Year: 2019
Type: article
Abstract: Fuzzing is a well-known technique for automatically testing the robustness of software and its susceptibility to securitycritical errors.Recently, many new and improved fuzzers have been presented.One critical aspect of any new fuzzer is its overall performance.However, given that there exist no standardized fuzzing evaluation methodology, we observe significant discrepancy in evaluation results making it highly challenging to compare fuzzing techniques.To tackle this deficiency, we developed a new framework, called FETA, which automatically evaluates fuzzers based on a fixed and comprehensive test set enabling objective and general comparison of performance results.We apply FETA to various recently released academic and non-academic fuzzers, eventually resulting in a large-scale evaluation of the current state-of-the-art fuzzing approaches.
Institution: University of Duisburg-Essen
Language: English
Subfield: Software
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
Funder: Deutsche Forschungsgemeinschaft
Award: G4351222689
OpenAlex ID: W2914981908
