations responsible for making sure the research is done safely and properly, including the University of Michigan and government offices.

All research records will be kept confidential to the extent provided by federal, state, and local laws. Data from the study will be kept in a secure location. These data may be used again in future research studies. Your real name will not appear in notes, transcripts or audio file names. Photographs, with permission, may be used in publications. You will be given a copy of this document for your records and one copy will be kept with the research records.

Consent.

Interviewees will get no direct benefits from this research. With your consent, this session will be audio taped.

Your participation in this study is voluntary. Subsequent to your consent, you may refuse to answer specific questions, withdraw from the study at any time, or ask that information be removed from our data set. You may also ask questions concerning the study before, during, or after the study.

Frank

49

Contact Information.

If you have questions about this research you may contact Rebecca D. Frank, University of Michigan, School of Information, (248) 854-0319, frankrd@umich.edu or Elizabeth Yakel, University of Michigan, School of Information, (734) 763-3569, yakel@umich.edu, fax (734) 615 - 3587.

This study has been reviewed by the Institutional Review Board at the University of Michigan and granted “Not Regulated” status. If you have any questions about your rights as a research participant, or wish to obtain information, ask questions or discuss any concerns about this study with someone other than the researcher(s), please contact the University of Michigan Institutional Review Board Health Sciences and Behavioral Sciences, (734) 936-0933, or toll free (866) 936-0933 540 E. Liberty St., Suite 202 Ann Arbor, MI 48104-2210, irbhsbs@umich.edu. (For international calls include the US Calling Code 1 and the exit number for the country of origin XXX+1+734-936-0933.)

Consent.

By signing this document, you are agreeing to be in the study. You will be given a copy of this document for your records and one copy will be kept with the study records. Be sure that questions you have about the study have been answered and that you understand what you are being asked to do. You may contact the researcher if you think of a question later. I agree to participate in the study. _____________________________________ Printed Name _____________________________________ _____________________________ Signature Date

Frank

50

AppendixB:QuestionsforSemi-StructuredInterview

1. What is the nature of your organization’s attitude toward disaster planning for digital collections? What types of activities has your organization undertaken regarding disaster preparedness/business continuity/continuity of service/etc.?

2. Does your organization have a formal disaster plan? a. For how long have you had a formal disaster plan? b. What was the impetus for the creation of this plan? c. Who is responsible for the maintenance of the plan?

3. If not, why? What documentation do you use in lieu of a formal disaster plan?

a. Does your organization intend to create a disaster plan? b. What has prevented you from developing a plan thus far?

4. Please describe the process by which your disaster plan was created.

a. Please discuss any of the following resources used in the creation of your organization’s plan: standards, other organizations, other documentation and/or models, original research, professional organizations.

b. What departments/individuals were involved in the creation of your disaster plan? Administration, digital preservation, systems, facilities?

c. Did you consult with any external organizations? Backup sites, vendor agreements, power company, internet service providers, etc.?

d. How long did it take? Were any parts of the process particularly time-consuming or challenging?

e. Did the plan require any formal approval in your organization? Please discuss.

5. Discussion of specific elements of the plan. (Questions will vary depending on the organization’s plan/documentation).

6. Please describe the process by which your disaster plan is updated. Do you have a

formal review schedule to update the plan? Who is involved in this process? a. Who is responsible for maintenance of the plan? For updating the plan?

7. Is your organization TRAC Certified? Data Seal of Approval? Any other

certifications? a. What impact has this certification had on your organization’s disaster planning

efforts? b. Did you have a disaster plan prior to certification?

8. What obstacles did you encounter during the development of your disaster plan?

How did you overcome these obstacles/difficulties?

9. Who has access to the disaster plan? a. Are different versions available (i.e. a version for the general public vs. a

restricted-access version for staff)? b. How is this plan made available (online, hard copy, etc.)? c. How are staff members (and possibly others) made aware of the plan?

Frank

51

d. Does your organization conduct training regarding the plan?

10. Have you had occasion to use the plan? a. Please describe the event(s). b. What elements of the plan were most useful? Least useful? c. How did your organization access the plan? Which method of access was most

useful? d. How did this event affect your organization’s view of the plan? e. Did this event prompt changes to the plan?

11. Please discuss how disaster planning fits into your organization’s budget.