Performance study on task offloading strategy with cloud-edge-device collaboration based on hybrid access networks

Haoran Du, Qi Zhang, Weiman Sun, Shunfu Jin

Abstract: Task offloading strategy with Cloud-Edge-Device collaboration has attracted extensive attention in diverse network applications. However, most existing studies focused on homogeneous access networks, while ignoring actual scenarios where user devices access strong Wi-Fi, weak Wi-Fi and cellular alternately. This paper aims to fill this gap by developing a method to evaluate task offloading strategy with Cloud-Edge-Device collaboration based on hybrid access networks. In our study, we establish a user device model with variable transmission rate related to access network type and an edge layer model with admission threshold. Then, for Cloud-Edge-Device collaboration, we present an analysis method to derive long-term performance measures in terms of system throughput and average task latency. Numerical and simulation results exhibit that access network type have an impact on system performance of task offloading strategy with Cloud-Edge-Device collaboration. Furthermore, we develop an improved artificial bee colony algorithm to search for the jointly optimal results of offloading probability and edge layer threshold.

External IDs:dblp:journals/winet/DuZSJ25