Date Added: Aug 2010
Accurately modeling server power consumption is critical in designing data center power provisioning infrastructure. However, to date, most research proposals have used average CPU utilization to infer the power consumption of clusters, typically averaging over tens of minutes per observation. This paper demonstrates that average CPU utilization is not sufficient to predict peak power consumption accurately. By characterizing the relationship between server utilization and power supply behavior, the authors can more accurately model the actual peak power consumption. Finally, this paper introduces a new operating system metric that can capture the needed information to design for peak power with low overhead.