Research On Key Technologies Of Watermarking Databases

Data sharing technologies, such as database outsourcing or online database services, have brought us great convenience in information exchanging. Meanwhile, problems of data piracy, data tampering and illegal dissemination are becoming increasingly serious. Database copyright protection issue in network enviroment has become a pressing problem. This thesis aims to study the key technologies of database watermark and its' applications in database copyright protection. Beginning from the problem of database copyright authentication, integrity authentication, traitor tracing and watermarking protocol etc., we studied several key issues in the database watermakring, such as, robust watermarking, fragile watermarking, digital fingerprints and watermark protocols. The main contributions of this dissertation are summarized as follows:Firstly, robust database watermark is studied. In the first place, a location-based watermark embedding and extracting method is presented, and use decision-making method based on pattern matching to determine whether there was a piracy. Then, a chaotic sequence was introduced into database watermark, an effective algorithm for embedding chaotic watermark into numerical relational database was proposed, together with a novel watermark detection algorithm and a decision making algorithm to determine piracy.Secondly, the significance of different attributes in a databse is evaluated using method of Analytic Hierarchy Process (AHP), and each attribute is assigned a weight accordingly. An attribute weighted algorithm for watermarking database is proposed thereby. The basic idea of this algorithm is to embed more watermark information into the more important attributes. The advantage of this method is taking the different significance of attributes into account. Thus achieves the idea of the more significant an attribute is, the more information it is embedded.Thirdly, based on the technology of fragile database watermarking and survival database DBMS intrusion detection, a survival database oriented fragile watermark based temper detection model is presented. In this model, by embedding fragile digital watermarks into databases and detecting them, the system can detect intrusion detection for tamper sensitivity of fragile digital watermarks. Compared with the existing intrusion detection schemes, FWTDS model has higher adaptability and security.Fourthly, based on the analysis of principles and applications of digital fingerprint technology, the the idea of attribute weight is introduced, and an algorithm for fingerprinting relational database based on attribute weight was presented. This algorithm consists of the digital fingerprint generation algorithm, embedding algorithm, extraction algorithms and traitor tracing algorithm. The algorithm determines the strength of digital fingerprint according to the importance of different attributes, and enhances robustness of digital fingerprints by a majority electoral system. Both the accuracy of digital fingerprint detection and traitor tracing accuracy are improved.Finally, on the basis of a thorough analysis of several well-known digital watermarking protocols, also refering to digital watermarking applications and transaction mode of digital products, combined with public key cryptography, a Trusted Third Party (TTP) based Buyer-Seller Watermarking Protocol (TBDWP) was presented. According to this protocol, the buyer and the seller are eqaul. TTP is responsible for monitoring transactions, and is responsible for arbitration. The users'rights problem, non-binding problem are well solved in this protocol. The total communication overhead is also reduced. The protocol is secure, fair, flexible and efficient. In addition, the protocol is suitable for a variety of digital media such as images, audio, video and databases.