微电子技术论文7435271.pdfVIP

Thisarticlehasbeenacceptedforpublicationinafutureissueofthisjournal,buthasnotbeenfullyedited.Contentmaychangepriortofinalpublication.Citationinformation:DOI10.1109/TC.2016.2543220,IEEE

TransactionsonComputers

1

APrivacy-PreservingOutsourcedFunctional

ComputationFrameworkAcrossLarge-Scale

MultipleEncryptedDomains

XimengLiu,Member,IEEE,BaodongQin,RobertH.Deng,Fellow,IEEE,RongxingLu,SeniorMember,IEEE,

andJianfengMa,Member,IEEE,

Abstract—Inthispaper,weproposeaframeworkforprivacy-Datastoredinthecloudaremaintainedbymanyservice

preservingoutsourcedfunctionalcomputationacrosslarge-scaleprovidersandusersfrequentlyusemultiplesourcesofin-

multipleencrypteddomains,whichwerefertoasPOFD.Withformationintheirdecisionmakingprocess.Forexample,let

POFD,ausercanobtaintheoutputofafunctioncomputed

overencrypteddatafrommultipleencrypteddomainswhileusconsiderastockbrokeragefirmwhichselectspreferred

protectingtheprivacyofthefunctionitself,itsinputanditsstocksorbondsbasedonaproprietarypredictionfunction

output.Specifically,weintroducetwonotionsofPOFD,thebasicwithnumerousdimensionalinputvalues.Thetimeconsuming

POFDanditsenh