Energy-efficient computation offloading strategy with task priority in cloud assisted multi-access edge computing

Abstract

Multi-access edge computing (MEC) provides cloud-like services at the edge of the radio access network close to mobile devices (MDs). This infrastructure can provide low-latency services to MDs and significantly reduce the pressure on the backbone network. However, the computing resources configured on an edge server (ES) are limited compared to a cloud data center (DC). It is difficult for ESs to satisfy the demands of MDs anytime and anywhere. Thus, a new paradigm that combines DC with ESs has been proposed to provide better capability and flexibility, namely, cloud-assisted MEC (CA-MEC). In CA-MEC, MDs can offload tasks to ESs and the DC, which means more elasticity and more complicated offloading decisions. This paper studies MDs’ energy-efficient computation offloading strategy in CA-MEC, which considers two different priority tasks. First, we establish mathematical models to characterize the CA-MEC environment. Second, we mathematically analyze the MD’s average task response time and average power consumption. Third, we propose efficient numerical algorithms to obtain a computation offloading strategy to optimize the energy efficiency of the target MD. Finally, we demonstrate several numerical examples and construct a comparative experiment to show the effectiveness of our algorithms.