D4V: a peer-to-peer architecture for data dissemination in smartphone-based vehicular applications

Vehicular data collection applications are emerging as an appealing technology to monitor urban areas, where a high concentration of connected vehicles with onboard sensors is a near future scenario. In this context, smartphones are, on one side, effective enablers of Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure (V2I) applications and, on the other side, highly sophisticated sensing platforms. In this paper, we introduce an effective and efficient system, denoted as D4V, to disseminate vehicle-related information and sensed data using smartphones as V2I devices. D4V relies on a Peer-to-Peer (P2P) overlay scheme, denoted as Distributed Geographic Table (DGT), which unifies the concepts of physical and virtual neighborhoods in a scalable and robust infrastructure for application-level services. First, we investigate the discovery procedure of the DGT overlay network, through analytical and simulation results. Then, we present and discuss an extensive simulation-based performance evaluation (considering relevant performance indicators) of the D4V system, in a 4G wireless communication scenario. The simulation methodology combines DEUS (an application-level simulation tool for the study of large-scale systems) with ns-3 (a well-known network simulator, which takes into account lower layers), in order to provide a D4V proof-of-concept. The observed results show that D4V-based information sharing among vehicles allows to significantly reduce risks and nuisances (e.g., due to road defects and congestions). Subjects Computer Networks and Communications, Distributed and Parallel Computing, Mobile and Ubiquitous Computing