Metrics
"Miles per gallon", "dollars per pound", "income per capita", ... metrics are part of every day life and conversation. Metrics enable evaluation, and evaluation enables decision-making. Business executives often say that "you can't manage what you don't measure." [download] There is no single correct metric. The appropriate choice of metric depends on the intended use.
For data centers, metrics can be physical (e.g. energy use per unit floor area) or economic (e.g. energy costs per unit floor area). They can apply to the whole-facility, a system, subsystem, or single component. As discussed elsewhere in this site, design intent documentation is enhanced by the assignment of metrics that help validate whether or not a best practice has in fact achieved its intended goal. Examples follow:
Whole-Building Metrics
- Total electrical power demand per unit of floor area (W/ft2 [example], kWh/ft2-year, MBTU/ft2-year, $/ft2-year)
- Application can vary, e.g. based on gross floor area versus "loaded" floor area
System Metrics
- HVAC effectiveness — a normalized efficiency metric (kW cooling/kW servers) [example]
- Mechanical: Airflow Management (return air temperature)
- Mechanical: Air Handler Systems (fan power Watts / CFM under typical conditions; Economizer Presence and Lockout Band)
- Mechanical: Humidification (kWh or therms per pound of humidification)
- Mechanical: Plant Optimization (kW/ton)
- IT Equipment Selection (watts cooling device / rack watts cooled)
- Electrical Infrastructure (utilized electricity and heat / input fuel Btus)
Component Metrics
- IT Equipment Selection (power supply efficiency, %)
- Electrical Infrastructure (UPS efficiency)
- Fan efficiency (%)
- Server (various measures of energy performance)