Toward Performance and Energy-Efficient Edge-of-Things

Abstract

Edge computing is a recent trend wherein a (large) number of Internet of Things (IoT) devices are being connected to the cloud via edge devices. Although edge devices generally have (slightly) more computing power than IoT devices, several strategies can help to improve performance and reduce energy consumption. The key research challenge here is how to ensure efficient resource utilization, particularly for real-time delay-sensitive applications, such as military and battlefield contexts, and real-time disaster recovery and response efforts. Focusing on edge computing with a (large) number of IoT devices, we propose a dynamic policy-based computing approach to improve the overall performance of Edge-of-Things (EoT). We evaluated the energy consumption of heterogeneous edge-only resources for given policies to demonstrate their impact on an EoT computing infrastructure. Finally, we discussed open issues and future research directions in green EoT computing.