An NFV-Based Edge Platform for Low-Latency V2X Services Supporting Vehicle Mobility-Driven Auto Scaling

摘要

To guarantee low-latency and reliable end-to-end connectivity and provide suitable resources for vehicle-to-everything (V2X) services, the network function virtualization (NFV)-based multi-access edge computing (MEC) technology has been recognized as an effective solution. In the literature, few studies have proposed corresponding NFV-based MEC frameworks to provide V2X services. However, these works only discuss the concept of applying the NFV to the MEC to offer V2X services without implementing it in a real testbed. On the other hand, within the NFV-based MEC framework, several studies have investigated the NFV auto scaling mechanism to provide suitable resources for different services. However, these works are not suitable for V2X services, which exhibit unique traffic characteristics. In this paper, we (1) implement an NFV-based edge platform to provide vehicles with low-latency V2X services, (2) implement a vehicle mobility-driven NFV auto scaling mechanism that forecasts the resource requirements for V2X services and triggers the scaling operations if needed, and (3) use the open source software to build an NFV testbed and deploy our platform as virtual network functions to the NFV testbed. Finally, we implement a simple V2X service to validate our proposed edge platform, justifying that it, together with our NFV auto scaling mechanism, requires affordable processing delay and supports low V2X service latency.