5G AutoMEC – Boosting edge-to-edge service continuity for CAM in a sliced network

Network function virtualization and edge computing in mobile networks are key enablers in the evolution of recent standards for a 5 Generation (5G) mobile communication system towards a comprehensive 5G ecosystem, which enables the deployment of customized networks and service access for various industry verticals by means of clearly defined and deployed network slices. In particular, the automotive industry can leverage low-latency access to services hosted along the distributed network edge, in support of a large variety of use cases. Network slices typically implement the requested service and network according to a set of defined requirements, as well as performance and latency bounds, while using a defined resources budget. With connected cars following different mobility patterns, the automotive sector represents a pretty agile customer of such network slices. This makes the management of network and edge computing resources a key challenge to tackle in order to balance the resource utilization, and the previously agreed and expected service levels. In this paper, we analyze the benefit of smart mobile edges and the use of machine learning to anticipate resource demand at distributed mobile edges in an automotive scenario. Finally, we experimentally show the feasibility to treat μ-slice resources efficiently by using an OSS-based prototype for the orchestrated edges, which is currently being developed for an automotive trial in Europe.