学术文献库

Fixed edge processing has become a key feature of 5G networks, while playing a key role in reducing latency, improving energy efficiency and introducing flexible compute resource utilization on-demand with added cost savings. Autonomous vehicles are expected to possess significantly more on-board processing capabilities and with improved connectivity. Vehicles continue to be used for a fraction of the day, and as such there is a potential to increase processing capacity by utilizing these resources while vehicles are in short-term and long-term car parks, in roads and at road intersections. Such car parks and road segments can be transformed, through 6G networks, into vehicular fog clusters, or Fogbanks, that can provide processing, storage and sensing capabilities, making use of underutilized vehicular resources. We introduce the Fogbanks concept, outline current research efforts underway in vehicular clouds, and suggest promising directions for 6G in a world where autonomous driving will become commonplace. Moreover, we study the processing allocation problem in cloud-based Fogbank architecture. We solve this problem using Mixed Integer Programming (MILP) to minimize the total power consumption of the proposed architecture, taking into account two allocation strategies, single allocation of tasks and distributed allocation. Finally, we describe additional future directions needed to establish reliability, security, virtualisation, energy efficiency, business models and standardization.