A recent study by the National Mining Association and American Coalition for Clean Coal Electricity revealed that global IT infrastructure consumed 10% of total energy produced by greenhouse gases in 2013. As power densities continue to skyrocket, data center managers must increase cooling resources to match soaring IT loads. Given that cooling resources account for roughly 60% of total energy consumed in data centers, one might think cooling efficiency would be high on the priority list of stakeholders. On the contrary, most data centers suffer from a silent, ubiquitous epidemic of overcooling that is driven by fear of thermal-related downtime. As HP Labs Fellow Chandrakant Patel put it, overcooling is a costly method of hotspot mitigation that is akin to "using a chain saw to cut butter."
In most cases, cooling resources are overabundant but distribution is not uniform due to poor airflow conditions. In the data center, airflow is the key player behind cooling delivery. For the most part, thermal problems in data centers are localized to particular racks outside of the flow range (CFM) of CRAC units. The cure for an overcooled data center is simple; it's called Intelligent Airflow Management – a real-time cooling delivery solution that adapts airflow to meet IT demand at the rack level. By controlling airflow in your data center, you can reduce operational costs and wasted energy resources by up to 40% without sacrificing thermal safety or server function. At AdaptivCOOL, we use a comprehensive system of networked hardware and software called Demand Based Cooling (DBC) to intelligently manage cooling resources in data centers.