Symbolic Input Unification and Minimization
Abstract
This effort investigated technology to perform unification and minimization of test cases produced by fuzz testing. Empirical evidence gathered in prior work showed that existing, commonly used unification techniques stack hashing and coverage profiles failed to identify cases that are evidence of the same bug, resulting in (sometimes dramatic) over counts. Less often, they can mistakenly unify two test cases when they shouldn't. Techniques for minimizing test cases, which rely on fuzzing-style mutations, may also suffer from these problems. Most of the work that was accomplished focused on the first two of the original three proposed tasks: 1) exploratory research with existing data; 2) develop test case unification algorithm; 3) test case minimization algorithm.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 09, 2019
- Accession Number
- AD1070382
Entities
People
- Andrew Ruef
- Michael Hicks
Organizations
- University of Maryland