Research On Asmuth-Bloom System Based On Software Watermark

Most traditional software protection techniques rely on encryption technology, but with the rapid growth of calculations by computer, the security of cipher code has always faced new challenges. A new area is exploited for the research of software's safety with the development of software watermarking technology. Software watermarking is a new way to protect software.Software watermarking embeds some special secret messages in software, and these messages can be extracted when it is necessary to improve authority ownership. They can be classified into two categories: static watermarks and dynamic watermarks. A static watermark is stored in program code with a certain format, and it does not change during the program execution. A dynamic watermark is built during the period of program execution, perhaps only after a particular sequence of input.Dynamic graphic watermark is a watermark that is represented in a graph topology, which is created only at the runtime of a program, triggered by a special sequence of input. Because of the hardness of pointer alias analysis such watermarks are difficult to attack automatically.In this thesis, we present a new robust dynamic graphic watermark algorithm based on Asmuth-Bloom system. In this algorithm, the idea of the secret sharing to the software watermarking is introduced, which divides the watermark information into several parts. The watermark recovering based on incomplete content is realized.A watermark system based on Java bytecode files is realized, which we call the ABMark1.0. The comprehensive system mainly consists of two parts, which are static data software watermarking sub-system and AB algorithm software watermark sub-system.Finally, a series of tests are made on the basis of the system, we evaluate the AB algorithm with respect to data rate, stealth, robustness and performance overload caused by watermark embedding. The experiments show that the AB algorithm is resilient against dewatermarking attacks such as code optimization and code obfuscation, but it has a lower data rate, more time and space are needed during the progress of recovering the watermark information.