A Big Sensor Data Offloading Scheme in Rail Networks

In this paper, we propose an offloading scheme to transfer massive stored sensor data from rolling stock to railway data centers. We apply a delayed offloading strategy for non-critical stored data assuming that the critical data has been already separated through an appropriate edge processing task and has been sent via a real-time communication such as cellular networks. We propose train stations as potential and feasible spots for data offloading via available wireless local area networks (WLAN) such as existing WiFi network at stations. Thus, stations will not only be the places of passenger exchange but also data exchange. We develop an analytical model customized for the proposed offloading strategy in rail applications. Then we validate the performance of our model through simulation in various scenarios in Omnet. The simulation results shows an accuracy of %98.67 for the proposed analytical model with reference to the simulation results in Omnetpp. Additionally, by using our proposed scheme, we can theoretically offload up to 5.43 GB per each stopping station.