Data Scheduling For Vehicular Fog Computing

Abstract

The Vehicular Adhoc NETworks (VANET) have become a crucial research area in transportation systems, thanks to technological development, in recent years. These networks generally consist of fixed nodes named Road Side Units (RSU) and mobile nodes (vehicles). The main objective is to ensure a reliable level of road safety and a comfortable aspect of traveling. With the evolution of Internet capabilities, vehicles have become more intelligent on the road, where a new range of modern applications has emerged in the form of communication, management and control. These services allow vehicles to communicate and disseminate safety information through messages. Indeed, these services are hosted in a modern infrastructure designed for communication such as cloud computing. This convergence leads to the appearance of new paradigms linked to vehicles, namely Vehicular Cloud Computing (VCC) and Vehicular Fog Computing (VFC). Despite these advances, the dissemination of messages in most vehicle scenarios remains a major challenge due to packet delays due to the high mobility of vehicles on the road, frequent changes in topology and the high density of vehicles, which leads at road events and frequent packets loss. The aim of this thesis is subscribed in the context of Vehicular Fog Computing (VFC) applications. Although video streaming is the most supported application on driving and the most network sensitive application, we select the video streaming application for our study. We propose a video data scheduling approach based on the creation of virtual networks. This virtual network is a subset (cluster) of the nodes connecting to each other to achieve efficient routing of information for video streaming packets. This mechanism adapts to the change in topology and the high mobility of the nodes. At the evaluation stage, we simulate the performance of this mechanism using an Omnett ++ simulator. The network performance of the proposal showed that the simulation was a precise model and that the best virtual fog nodes for the broadcast request and the video event.