IoT is one of the most prolific origins of data that is collected from sensory inputs. IoT devices are characterized by low computational power, thus motivating the computation offloading scheme, a promising technique that mitigates energy performance issues in limited power devices. Instead of offloading to fog or cloud, a new research track is emerging where devices unload their computational needs to vehicles roaming around, and this particular type of offloading is called vehicular fog offloading. In this paper, we propose the use of vehicles as Mobile Computing Nodes (MCNs) roaming around within an area to offer computational services to IoT devices. Assigning tasks to an appropriate MCN by an IoT device that wishes to offload specific communication tasks to it will be formulated based on the quality of the communication channel and on the time needed to offload the computation request. Specifically, the vehicle which will serve as an MCN must be within the range of the requesting fixed IoT device and offer suitable SINR and the estimated duration of connection values, while taking into consideration its mobility. Another challenge would be delivering the results back to the IoT device or another party, knowing that the vehicles will not be stationary. The performance results show that the proposed system represents a promising solution by offering computation offloading services and delivering the results within acceptable times, regardless of where the vehicle might be when it wants to return the results.
展开▼
