OTRE-POWER: Open Tool for Reproducible Characterization of Edge Power Consumption

Understanding the actual energy consumption of edge devices, such as Raspberry Pi boards, is becoming increasingly important as these platforms play a central role in real-world IoT applications. Power consumption information is typically sourced from manufacturer datasheets or coarse approximations, which often fail to reflect operational dynamics under realistic workloads. In this paper, we present an open tool for reproducible empirical characterization of edge device power consumption across varying computational workloads. The tool exploits affordable, off-the-shelf hardware for power measurement in combination with system-level metrics and automated workload generation through the stress-ng framework. Power data are collected at high temporal resolution and analyzed using regression models to derive load-dependent energy profiles suitable for integration into system-level energy modeling. Alongside this, we provide a toolchain and a set of Jupyter notebooks to facilitate reproducibility, cross-platform comparison, and future expansion by the community.