QoS-aware Task Offloading with NOMA-based Resource Allocation for Mobile Edge Computing
Luyuan Zeng, Wushao Wen, Chongwu Dong · 2022 IEEE Wireless Communications and Networking Conference (WCNC) · 2022
Task offloading can scale the service capacity of IoT devices. However, IoT devices should go through wireless networks to connect with edge computing servers. The wireless network’s performance can not be guaranteed in the dynamic scenario of the mobile environment. Devices would obtain diverse channel quality in frequency, time, and space, which is affected by many factors, such as selective channel fading and path loss fading. The network experience varies significantly between devices, even allocating the same amount of resources for all devices. Besides, too many tasks offloaded to one edge server simultaneously could exhaust the network resources between devices and base station and computing resources in the edge server. So, allocating the communication resource and determining task offloading among devices is a critical issue that should be considered appropriately and comprehensively. Aiming at this problem, we propose a QoS-aware task offloading strategy by decomposing the original problem into two sub-problems: bandwidth resource block allocation and task offloading scheduling. In our approach, the bandwidth resource allocation from the 5G network and task offloading scheduling between multiple edge servers in one edge cloud are jointly considered in two successive phases. Our strategy enables the acceleration of task computation by fine-grained management of network resources in real-time. Simulation results show that our algorithm significantly improves task offloading utility and improves the utilization of network symbol resource.