A Fuzzer Based on a Fine-Grained Deeper Strategy

Software nowadays suffers much danger from vulnerabilities, threatening much valuable thing, e.g. the security of private information. Fuzzing is a successful tool in bug-detecting without resorting to much prior knowledge, and has actually discovered many bugs. However, traditional fuzzing has a common drawback that it is limited in a superficial level, and nearly cannot drill deep into the program. This paper proposes DO-Fuzzer (Depth-Oriented-Fuzzer) based on an evolutionary skeleton, employing a fine-grained deep strategy to guide the fuzzer deep into the program. Experiment shows that the deep strategy can augment the fuzzer to detect program-path and consequently program-bugs.