Fog-Enabled Cooperative Offloading for Intermittently Connected Vehicular Networks

Abstract

In this paper, we investigate how to utilize the surplus resources in intermittently connected vehicular networks (ICVNs) to help target vehicles downloading and processing files from remote server. To this end, we propose a fog-enabled scheme for ICVNs to make the pooled resources distributed among moving vehicles work cooperatively. Specifically, each target vehicle is served by a virtual mobile fog node whose resources are extracted from a set of supporting vehicles coming from the opposite direction. The corresponding supporting vehicles instead of the target one download chunks of raw data from the remote server and pre-process them respectively. The pre-processed data are further carried on by the supporting vehicle and transferred to the target until they encounter. Considering the V2V communication interference as well as the computing and storage capability constraints of each supporting vehicles, we formulate the offloading problem as a mixed integer programming problem which minimize the average time cost of target vehicles receiving entire processed data. Since the problem is NP-hard, we design a heuristic algorithm to schedule the resource pool to help target vehicles. The simulation results confirmed the effectiveness of our algorithm in reducing the time cost. Furthermore, the proposal is more robust in different file sizes compared with choosing supporting vehicles by encounter order.