WOWAF: Enhanced Dynamic Binary Analysis Framework Targeting Windows WOW64 Environments

Dynamic binary program analysis plays a crucial role in software vulnerability discovery and malicious code analysis. While 64-bit computing environments have become prevalent, numerous applications still utilize their 32-bit counterparts due to factors such as compatibility considerations, including both regular software and malware. Traditional dynamic analysis frameworks are not optimized for the analysis of 32-bit and mixed-mode programs. As a result, they often encounter issues such as anomalies and performance problems when applied to specific application analysis. To address these issues, this paper proposes an enhanced dynamic binary analysis framework tailored for Windows on Windows 64-bit (WOW64) environments that enables efficient fine-grained analysis of target applications. The framework is implemented on the built-in emulator in 64-bit Windows operating systems, facilitating effective and stable analysis of both pure 32-bit programs and mixed applications. By leveraging kernel features, the framework facilitates dynamic instrumentation of target programs, while incorporating a novel shadow memory allocation management scheme to minimize impact on program execution, and demonstrates its good deployment feasibility. The effectiveness of the analytical framework is validated through comprehensive evaluations of diverse applications in real-world environments, such as exploit programs and evasive malware samples. Additionally, systematic benchmark experiments further demonstrate its strong analytical performance.