STATIC AND DYNAMIC ANALYSES FOR PROTECTING THE JAVA SOFTWARE EXECUTION ENVIRONMENT

In my thesis, I present three projects on which I have worked during my Ph.D. studies. All of them focus on software protection in the Java environment with static and dynamic techniques for control-flow and data-dependency analysis. More specifically, the first two works are dedicated to the problem of deserialization of untrusted data in Java. In the first, I present a defense system that was designed for protecting the Java Virtual Machine, along with the results that were obtained. In the second, I present a recent research project that aims at automatic generation of deserialization attacks, to help identifying them and increasing protection. The last discussed work concerns another branch of software protection: the authentication on short-distance channels (or the lack thereof) in Android APKs. In said work, I present a tool that was built for automatically identifying the presence of high-level authentication in Android apps. I thoroughly discuss experiments, limitations and future work for all three projects, concluding with general principles that bring these works together, and can be applied when facing related security issues in high-level software protection.