Understanding PUE is a Big Step in Federal Sustainability Compliance

January 13, 2016 Tony Evans

 

On top of complying with the 2010 Federal Data Center Consolidation Initiative (FDCCI) requirements, the 2015 Executive Order 13693 adds additional conditions for reduced energy consumption that must be met.

Implementation instructions are contained in a 77-page document issued by The White House Council on Environmental Quality/Office of Federal Sustainability, but the bottom line is, moving forward, no federal data center can operate with a power usage effectiveness (PUE) higher than 1.5.

Understanding PUE

Since, you can’t control something you have not measured, the first step in managing efficiency is knowing and understanding your PUE, which is defined as: the ratio of total data center input power to the IT load power. It’s an indication of how much power is going to the IT load versus “everything else” in the data center.

The benefits of determining data center infrastructure efficiency as part of an effective energy management plan are widely recognized and the metrics of PUE (and its reciprocal Data Center Infrastructure Efficiency (DCIE)) have emerged as recognized standards.

But therein lies the complexity. Data centers are often part of multi-purpose buildings, so it can be quite challenging to find either a single point to measure the total data center input power or to measure the IT load.

My colleague, Victor Avelar, Senior Research Analyst at Schneider Electric, explained the dilemma in White Paper 158, “This means that the appropriate measurement points to capture all data center energy use need to be identified, and typically the power use data from different subsystems needs to be combined to find the total data center input power. To further complicate the matter, some of these power measurements may be impractical to obtain, or the devices may be shared with non-data center applications.”

Methodology in Three Parts

The kinds of equipment, network of devices and maze of subsystems vary depending on the data center itself, so in order to be accurate there must be a common denominator for measurement. No official standard currently exists, but the associated problems are recognized as standard by organizations that create industry guidelines.

Based on this premise, the paper goes on to outline a three-part methodology, which can be used as a standard approach to address the challenges of measuring PUE.

  1. Establish a standard to categorize data center subsystems as either

(a) IT load

(b) Physical infrastructure or

(c) Not included in the calculation

  1. If a subsystem’s power consumption cannot be directly measured because it is shared with non-data-center loads, estimate the power using a standardized methodology for that subsystem type.
  1. If a subsystem’s power cannot be directly measured because of technical barriers to measurement, estimate the power using a standardized methodology for that subsystem type.

 

I have presented an overview here, but read the Guidance for Calculation of Efficiency (PUE) in Data Centers White Paper to get the full picture. It provides a list of the various data center subsystems that need to be included in energy use analysis, shows how the energy use of these subsystems should be incorporated into PUE calculations and describes in more detail the practical approaches you should use to get a proper measurement.

 

The post Understanding PUE is a Big Step in Federal Sustainability Compliance appeared first on Schneider Electric Blog.

 

Read more...

Previous Video
Increasing Energy Efficiency, Cardiff University
Increasing Energy Efficiency, Cardiff University

Managing the cooling cost and decreasing energy consumption at Cardiff University's HPC and general IT-use ...

Next Video
InfraStruxure - Simple, Adaptable, Manageable
InfraStruxure - Simple, Adaptable, Manageable

InfraStruxure is integrated, rule-based system architecture that dramatically reduces time and complexity f...

×

Become a Data Center Expert - get the latest content before anyone else

Thank you
Error - something went wrong!