The Multi-access Edge Computing (MEC) paradigm allows several automotive applications to be offloaded from the vehicles to the edge. Besides a higher computation capability, compared to the on-board vehicle, and the shorter latency, compared to the remote cloud, the edge offers additional (context) information that is not directly available at the vehicle, e.g., via data fusion from multiple sources. In this paper we propose a high-level architecture for MEC-assisted platooning control. Within the architecture, the longitudinal controller is conceived as a virtualized application running on an edge server, and aligned with the European Telecommunications Standard Institute (ETSI) MEC reference framework. Performance assessment conducted through a realistic simulation framework, coupling a vehicular mobility simulator and Docker containers, showcases the feasibility and effectiveness of our proposal.