WP#76 - Power Metrics for ITE

21 December, 2017 | White Paper

Editor: Phil Isaak, Individual Member Contributors: Robert Bunger, Schneider Electric; Jay Dietrich, IBM; Donald Goddard, NetApp; James Halpin, Sky UK; Robert Landstrom, Interxion

This paper describes the outcome of an initiative undertaken by The Green Grid in early 2017 (WI #17-002) to define a set of metrics that can quantify the power demand and energy consumed of ITE platforms within a data center.  The ITE systems are defined as the compute processing, storage and network hardware within the data center. This definition of ITE systems aligns with The Green Grid’s WP#72 tuples that define ICT capacity and utilization metrics. The ITE platforms are a subset of equipment within each ITE system to differentiate between various type of equipment by function, form factor, class, generation, etc. The power demand will provide an indicator of the share of the total capacity that is actually utilized (“ITE power utilization”).  The energy (“ITE energy”) will provide an indicator of the total electrical energy consumed over a defined period (typically one year).  Trending the power demand over a multi-year timeframe can serve as a useful indicator for future capacity requirements (capital expenditures), while trending energy over a multi-year timeframe can serve as a useful indicator for future operational expenditures.  ITE power utilization also serves as a useful proxy for the energy efficiency with which the electrical distribution is supporting the ITE platforms because sub-systems within the electrical distribution are generally more energy efficient at higher utilization. The total power demand metrics across all ITE deal with expected maximum, which means that the peak and minimum demand of a single ITE device may be substantially different at any given time. The goal of these metrics is to better track and communicate how particular ITE platforms are consuming power and energy in pursuit of efficiency, utilization, and the ability to facilitate forecasting future capacity and energy requirements in specific IT data centers.