Brandon Aaron Rubenstein

Data center cold aisle set point optimization through total operating cost modeling

Operational expenses of servers and the data center to support them account for a significant part of the total cost of ownership of a data center and potentially the total business. Manipulating the inlet temperature to the servers can provide a possible method of lowering the operating cost of a data center. Conventional wisdom regarding management of the cold aisle temperature in the data center has traditionally been that colder temperatures kept the failure rate of the equipment down and higher temperatures reduced energy costs. While it is true that in most cooling strategies, increasing the cold aisle set point temperature reduces the power consumed by the facilities, the overall cost savings may not be realized due to increases in server power and decreased reliability. An investigation using the total operational expense of the data center would be required to find the optimum operating point of the environmental controls. The simulation described in this paper demonstrates that when accounting for total operational cost including component reliability and servicing in addition to energy and water resource costs, there is an optimal data center cold aisle set point. This point will vary with resource rates and climate variations. The run cost effectiveness metric, RCE, allows data center designers and managers to monitor and optimize their data center for the given static costs of the region. Its a measurable value for many data centers that can break out these costs either in real time, or more likely tabulated form collected from monitoring data over time.