Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore the fundamental mechanisms behind fuzz testing through this 39-minute conference talk that demystifies how fuzzers actually work beyond the common misconception that they are simply "magic bug printers." Trace the evolution of fuzzing from its 1988 origins with Miller's random byte sequences to Unix utilities, through early predecessors like Breuer's 1971 fault detection work and Purdom's 1972 parser testing, up to the revolutionary impact of American Fuzzy Lop (AFL) in 2013. Understand how AFL combined coverage observations with evolutionary novelty search concepts to create a tool that could evolve inputs through mutations to discover untested code paths. Examine the challenges that arose from AFL's success, including the opacity of modern fuzzer operations that has led to research stagnation and inadequate harness writing practices. Analyze recent research papers and case studies that reveal the true relationship between fuzzers, their inputs, and target programs, moving beyond surface-level understanding to uncover the underlying principles. Learn why modern fuzzers work, what their fundamental limitations are, and how to write more effective fuzzers and harnesses by developing a proper mental model of fuzzer-program interactions across different classes of subjects under test.
