“…1 Numerous other research works are exploring alternative approaches to develop energy-efficient systems across all the layers of the hardware-software-application/service stack-from processors, servers, and energy management systems to applications and service pricing policies-using modeling, simulation, experiments or observation. In particular, research and industrial efforts in computer architecture are focusing on the design of energy-efficient processors, accelerators, and servers, 13 which can reduce their power consumption by switching-off completely during periods of inactivity or by switching between high and low-power states in response to fluctuating workloads, 7,64,116,[118][119][120] provisioned power constraints ("power caps"), 121,122 or intermittent availability of "green" energy. 123 In the area of data center middleware, many studies focus on the energy-efficient management of a data center's workload using various algorithms from machine learning, 19,124,125 self-adaptive scheduling and optimization, 126 to resource allocation, task placement and scheduling.…”