This item is available under a Creative Commons License for non-commercial use only
1.2 COMPUTER AND INFORMATION SCIENCE
Whitebox fuzz testing is a vital part of the software testing process in the software development life cycle (SDLC). It is used for bug detection and security vulnerability checking as well. But current tools lack the ability to detect all the bugs and cover the entire code under test in a reasonable time. This study will explore some of the various whitebox fuzzing techniques and tools (AFL, SAGE, Driller, etc.) currently in use followed by a discussion of their strategies and the challenges facing them. One of the most popular state-of-the-art fuzzers, American Fuzzy Lop (AFL) will be discussed in detail and the modifications proposed to reduce the time required by it while functioning under QEMU emulation mode will be put forth. The study found that the AFL fuzzer can be sped up by injecting an intermediary layer of code in the Tiny Code Generator (TCG) that helps in translating blocks between the two architectures being used for testing. The modified version of AFL was able to find a mean 1.6 crashes more than the basic AFL running in QEMU mode. The study will then recommend future research avenues in the form of hybrid techniques to resolve the challenges faced by the state of the art fuzzers and create an optimal fuzzing tool. The motivation behind the study is to optimize the fuzzing process in order to reduce the time taken to perform software testing and produce robust, error-free software products.
Mahajan, R. (2019). Augmenting American Fuzzy Lop to Increase the Speed of Bug Detection. Dissertation M.Sc. in Computing (Advanced Software Development), Technological University Dublin, 2019.