An Experimental Study on Optimizing Energy Consumption of Edge IoT Devices
MetadataShow full metadata
Recent advances in the world of IoT have significantly improved what one can do with the help of personal gadgets such as smartwatches and other edge devices. This improvement has lead to the inclusion of additional functionality like fall detection, temperature and motion sensing, etc. using an edge device. This in itself is not a bad thing, however, this does pose the problem of increased energy consumption on devices which are already constrained by limited battery. A portion of this energy is consumed during client-server communication which is an essential part of applications that either perform major computations from sensor data on the cloud or use the server to backup user data. We present an experimental study which discusses the possibility of optimizing energy consumption for edge IoT devices by reducing communication cost of applications by determining an optimal chunk size. We present a metric energy-per-byte or EPB which helps determine the optimal chunk size for respective edge devices. We also discuss a case study to (1) determine whether containerized applications are energy efficient and (2) if a coordination language framework like Lingua Franca can help design energy efficient applications. We evaluate our approach with the help of two applications running on different edge devices and softwares, and we show that we achieve considerable energy optimization depending on the application functionality and the hardware involved.