Computation Offloading in the Internet of Connected Vehicles: A Systematic Literature Survey

Nowadays, there is a rapid development in vehicles world. Vehicles are equipped with smart systems as well as infotainment applications. But such systems consume vehicles' computation or storage capacity. However, when the vehicle encounters a computation and/storage hungery applications or near real time applications that need high Quality of experience (QoE), it must offload it, either partially or entirely, to a more powerful and resourceful entity. At the beginnings this entity was a remote cloud. Although clouds are powerful in terms of computation and storage capacities, the process of task offloading to a remote cloud consumes the network bandwidth, which is not suitable to delay sensitive applications. As a solution, researchers propose to use cloudlets as third entity closer to the network edge. This will make the offloading much faster, but unfortunately due to the fact that cloudlets less computation and storage capacity than clouds, offloading will cause resource starvation. These factors motivate the appearance of Vehicular Cloud Computing (VCC). VCC proposes collecting the on-board units of multiple vehicles to form an on-ground cloud. This allows vehicles to offload their computational task to other vehicles in the vicinity. In this paper, we first provide a summery on concepts that are related to edge computing and task offloading process, and then we review a set of papers that use different approaches to execute computation offloading and scheduling.