An improved fuzzing approach based on adaptive random testing

Abstract

Fuzzing is a highly automated testing technique. It has been widely used in software vulnerability mining. American fuzzy lop (AFL) is one of the most effective fuzzing tools, with low resource consumption and a variety of efficient fuzzy test strategies. However, because it uses a random testing (RT) algorithm when generating test cases, there is a problem of low quality and low test efficiency. In this paper, we propose an improved fuzzing testing approach based on adaptive random testing (ART) to enhance the effectiveness of AFL test case generation. We also introduce AFL-ART, a new fuzzing tool based on ART. According to the experimental results, AFLART can enhance AFL test case generation, and improve fuzzing testing efficiency.