An Economy-mode Framework for Task Offloading in Fog Computing Networks

Abstract

In the fog network, by offtoading the computing tasks to the fog nodes (FNs) located at the edge of the network, the task processing delay of the terminal nodes (TNs) can be reduced significantly. However, not all the FNs are willing to help TNs for free. Therefore, an economy-mode framework for task offtoading in the fog network is required to motivate the FNs. In this paper, we formulate the task offtoading problem based on Stackelberg game. The TN with computing tasks acts as the leader and its strategy is the optimal unit computing price. Each FN with idle resources acts as the follower and its strategy is the sharable optimal proportion of computing capability. Firstly, the leader predicts the followers' strategies and provides the unit computing price. Next, given the price, the followers will do exactly what the leader wishes, which is the system optimal point. By maximizing the utility functions of both the TN and FNs, a win-win solution is obtained. The numerical simulations indicate the importance of our proposed economy-mode framework for task offtoading.