(r,QI)-Transform:Reversible Data Anonymity Based on Numeric Type of Data in Outsourced Database

(整期优先)网络出版时间:2017-03-13
/ 1
Anoutsourcedatabaseisadatabaseserviceprovidedbycloudcomputingcompanies.Usingtheoutsourcedatabasecanreducethehardwareandsoftware'scostandalsogetmoreefficientandreliabledataprocessingcapacity.However,theoutsourcedatabasestillhassomechallenges.Iftheserviceproviderdoesnothavesufficientconfidence,thereisthepossibilityofdataleakage.Thedatamayhasuser'sprivacy,sodataleakagemaycausedataprivacyleak.Basedonthisfactor,toprotecttheprivacyofdataintheoutsourcedatabasebecomesveryimportant.Inthepast,scholarshaveproposedk-anonymitytoprotectdataprivacyinthedatabase.Itletsdatabecomeanonymoustoavoiddataprivacyleak.Butk-anonymityhassomeproblems,itisirreversible,andeasiertobeattackedbyhomogeneityattackandbackgroundknowledgeattack.Lateron,scholarshaveproposedsomestudiestosolvehomogeneityattackandbackgroundknowledgeattack.Buttheirstudiesstillcannotrecoverbacktotheoriginaldata.Inthispaper,weproposeadataanonymitymethod.Itcanbereversibleandalsopreventthosetwoattacks.Ourstudyisbasedontheproposedr-transform.Itcanbeusedonthenumerictypeofattributesintheoutsourcedatabase.Intheexperiment,wediscussedthetimerequiredtoanonymizeandrecoverdata.Furthermore,weinvestigatedthedefenseagainsthomogeneousattackandbackgroundknowledgeattack.Attheend,wesummarizedtheproposedmethodandfutureresearches.