Reliability analysis of 5G-VANET using cloud-fog-edge based architecture

Abstract

Intelligent Transportation System (ITS) is an application which focuses on building and improving road safety and transportation system of all kinds. To contemplate the capabilities of the ITS applications, data transmission between vehicle to infrastructure (V2I) should be highly efficient and reliable. One of the important technologies that facilitates ITS to achieve its goal is Vehicular Ad hoc network (VANET) which has envisioned benefits ranging from autonomous vehicles, improving road safety and reducing traffic congestion to entertainment services for passengers’ convenience and comfort. However, with the emergence of 5G networks, it is imperative to integrate 5G and vehicular networks. To provide the needed resources for supporting these myriads of emerging applications, fog and edge computing have further been put into action at par with cloud computing. The technology of fog computing in 5G has turned out to be an adequate solution for faster processing in delay sensitive applications, such as VANETs, being a hybrid solution between fully centralized and fully distributed networks. Given the rise in popularity of ITS, which exhibit similarities to other intricate, interconnected systems, a substantial volume of data will be generated via vehicular networks. This data necessitates reliable and secure processing, highlighting the need for dedicated research on reliability. Therefore, it is crucial to assess the reliability and architectural issues of vehicular networks. In this paper, a novel architecture for fifth-generation VANETs (5G-V) is suggested which provides a seamless integration of 5G-V with cloud-fog-edge computing. A three-level hierarchical model is developed and its reliability metrics are obtained using analytical models. Stochastic modeling techniques like Markov chains and reliability block diagrams are used to develop these models. To demonstrate the viability of the given approach, numerical illustrations of the proposed models are presented graphically.