Platform agnostic binary coverage maximization
Frank Adkins · 2015
Software security hinges on the need for developers to find vulnerabilities in their software before malicious users do. However, input test suites are manually intensive to produce and often fail to maximize coverage. This generally prescribes the use of a fuzzer. Yet, modern techniques also fail either because they lack the information necessary to make progress or require too much to be practical. In our research, we aim to strike an ideal balance in this spectrum and make an extra effort to address architectures that are commonly excluded from security research.