ETAP: Energy-aware Timing Analysis of Intermittent Programs

Energy harvesting battery-free embedded devices rely only on ambient energy that enables stand-alone and sustainable IoT applications. These execute programs when the harvested in their reservoir is sufficient to operate stop execution abruptly (and start charging) otherwise. intermittent have varying timing behavior under different conditions, hardware configurations, program structures. This article presents Energy-aware Timing Analysis of Programs (ETAP), a probabilistic symbolic approach analyzes at compile time. ETAP symbolically executes given while taking time cost models for dynamic consumption into account. We evaluate by comparing compile-time analysis results our benchmark codes real-world application with executions real hardware. Our evaluation shows ETAP’s prediction error rate between 0.0076% 10.8%, it speeds up least two orders magnitude compared manual testing.