Design And Implementation Of Relational Sensitive Data Protection System

Data is the key factor of production in the development of digital economy.The sharing and circulation of data is an important way to exert the vitality and value of data factors of production.For example,the China-Customs data sharing platform promotes information construction and improves economic vitality.At present,although there are many kinds of data,relational data is still the mainstream data form.And there is a lot of sensitive data in it.The leakage of sensitive data seriously damages the right of use and ownership of data owners,so it greatly affects the availability and security of data in sharing and circulation.In this paper,from the perspective of the protection of sensitive data,I have analyzed the research status at home and abroad.And I have been clear about the data masking and database watermark two technologies of existing problems.According to different situations of sensitive data protection requirements,from the Angle of the active protection of sensitive data,design has realized the sensitive data protection system for relational data.The system can not only actively protect personal privacy,corporate secrets and other sensitive information,but also have the ability to confirm the rights and trace the data leakage behavior,which provides theoretical basis and practical application for the protection of relational sensitive data.The main research contents include:Firstly,in view of the lack of flexibility in the configuration of basic static desensitization technology under multi-demand scenarios,nearly 20 kinds of existing basic masking technologies are analyzed and implemented.According to the reversibility of different masking technologies,technical characteristics and data types of database attributes,personalized masking strategies under different desensitization levels are designed.Secondly,dynamic masking in interactive environment is realized through dynamic real-time masking of whole-process data.For the original data query and statistical data query two types of protection respectively.The SQL statement was rewritten according to the data type of the attribute.The original data query protection was realized by using the rewritten SQL statement to query the database,and the statistical data query protection was realized based on the histogram satisfying differential privacy.Thirdly,the right confirmation and traceability method of database watermarking technology based on multi-type data are studied.A secure and reversible watermarking algorithm based on semantics is proposed for numerical attributes and a watermarking algorithm based on Chinese word segmentation and word embedding is proposed for non-numerical attributes.Both algorithms use double key and double encryption,which makes the algorithms have high security and strong robustness.In addition,the embedding is implemented for numerical,Chinese and English attribute types respectively,which makes full use of the data and improve the watermark capacity.Finally,a relational sensitive data protection system is designed and implemented.The system can manage the data resources owned by users,realize static masking of the whole table,dynamic masking in real time,and confirm and trace the data rights.Verification through system testing shows that the system can effectively realize the dynamic and static masking operation as well as the data right confirmation and traceability function,protect the data availability and security,maintain the data owner's right of use and ownership and other legal rights,and achieve the system requirements.