A unified approach to coordinated energy-management in data centers

Energy consumption has become a critical issue for data centers, triggered by the rise in energy costs, volatility in the supply and demand of energy and the widespread proliferation of power-hungry information technology (IT) equipment. In response, researchers are developing energy-efficient data centers by incorporating energy-aware systems both at the IT level (shutting down servers or temperature-aware workload placement) as well as at the facilities level (shutting down air-conditioning units, increasing temperature of air supplied to the plenum). In this work, we explore a novel approach to coordinated-management of IT systems and its cooling infrastructure to joint power and temperature objectives. In particular, for a given total IT workload, we present a method to determine the optimal settings of computer room air-conditioning units (CRACs) in a data center so as to minimize overall energy consumption in the data center while satisfying specified temperature constraints. Using potential fluid-flow theory, our approach identifies distinct thermal zones associated with each CRAC and then provisions cooling power to match the heat generated from the racks in that zone. We illustrate the resulting range in behavior and potential for energy savings in a large 10,000 sqft commercial data center with 10 CRACs and 186 racks.