The Research On The Relational Database Zero-Watermarking Algorithm Based On Text Attributes

With the network technology growing rapidly, e-commerce, e-government and other information management systems based on database are applied more and more to provide services to public. How to protect the data of database from illegal spreading becomes the key issue of database security.Currently, digital watermarking technology is one of the most effective methods to solve the problem. We analyzed existing database watermarking algorithms and designed a database watermarking algorithm, which takes advantage of the attribute that most characters attribute data items in database is used to store text. The main works and achievements in this paper are as follow:First, based on analysis of existing database watermarking algorithms, we designed a database watermarking algorithm which focuses on the text attributes. There are two parts of the algorithm—construction procedure and detection algorithm.Then, the construction procedure is designed. Generally, there are some frequent phrases in different tuples. Therefore, we obtain the feature keys of all tuples by obtaining the feature keys of each tuple. Then the zero-watermarking sequence is constructed by using these feature keys. After that, we did experiments with the different ratio of feature keys and the length of zero-watermarking sequence, thus, we get the affection to zero-watermarking construction of them two. The results show that the algorithm is available.Finally, a detecting algorithm is designed based on ZWMC. The detecting algorithm combines synonym adjusting and partial orders of feature keys. The robustness and false-positive rate of the detecting algorithm is also analyzed through experiments. The experiments results show that the algorithm is effective.