Reserarch On Security Of Publishing Views Based On Probabilistic Database

With the development of database technology, data exchange frequency and quantity increase continually, the problem of information disclosure is outstanding day by day in the view publishing process, so guaranteeing security of published view becomes a new subject of database security.This paper analyzes the current situation of the domestic and international security problem in view publishing process, and researches for the problem of security deciding of views.At first, by introducing several different attack methods, such as multi-party collusion and information inference, and the ways to protect sensitive data is introduced in detail, including query retriction, data perturbation, data swapping and K-anonymity protection model etc, it analyzes the existing method of security deciding of views.Secondly, on the basis of the above research, a probability model of security deciding in publishing views is erected. In this model, publishing views and sensitive information are expressed by boolean query. Then the formularized definition of publishing views, private information, prior knowledge is given, and a new deciding theorem based on entropy is provided.Moreover, according to the above model and deciding theorem, the algorithm of security deciding of views based on entropy and the algorithm of security deciding of views based on asymptotic conditional probabilities are introduced. They both solve the NP-complete problem in deciding security of multi-views.Finally, the two algorithms we mentioned above are certified by experiment. They are compared in measure precision and overhead for time and space.