Service Migration in Edge Computing Environments for Connected Autonomous Vehicles

In Connected Autonomous Vehicles scenarios or CAV, ubiquitous connectivity will play a major role in the safety of the vehicles and passengers. The extensive amount of sensors in each vehicle will generate huge amounts of data that cannot be processed promptly by onboard units. Edge computing is a crucial solution to provide the required computation power and extremely low latency requirements for the future generation of CAVs. However, the high mobility of vehicles, together with dynamic 5G networking scenarios, poses a challenge to keep the services always close to the users, and therefore, keep the latency very low, such as expected by CAVs. In this paper, we propose MILT, a service migration algorithm for edge computing to perform predictive migration of services based on mobility prediction, available resources, and the quality level of the networks and applications. MILT supports a mobility-based handover prediction scheme to perform a pre-migration to the best available edge server while reducing the latency and increasing the processing capacity of the services of CAVs. Simulation results show the efficiency of the proposed algorithm in terms of latency, migration failures, and network throughput.