Research On Data Provenance Technology Based On Database Watermarking

In recent years,with the rapid development of data exchange,data security has become more and more important.Once the data is leaked,it will pose a great threat to the privacy and property security of users.Copyright authentication and data provenance have become an important requirement of the data security defense mechanism.The major security manufacturers also focus on data provenance and have achieved certain results.However,under the temptation of huge profit caused by stealing data,the attack mode and anti-detection ability of malicious personnel are also developing,which undoubtedly brings great resistance to the research of data provenance technology.Therefore,it is more important to study effective data provenance algorithms.Based on the analysis and research of the current common data provenance technology,this paper proposes a data provenance technology based on database watermarking,which is mainly used to solve the problem of embedding provenance information and low provenance accuracy when the data types are complex.The main contents and innovations are as follows:(1)We analyze the structure and definition of the data provenance system,then the existing data provenance technology is sorted out.We have a brief introduction to current tag-based data provenance techniques,blockchain-based data provenance techniques,and database watermark-based data provenance techniques.This section focuses on several data provenance techniques based on database watermarking;(2)Based on the analysis and research of the current data provenance technology based on database watermarking,a data provenance algorithm based on virtual tuple is proposed for the limitations of current database watermarks on data types.The algorithm uses the data dictionary and the pseudo-random generator to generate non-numeric attributes and numerical attributes in the virtual tuple respectively,and embeds the watermark information through high-simulation data.After the data is leaked,the virtual tuple can be extracted and compared to track the traitor.The experimental results show that the algorithm has high traceability accuracy under common database attacks.(3)Based on the research of various data provenance techniques based on database watermarking,a data provenance algorithm based on reference relations preservation is proposed for the problem of large storage space.The algorithm can solve the limitation of data types in general algorithms,and it does not need extra space to store the watermark information.It can use different keys to encrypt the added primary foreign key to achieve watermark addition for different users,thereby embedding and detecting the watermark information.(4)According to the above two algorithms,the experimental simulation environmentis built to test the two algorithms.It can be seen from the experimental results that the algorithms in this paper are applicable to the database of various data types,and the robustness of them are both good.The algorithms provide more effective protection for data distribution and storage.