Database security comprises a set of measures, policies and mechanisms to provide secrecy, integrity and availability of data and to combat possible attacks on the system from insiders and outsiders, both malicious and accidental. The increasing development of information technology in the past few years has led to the widespread use of computer systems in various public and private organizations, such hospitals, universities, banks, libraries and so on. The increased reliability now offered in hardware and software technologies, coupled with the continuous reduction of costs, the increasing technologies, the increasing professional expertise of information specialists and the availability of support tools, have all contributed to encourage the widespread use of computing services. This has meant the more data than ever before is now stored and managed by computer systems, or rather by the tools and techniques capable of supporting and meting these application requirements. A great attention is according to medical databases. These databases are statistical databases because we want to have information about an interrogation set and not for a specifically record. Is very important for medical statistically databases to keep the secrecy about a separated records, because in most situations these information are confidentially. The paper presents a new approach to the notion of protection for statistical databases. This approach is inspired by the idea of integrity in the database semantics of all its defining characteristics, on one hand, and by a previous approach, which introduced statistical elements in the protection of the database, on the other hand. The proposed protection mechanism, based on the perturbation of answers, allows time and space reduction, compared to previous approaches. The modularity of the protection mechanism that we propose facilitates its implementation. Beside the protection of the database, the central element in this study is the utility (the accuracy of the answers altered by statistical means).


Statistical database, Modularity, Utility, Random variable.