Optimal Offloading Percentage for Fresh Status Updates in Mobile Edge Computing Networks

Abstract

In this paper, we focus on the age of information (AoI) performance of mobile edge computing (MEC) networks, and investigate the optimal offloading percentage to achieve the minimal AoI concerning the partial computing. We study the impact of parameters, including packet size, required CPU cycles for computation, average data rate, and computing capacity of MEC server, on the optimal offloading percentage in MEC network, where a single-source First-Come-First-Served(FCFS), GI/M/I queuing model is considered. Our work verifies the existence of the optimal offloading percentage and yields several interesting insights. Due to the effect of queue, the optimal offloading percentage does not always increase with respect to average data rate, but first increases and then slightly decreases, and the required number of CPU cycles for computation has a similar impact on the optimal offloading percentage.