By: Richard Johnson

Scheduled on: June 19 at 16:00

The past few years have seen a leap in fuzzing technology. The original paradigm established a decade ago resulted in two widely deployed approaches to fuzzing: sample based mutation and model based generation. Thanks to ever-increasing computational performance and better engineering, newer guided fuzzing approaches have proven to be supremely effective with a low cost of deployment. This talk will explore a few different approaches to guided fuzzing through dynamic analysis including code coverage analysis, constraint solving, and sampling/profiling based feedback mechanisms.

Novel contributions in this talk include: - Opensource Windows Driver enabling Intel “Processor Trace” - DBI based tracing engine for Windows/Linux/OSX binaries - American Fuzzy Lop with full support for Windows binary targets