The Green Grid Expands The ‘PUE’ Metric
Integrating measures for efficiency, capacity, and availability of data center cooling.
Once again The Green Grid (TGG) is at it, with new and better metrics for the improvement of your data center cooling performance. Their most recent standards and best practices are truly leading the industry to achieve optimal total cost of ownership performance while ensuring that the necessary levels of IT availability and facilities uptime are maintained. Best known for their introduction and promotion of the power usage effectiveness (PUE) energy efficiency measure, they have established several other meaningful performance metrics to the industry as well, and have earned a lasting reputation as possibly the greatest group of thought leaders ever to address the global data center community.
As many of us know, TGG is a nonprofit, industry consortium of endusers, policy-makers, technology providers, facility architects, and utility companies collaborating to improve the resource efficiency of data centers. Between April 2006 and April 2007, the world’s largest data center technology companies and vendors came together to create TGG.
Since then they have enjoyed accolades following the global adoption of PUE as the most meaningful metric of data center energy efficiency anywhere. They have become a real think tank, churning out a multitude of meaningful standards and guidelines, including one of my favorites, “The Data Center Maturity Model.”
At their most recent symposium in Seattle in March of this year TGG announced a change in leadership and reassessment of their direction with a focus on “The Economics of Sustainable IT.” There they brought together data center practitioners from across the globe to challenge the theme of effective and accountable resource efficiency. They really did a great job of examining the benefits and challenges of pursuing sustainability while remaining focused on emerging solutions for achieving data center efficiency. They also provided a thorough review of soon-to-be-released technically progressive projects from TGG, including:
Data Center Utilization and Capacity
DCMM 2.0 – Revision of the Data Center Maturity Model
Data Center Automation with a DCIM System
Liquid Cooling Technology Update
Historically, standards and guidelines like these have been made available to members only, but many can now be found available to the public on their website at http://bit.ly/29L0jJk.
Examples of these recent white papers and works include:
Data Center Cooling Performance Indicator (June 2016)
Analysis of the Server Efficiency Rating Tool (Nov. 2015)
Space Usage Effectiveness — SpUE (Dec. 2014)
Data Center Power Systems Harmonics (Oct. 2013)
Electronics Disposal Efficiency: An IT Recycling Metric for Enterprises & Data Centers (March 2013)
Data Center Life Cycle Assessment Guidelines (Nov. 2012)
Power Usage Effectiveness: A Green Grid Data Center Sustainability Metric — PUE (Oct. 2012)
They are absolutely worth checking out.
The Performance Indicator: Assessing and Visualizing Data Center Cooling Performance
The newest TGG white paper introduces the Performance Indicator (PI), a takeoff on the original PUE metric that provides data center operators with a tool to measure and improve cooling capacity and availability, as well as efficiency. It is probably best thought of as a way to holistically assess and visualize how closely a cooling system performs relative to the original design intent. By addressing availability, capacity, and sustainability simultaneously, the new PI helps maintain safe and suitable thermal IT operations without losing sight of efficiency.
According to Mark Seymore, lead developer of the PI concept, there are three key performance measures along with four levels of assessment to be considered when using the new tool. The three performance measures, when considered simultaneously, provide a visualization and calculation method for continued improvement of cooling performance, while the four levels of assessment allow operators to establish a complete and accurate view of the facility through predictive calculation of both current and future environmental states.
The three metrics for efficiency, capacity, and availability are defined in the PI model as follows:
The Efficiency Metric: “PUE ratio,” PUEr(X) = PUEref(X) ÷ PUE actual as the ratio of design PUE divided by measured PUE, defining the percentage of cooling used as intended by design.
The Capacity Metric: “IT Thermal Conformance” as the percentage of IT equipment operating within the recommended inlet temperature spec at a given IT load and distribution and during normal operations.
The Availability Metric: “IT Thermal Resilience” as the percentage of IT equipment operating within the allowable inlet temperature spec during a cooling failure or planned maintenance activity, where redundant cooling units are offline.
As a rule of thumb for each metric above, the higher the calculated percentage the better the cooling performance. (Note that more explicit definitions for each variable are available at the link below in this column.)
As you can see in Figure 1, any combination of results for a given load state will define a “performance triangle.” This triangle defines a unique operating status and measure of overall performance. The ideal performance is where each point of the triangle reaches 100% simultaneously to make a full and perfect triangle, and that would likely occur at 100% load with excellent airflow and temperature controls.
According to TGG, the PI may be used to:
Visualize the balance between the three metrics
Assess a facility’s performance in relation to the facility’s target range
Track a facility’s progress over time, as the load grows and as changes to the facility are implemented
Assess effects of changes before implementation, from IT deployments to installation of containment
Compare alternative configuration options
The above PI model is intended to help an operator assess data center cooling performance for any number of current and future operating states. In order to effectively perform the assessments, TGG recommends the use of several tools to monitor, predict, and calibrate the data to be collected, and their white paper outlines four levels of assessment using some or all of the tools at each level.
Assessment Level 1 is the most fundamental assessment of current operations based upon rudimentary measurements of temperature and power consumption taken from room level monitors at HVAC equipment supply and return points and readings on UPS system displays.
Level 2 is a more accurate assessment of current operating conditions using advanced temperature and power monitoring methods to take data at the chassis and IT equipment levels from environmental and branch line sensors, thermal images, and even SMPT network data taken from the servers.
Level 3 is a basic predictive analytics assessment using computational fluid dynamics models to determine how airflow and temperature can vary with changes in the “power, space, and cooling” infrastructure.
Level 4 is an advanced computational fluid dynamics model that is verified and calibrated to represent the real time characteristics of the facility environment while in operation.
Current operations are analyzed using data collected by temperature monitoring systems such as those found in a data center infrastructure management (DCIM) system suitable for monitoring of environmental and power utilization trends. Future operations are assessed by using data developed with sophisticated predictive analytics tools for environmental and power performance, such as computational fluid dynamics (CFD) models.
The highest level of assessment uses a holistic model where simulation tools are checked against and calibrated with real time data taken from the operating data center environment. This model defines what may be the most accurate and reliable tool you could develop to maintain and improve your data center cooling performance.
The completed tool provides a living model of your data center with the ability to help you develop and manage both specific improvements and an effective cooling strategy over time. Some of these predictive “what-if” scenarios that can benefit your operations are listed below as:
Improving PUE by increasing server inlet air temperatures
Managing power utilization for a variety of different IT loads and load distributions
Knowing the effects of change before implementing change
Achieve higher availability with a higher certainty of capacity and more efficiency and savings
For those of you interested in seeing how the systems works for different situations, TGG has made a copy of their webinar presentation available to the public on the TGG web site at http://bit.ly/29L19Wk.
The webinar will take you through several examples and case studies that explain how to use the new tool to predict overall cooling performance while planning for change in the form of capacity and operational improvements.
According to the author, the performance indicator has been defined in terms flexible enough to allow a user to add metrics and to develop a graphical model in the shape of a square or a polygon. An example of adding metrics could broaden the aspect of cooling performance with measures of water utilization effectiveness and of air quality and humidity. Another example might be to add measures of the capacity, availability, and efficiency of electrical distribution systems to provide a broader perspective of data center PUE performance.
Credits for the webinar and white paper go to Mark Seymour and Maira Bana of Future Facilities, David Wang of Teradata, Danny Cummins of Siemens, and Veerendra Para of IBM. For information related to this work contact Sarah Otterstetter at 781-418-2416 or firstname.lastname@example.org.
The Green Grid Membership
TGG’s mission is to drive accountable, effective, resource-efficient, end-to-end IT ecosystems in data centers. They do this with a well-defined process that provides frameworks for organizations to realize operational efficiency and maturity across the IT infrastructure. TGG’s vision for their membership is to enable them to improve operations, mitigate risk, enhance thought leadership, and strengthen the value chain as it relates to energy/asset recovery and resource efficiency. Information related to membership can be found at URL http://www.thegreengrid.org/become-a-member.aspx.
Critical Facilities Roundtable
Critical Facilities Roundtable (CFRT) met with the German American Chamber of Commerce on November 3, 2015, in Mountain to view Germany’s recent award winning data center products and services including superconductor bus-bar, low temperature adsorption chillers and global DCIM applications. CFRT is a non-profit organization based in the Silicon Valley that is dedicated to the sharing of information and solutions amongst our members made up of critical facilities owners and operators. Please visit our website at www.cfroundtable.org or contact us at 415-748-0515 for more information.