FogJam: A Fog Service for Detecting Traffic Congestion in a Continuous Data Stream VANET

In a continuous data stream vehicular network environment, a Traffic Congestion Detection Service (TCDS) receives many periodic information to update and discover road segments with low speeds and high vehicular density. The use of Cloud computing to support the massive amounts of traffic data received from multiple vehicles significantly increases network traffic. Therefore, we propose FogJam, a Fog service to detect traffic congestion directly at the edge of the vehicular network. In the network edge, FogJam leverages sampling and clustering-based methods to reduce the traffic data stream transmitted by all vehicles on the network links to the Cloud. In the Cloud, FogJam is used as a macro-control of all vehicles geographic positions, acquiring traffic flow data to detect traffic congestion. We evaluate FogJam using OMNeT++, Veins and SUMO simulators. The results suggest that FogJam is highly accurate in detecting traffic congestion at a lower cost, even in a high vehicular density scenario. Furthermore, using clustering-based methods, FogJam is able to, on average, reduce the impact on network usage by approximately 70% compared to the sampling methods, while maintaining an acceptable level of congestion detection accuracy.