Cost and QoS-Aware Service Migration in Vehicular Edge Computing Networks with Limited Backhaul Capacity

Vehicular edge computing (VEC) is a key enabler for providing low-latency vehicular service, by deploying the services at the edge servers associated with wireless nodes distributed in vehicle networks. Due to limited coverage of wireless nodes and high vehicle mobility, vehicular services need to be adaptively migrated over the backhaul network to follow the vehicles. Given an ever-increasing amount of vehicles served by finite backhaul bandwidth, the backhaul network can become the bottleneck for timely service migration. However, few studies on service migration considered optimizing the utilization of backhaul network bandwidth to alleviate such bottleneck. In this study, we focus on the service migration problem in VECenabled small cell networks with distributed wireless backhaul. Specifically, we formulate the service migration in VEC networks as an optimization problem with the objective of minimizing the migration cost while guaranteeing the desired quality of service (QoS), taking into account the limited time, bandwidth resource, and computation capacity. We develop an efficient service migration algorithm, called OMELET, which dynamically makes service migration decisions based on Lyapunov technique via jointly optimizing migration decision, path selection and rate allocation. Simulation results validate that OMELET can achieve a more effective trade-off between migration cost and QoS guarantee compared with the state-of-the-art service migration schemes.