User Guide

Consumption
Consumption is a measure of how much of a particular metric is being consumed. For example, a 60-watt light
bulb has a power consumption of 60 watts.
Faceplate - Metric information as provided by the manufacturer. This is the maximum consumption of the device
when it is fully loaded with modules.
• User-Defined - Metric information provided by the user, either broadly on the asset type or specifically on a
particular asset. This is commonly referred to as derating.
• Measured - Metric information provided by devices and sensors in the data center and regarded as the most
accurate consumption metric. Measured consumption is provided through integration with certain
technologies.
• Aggregate - The consumption value for a floor level asset is the consumption for the asset itself, aggregated
with the consumption of all of its contained assets. In many cases, the user-defined and measured data are
pre-aggregated. This means you do not aggregate them with contained assets again.
Consumption details
When computing the consumption of a floor level asset, there are two primary factors to consider.
Consumption metric policy: Many consumption metrics are available. The consumption metrics are in order of
priority.
• Measured Consumption - This is considered the most accurate source of consumption data.
• User-Defined Consumption (Asset) - This is the next best source for consumption information. When defined
on an asset, it is likely that you obtained the information through your own measurements. When defined on
an asset, you can take into account configuration and usage.
• User-Defined Consumption (Asset Type) - When exact consumption data for a particular asset is not
available, then user-defined consumption for the asset type is the best source.
• Faceplate Consumption - When measured consumption is not available, resort to user-defined data (first on
the asset, then on the asset type). When none of these are available, resort to the faceplate data.
Aggregate Calculations Policy: When the chosen metric does not represent the aggregate consumption of the
asset and its contained assets, then the metric should be aggregated with the consumption metrics of its
contained assets at compute time.
• If the metric includes the consumption of the contained assets and you measure power at the power strip
feeding a rack, then you should enter a user-defined consumption value for that asset and disregard any
consumption values of contained assets.
• If the metric does not include the consumption of the contained assets and you measure the power
consumption of an empty blade chassis, and then measures the incremental power consumption of adding
additional blades, you should set user-defined consumption values for the blade chassis and blade asset
types. As blades are populated into the blade chassis, power consumption is aggregated to provide the
overall consumption of the populated blade chassis.
• Existing faceplate data works well by aggregating consumption metrics. However, faceplate metrics for
modules is usually not available. Instead, the faceplate metrics for the containing asset is the maximum
possible if it is fully loaded with modules. To account for this, metrics should indicate whether they are an
aggregate metric or not. All faceplate data can be assumed to be non-aggregate.
80 Data Center Planner Installer/User Guide