Estimating Worst-case Resource Usage by Resource-usage-aware Fuzzing

Abstract Worst-case resource usage provides a useful guidance in the design, configuration and deployment of software, especially when it runs under context with limited amount resources. Static resource-bound analysis can provide sound upper bounds worst-case but may too conservative, even unbounded, results. In this paper, we present resource-usage-aware fuzzing approach to estimate usage. The key idea is guide process using resource-usage together relevant coverage. Moreover, leverage semantic patch make use static information (including control-flow, function-call, etc.) instrument original program, for sake aiding subsequent fuzzing. We have conducted experiments various resources real-world programs, including heap memory, stack depths, sockets, user-defined resources, etc. preliminary experimental results show promising ability our estimating compared two state-of-the-art tools (AFL MemLock).