Design and management of data center effectiveness, risks and costs

Data centers have been cooled for many years by delivering cool air to the IT equipment via the room. One of the key advantages of this approach is the flexibility that it provides the owner / operator in terms of equipment deployment. In principle it seems that all that is necessary is to determine the maximum power consumption of the equipment and provide an equivalent amount of cooling to the data center. Why then, since we have been building and operating data centers for decades using air cooling, do data centers experience hot spots, operate inefficiently and fail to reach their design expectations for capacity? This paper explains some of the challenges faced in the search for the perfect data center. In particular it identifies why, given the variability of equipment design and the time varying nature of the data center load, Computational Fluid Dynamics (CFD) should be used for airflow and heat transfer modeling. It shows that although it is not the only tool to be used in design and / or management of a data center, it is an essential tool to avoid lost capacity due to inability to efficiently cool equipment and the resulting potential for overheating. Important themes include a strategy that encompasses the entire life cycle of the data center, the need for an appropriate level of detail and the critical requirement for model verification and calibration.