Temporal Power Model for Effective Usage in Data Center

Data center is considered as major contributor for CO2 emission in ICT sector. Power optimization approaches supported with power usage metrics were proposed and discussed in literatures. Recently, engaging green energy resources to power data center were proposed toward reducing data centers´ carbon footprints, called renewable-based data center. Apparently, most of these proposals were mainly considering various techniques to predict the renewable availability. Although utilizing green resources in data center would reduce the carbon footprints, the green energy has its temporal limitation on the around the clock operation. In this short paper, we are proposing a temporal power model for green energy usage in data center. The model is combining the power usage in data center and the green power supplied. Using the total derivative, it rates the instantaneous change of carbon footprints. A mathematical approach using the total derivative was applied to a PUE-based power model to achieve the instantaneous green cost of a request in data center. This model is applicable to accurately capture the power usage and carbon footprints per request. In addition, our model is expandable to precisely calculate the per request green incentive.