Effect of short, repetitive continuous fuzzing sessions on coverage improvement

Ascertain how short, repetitive continuous fuzzing sessions contribute to code coverage improvement in continuous fuzzing.

Background

Coverage is a primary performance metric for fuzzing, and prior work on traditional long-running fuzzing shows coverage growth, often logarithmic. Continuous fuzzing instead consists of many short sessions and also employs mechanisms like corpus pruning, making its coverage dynamics potentially different.

The authors identify that it is not established whether repeated short sessions, as used in continuous fuzzing, systematically improve coverage over time.