Multithreading for obfuscating code against dynamic analysis and reverse engineering

Abstract

Code obfuscation is applied to change the structure of software code that makes it hard to understand; in fact, it reduces the readability of code. Obfuscation techniques give protection to proprietary code from reverse engineering, tampering, and unauthorised usage. General code obfuscation techniques make code analyses difficult. Meanwhile, multithreading allows multiple threads to run concurrently within a single process, where each thread helds a separate execution flow, allowing various parts of code to execute simultaneously and independently. Code obfuscation techniques have been used in industrial applications for years, but the scope of multithreading in this area has still not been properly explored. The general code obfuscation techniques change the code structures as well as change the control flow format in a single thread. Multithreading obfuscation has the potential to utilise the complexities of multithreaded execution by adding layers of obscurity to the content and runtime behaviour of code. By applying M threads in a program code with N functions, it may provide N**M combinations, making the control flow harder to analyse. Applying this approach enables the randomness of code execution, and it can give protection against dynamic analysis. Finally, a synchronisation mechanism can increase enough complexity to hinder the reverse engineering. These approaches should make the reconstruction of the original sequential logic much harder. In this research, we aim to design an obfuscation tool for C++ programs that applies multithreading obfuscation techniques to large-scale source code and this would significantly enhance the software security by increasing complexity and unpredictability. And the next step of our research is to focus on the improvement of our obfuscator.