Reproduce “Multiple Targets Directed Greybox Fuzzing” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Hongliang Liang, Xianglin Cheng, Jie Liu, Jin Li
Researchers from Beijing University of Posts and Telecommunications and CNKLSTISS developed LeoFuzz, a directed greybox fuzzer designed to efficiently find multiple vulnerabilities in software simultaneously. This fuzzer outperformed six state-of-the-art tools, reproducing all 31 vulnerabilities 7.35 times faster than QSYM for multi-target scenarios and uncovering 23 new vulnerabilities.
Key insights
- A single instance of a multi-target fuzzer can be more efficient than running multiple parallel single-target instances, as it can leverage interdependencies between related bugs.
- Fine-grained energy scheduling that considers multiple relationships between a seed's execution and various target sequences significantly improves fuzzing effectiveness.
- Dynamically adapting the balance between exploration (finding new paths) and exploitation (deepening existing paths towards targets) is crucial for efficient multi-target directed fuzzing.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.