Improvement of the sample mutation strategy based on fuzzing framework peach

Fuzzing technology plays an important role in the process of vulnerability mining. Peach is a relatively mature tool in Fuzzing technology. In this paper, the sample mutation strategy in the Peach framework are studied, and the improved scheme is proposed. The scheme is shown in two aspects: first, on the basis of one-dimensional mutation strategy this paper puts forward a multi-dimensional mutation strategy; second, this paper optimizes the random number generation method of the mutator by introducing the two-dimensional Logistic chaotic mapping algorithm which could generate the chaotic sequence with better randomness and higher uniformity. Experiments were carried out to compare the improved scheme with the original one and the results shows that Peach frame with the improved mutation strategy discovered more abnormal points of software than Peach with original one in the same circumstances. This scheme refines the grain size of the variation of the test samples, and improves the efficiency of Fuzzing to a certain extent.