Fuzzing using Logic Specifications

Óbudai Egyetem

Alkalmazott Informatikai és Alkalmazott Matematikai Doktori Iskola

Csiszár Orsolya

informatikai tudományok

Fuzzers traditionally generate inputs to test software using methods such as input mutation, random input generation, or generation of inputs according to some formal grammar. It is sometimes possible to restrict the fuzzer to specific classes of inputs, for example by using a more specific grammar that generates a sub-language of the original one, or by using custom mutators. Both methods suffer from drawbacks. For example, in order to use grammars, the set of inputs must be expressible as a formal language, and custom mutators are usually relatively simple in nature. The purpose of this research topic is to design fuzzing solutions that leverage logical specifications of “interesting” fuzzer inputs. Logical specifications are declarative in nature. They can be used both to generate inputs and to validate inputs that have been otherwise generated. Finally, they can more naturally express logical relationships within the input (“if token X is present, then token Y must also be present unless token Z appears three times”).

1 fő

Óbudai Egyetem

2022-07-01

angol