Design And Implementation The Middleware Of Damaged Data Detection Based On Application Semantics

Damaged data detection, as one of the most important components of theIntrusion tolerance technique, is helpful for the maintenance of the existence ofdatabase application system. But the extant damaged data detection techniques canhardly reach the requirements of damaged data analysis of database application system.The over estimation issue of the damaged data detection result,the low serviceabilityfor recovery system and the architecture is not suitable for a variety of databaseapplication system.In order to solve the problems above, design the damaged data detectionmiddleware which based on the application of semantic, obtain the applicationsemantics information of database application system and analysis the data damage.Middleware consists of a user interface, transaction semantic analysis module anddamaged data analysis module. Database application system connects to the databasethrough the user interface, submitted to database operations, import semanticinformation based application which the database application system contains. Userinterface improves the compatibility issues of damage data detection technology anddatabase application system. Application semantics analysis module captures databaseoperations submitted by the user interface. Identify the database transactioncorresponding business, and record the transaction semantics analysis results throughthe application of semantic information after the conversion of SQL statement.According to the transaction semantics information and dependence, damaged dataanalysis module eliminates the damaged affairs without the business dependenceand improves the accuracy of damage assessment.It makes use of the semanticinformation of the business and affairs, determine whether the result will be differentafter the recovery of the damaged date related.Finally, the experiments for the exactness and performance of the middleware of damaged data detection based on论app lic文ation semantics are conducted. Theexperimental result shows that the performance of damaged data detection middlewarebased on application semantics is slightly worse than non-application semanticdamaged date detection method. But the performance of damaged data detectionmiddleware based on application semantics improves the over-estimation issue, raisesthe serviceability of the damaged data detection result.