Search

In order to investigate vulnerabilities in stateful protocols, AFL-NET has put forward stateful protocol fuzz testing. In such fuzz testing, the selection of seeds makes a major contribution to the exploration of paths. However, current stateful protocol fuzz testers often repeatedly execute the same several seeds, resulting in an inability to effectively explore more paths. To alleviate this problem, starting from the gain of seeds, this paper proposed an effective seed dynamic scheduling algorithm based on stateful protocols. The algorithm utilized the potential gain, actual gain, and cost of seeds as the gain, using this gain to dynamically schedule seeds and allocate the number of times seeds. Experiments show that this method significantly improves the number of vulnerabilities found and also has a certain degree of improvement in increasing coverage, indicating that the definition of this gain and the seed scheduling algorithm can effectively select seeds and explore more paths and vulnerabilities. [ABSTRACT FROM AUTHOR]