Software-defined vehicular networks: A cooperative approach for computational offloading

Conventional vehicular systems are gradually evolving into intelligent transportation systems. Multi-access Edge Computing (MEC) has become an important component of the software-defined vehicular networks that benefit from the introduction of reliable and low latency 5G networks and Wi-Fi. MEC offers a variety of smart services closer to the access networks leveraged by vehicles. The intelligent and connected vehicles demand access to compute-intensive applications that require compute and storage resources in addition to what is available in vehicles today. The computational burden on the connected vehicles can be significantly reduced by offloading to the MEC nodes and to the cloud. However, existing computational offloading schemes are challenged by the fast-moving vehicles, frequent handovers and subtle differences to the scenarios designed to support mobile phone handovers. The paper presents a software-defined vehicular edge computing architecture that copes with the mobility challenges by utilising the Software Defined Networking paradigm to perform traffic flow management for information collection and intelligent management of connected vehicles and networks.