Scope
In (Trihinas et al, 2022) the authors provide a comprehensive overview of the metrics and methodologies required to estimate energy consumption and carbon emissions in IoT settings. It covers: i) fundamental concepts such as power consumption (using equations for energy, power usage effectiveness, etc.) and carbon intensity, ii) experimental evaluation of AI-driven IoT services through a reference scenario involving edge micro-data centers, iii) analysis of factors like training time, geographical location, and resource heterogeneity on the environmental footprint. iv) identification of challenges and recommendations for integrating these assessments into existing IoT testing frameworks.
Summary
This work investigates the environmental impacts of AI-driven IoT applications. It highlights that while current IoT testing tools effectively emulate heterogeneous edge environments, they largely overlook the critical aspects of energy consumption and carbon emissions during the testing phase of AI services. The authors present key questions, observations, and challenges to address this gap, aiming to guide the development of next-generation testing and benchmarking suites that incorporate energy and carbon footprint assessments.
Relevance for EXIGENCE
Trihinas et al.’s study is highly relevant to EXIGENCE project for energy ecosystem, energy optimisation, green incentives and incentive mechanism – which focuses on reducing energy usage and optimising sustainability in technology systems. By providing a structured framework for quantifying the energy consumption and carbon footprint of AI-driven IoT services, the paper offers valuable insights and methodologies that can be adapted within EXIGENCE. These insights support the incentive mechanisms and technical solutions aimed at promoting energy-efficient practices, thereby directly contributing to EXIGENCE’s goal of aligning technological innovation with climate change mitigation.
D. Trihinas, L. Thamsen, J. Beilharz, and M. Symeonides, “Towards Energy Consumption and Carbon Footprint Testing for AI-driven IoT Services”. IEEE International Conference on Cloud Engineering (IC2E), 2022.
