A Double Auction-based MEC-Enabled IoT Task Offloading: A Green Framework

With the impressive growth of the Internet of Things (IoT), the need for offloading delay-sensitive and computing-intensive tasks to resource-rich cloud data centers has arisen. However, the long distance of cloud data centers poses a problem for delay-sensitive tasks. Mobile Edge Computing (MEC) lets IoT tasks be offloaded to proximate servers. This paper considers a MEC-enabled IoT network that User Equipments (UEs) compete against the MEC servers’ limited resources. Therefore, we propose a green task offloading framework in which an auctioneer decides to match UEs and MEC servers through a double auction. To increase the utility of the UEs, a Reinforcement Learning module predicts the best bidding prices at each step. In addition to prices suggested by the two parties, we deem the carbon emission of running tasks in MEC servers in the auction framework. The presented framework believes to maximizes social welfare and reducing the carbon footprint simultaneously.