Cooperative Offloading in D2D-Enabled Three-Tier MEC Networks for IoT

Mobile/multi-access edge computing (MEC) takes advantage of its proximity to end-users, which greatly reduces the transmission delay of task offloading compared to mobile cloud computing (MCC). Offloading computing tasks to edge servers with a certain amount of computing ability can also reduce the computing delay. Meanwhile, device-to-device (D2D) cooperation can help to process small-scale delay-sensitive tasks to further decrease the delay of tasks. But where to offload the computing tasks is a critical issue. In this article, we integrate MEC and D2D cooperation techniques to optimize the offloading decisions and resource allocation problem in D2D-enabled three-tier MEC networks for Internet of Things (IoT). Mobile devices (MDs), edge clouds, and central cloud data center (DC) make up these three-tier MEC networks. They cooperate with each other to finish the offloading tasks. Each task can be processed by MD itself or its neighboring MDs at device tier, by edge servers at edge tier, or by remote cloud servers at cloud tier. Under the maximum energy cost constraints, we formulate the cooperative offloading problem into a mixed-integer nonlinear problem aiming to minimize the total delay of tasks. We utilize the alternating direction method of multipliers (ADMM) to speed up the computing process. The proposed scheme decomposes the complicated problem into 3 smaller subproblems, which are solved in a parallel fashion. Finally, we compare our proposal with D2D and MEC networks in simulations. Numerical results validate that the proposed D2D-enabled MEC networks for IoT can significantly enhance the computing abilities and reduce the total delay of tasks.

Automated summary

This article integrates MEC and D2D cooperation techniques to optimize offloading decisions and resource allocation in D2D-enabled three-tier MEC networks. Experimental results validate that the proposed D2D-enabled MEC networks for IoT can significantly enhance computing abilities and reduce task delay.