Flexible Virtual Energy Sharing by Distributed Task Reallocation in IoT Edge Networks
Ruitao Chen, Xianbin Wang, Shuran Sheng · 2018
The convergence of Internet of Things (IoT) and edge computing provides a promising solution to many distributed IoT applications, which often involve real-time information gathering and complex processing. However, energy consumption of computational-intensive processing at edge devices becomes a main constraint due to limited battery capacity. In addressing this, computation offloading to a remote server has been utilized but leading to potentially increased latency due to network delay. In this paper, we propose a virtual energy sharing among edge IoT devices through a collaborative computing mechanism by a flexible and situation-dependent computation task reallocation coordinated by a smart gateway. Our key objective is to achieve energy aware task executions and energy sharing among collaborative edge devices by flexibly adjusting task volumes and CPU frequency based on the workload conditions. To implement this, trade-off between energy consumption and computation time of two collaborative edge devices is formulated to achieve a flexible and situation-dependent decision-making considering the time-varying resource conditions and application delay tolerance. Simulation results show that the proposed scheme could achieve a flexible virtual energy sharing by managing the trade-off between resource utilization and latency performance of tasks.