【作者】 崔新春；

【导师】 秦小麟；

【作者基本信息】 南京航空航天大学 ， 计算机应用技术， 2011， 博士

【摘要】 数据外包处理、在线数据库服务等数据共享技术为交流信息带来极大的便利。同时，数据的盗版、篡改和非法传播问题也日益严重。网络环境中的数据库版权保护问题成为亟待解决的难题。本文旨在研究数据库水印的关键技术及其在数据库版权保护中的应用。主要从数据库版权认证、完整性认证、叛逆者追踪以及水印协议等方面入手，系统研究了水印数据库中的几个关键问题，如：健壮性水印、脆弱性水印、数字指纹以及数字水印协议等。本文的主要贡献与创新点如下：⑴深入研究了健壮性数据库水印。首先，提出一种基于水印位置的向数据库中嵌入和提取水印的新方法，并且使用基于模式匹配的决策方法来判定是否发生盗版。其次，将混沌序列引入数据库水印，提出一种有效的将混沌水印嵌入到数值型关系数据库的算法和一种新颖的水印检测算法以及判定盗版的决策算法；⑵引入层次分析法，对数据库不同属性的重要程度做出评价并赋予不同的权重。据此，提出一种基于属性权重的数据库水印算法。其基本思想是：向重要程度高的属性中嵌入更多的水印。该算法的优越性在于将充分考虑了不同属性的重要程度，将水印强度与之结合，进而实现重要属性水印强度高的设计思想。⑶在脆弱性数据库水印技术和可生存数据库DBMS入侵检测的基础上，提出了一个面向可生存数据库的基于脆弱性数字水印的篡改检测系统（FragileWatermarkbasedTemperDetectionSystem,FWTDS）模型。该模型通过对数据库对象中的脆弱性数字水印的嵌入和检测，利用脆弱性数字水印对篡改的敏感性来检测入侵。和已有的入侵检测模型相比较，FWTDS模型具有更高的适应性和安全性。⑷在分析数字指纹技术原理和应用的基础上，引入数据库属性权重分析的思想，提出了一个基于属性权重的关系数据库数字指纹算法。该算法包含数字指纹生成算法、嵌入算法、提取算法以及叛逆者追踪算法。它依据数据库属性不同的重要性确定嵌入数字指纹的强度，并通过多数选举机制增强数字指纹的健壮性，提高了数字指纹检出的准确性和叛逆追踪的精确性。⑸在分析已有的几种著名数字水印协议的基础上，参考数字产品交易模式和数字水印应用，并结合公钥密码技术，提出了一种基于TTP的Buyer-Seller公平数字水印协议TBDWP。在TBDWP协议中买方和卖方地位对等，TTP负责监督买卖过程并负责仲裁。较好地解决了用户权利问题、非绑定问题并且有效减少了通信开销。该协议是安全、公平、灵活和高效的。另外，该协议适合于多种数字媒体，如：图像、音频、视频以及数据库等。更多还原

【Abstract】 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.更多还原