Queueing Model for Intermittent Communication and Computing of Battery-Less Edge Computing

Battery-less intermittent IoT system supports reliable computing without a battery or stable energy supply but with the drawback of low computation capacity. To overcome this drawback, intermittent edge computing systems are proposed to integrate Multi-access edge computing (MEC) with intermittent devices through wireless connections. The key challenge of such a system is determining the offloading timing: offloading could be inefficient when it is interrupted by power failures. An accurate estimation model of the offloading efficiency is the key to determining the optimal offloading decisions. We proposed a queueing-based estimation model to estimate the offloading and local computing efficiency of a Bluetooth Low Energy (BLE) intermittent edge computing system. The proposed Discrete Time Markov Chain (DTMC) model takes the energy arrival and task creation processes into account to capture the uncertainty in the battery-less devices. The simulation results suggested that the proposed model provides more accurate estimations and achieves the lowest system latency compared with the state-of-the-art model and baseline algorithms.