Mobility has always been a big challenge in cellular networks, because it is responsible for traffic fluctuations that eventually result into inconstant resource usage and the need for proper Quality of Service management. When applications get deployed at the network edge, the challenges even grow because software is harder to hand-over than traffic streams. Cloud technologies have been designed with different specifications, and should be properly revised to balance efficiency and effectiveness in distributed and capillary infrastructures. In this paper, we propose some extensions to OpenStack for power management and Quality of Service. Our framework provides additional APIs for setting the service level and interacting with power-saving mechanisms. It is designed to be easily integrated with modern software orchestration tools and workload consolidation algorithms. We report real measurements from an experimental proof-of-concept.