多級安全數(shù)據(jù)庫系統(tǒng)函數(shù)依賴和多值依賴推理研究.pdf

1、華中科技大學(xué)碩士學(xué)位論文多級安全數(shù)據(jù)庫系統(tǒng)函數(shù)依賴和多值依賴推理研究姓名：佘志東申請學(xué)位級別：碩士專業(yè)：計算機(jī)軟件與理論指導(dǎo)教師：吳恒山2003.5.10華中科技大學(xué)碩士學(xué)位論文AbstractWiththeproliferationofcomputerandnetworktechnology，databasessecurityhasbecomeanimportantresearchareaofinformationsecurityTh

2、eresearchofinferenceproblemsindatabasesisakeyfieldofresearchingmultilevelsecuredatabasesystem(MLSDBS)Accordingtothecurrentresearchstatusofinferenceproblems，wewillinvestigatethefunctionaldependencies(FD)inferenceandthemul

3、tivalueddependencies(MVD)inferenceinMLSDBSdeeplyForFDinferenceinMLSDBS，wefirstintroducethereasonofFDinferenceoccurrenceundertheattributeclassificationschemaThenWeproposeasufficientandnecessaryconditionforFD—compromisesde

4、tectionandelimination，anddesignanexac。talgorithmtOadjustattributelevelsSOthatFD—compromisesareavoided，since，thenumberofFDsandattributesinvolvedinthealgorithmissmallThisalgorithmwillincurminimuminformationlossFinallyweana

5、lysethecorrectnessandtimecomplexityofthisalgorithmTocomparewithotheralgorithmsproposedbyotherscholars，thisalgorithmisnotonlycorrectbutwillsavemorecomputingtimeForMVD’s，wefirstdiscussthereasonofMVDirifereneeoccurrenceunde

6、rtherecordclassificationschemaThenweproposeasufficientandnecessaryconditionforMVD‘compromisesdetectionandelimination，andgiveanalgorithmthatwillincurminimuminformationlosstO甜justtuplelevelsSOthatFDcompromisesareavoidedFin

7、allyweanalysethecorrectnessandtimecomplexityofthisalgorithmThetwoalgorithmsabovehavebeenimplemented，andhavebeenprovedtobecorrectandeffectivetKeywords1MLSDBS；FunctionalDependecies：MultivaluedDependenciesInferenceCompromis