data center maturity model - CiteSeerX

WHITE PAPER #36

DATA CENTER MATURITY MODEL LEAD EDITOR: Harkeeret Singh, Thomson Reuters

CONTRIBUTING EDITORS: Dan Azevedo, Symantec

The Green Grid Advisory Council

David Ibarra, DPR Construction

The Green Grid Technical

Rona Newmark, EMC

Committee

Steve O’Donnell, Enterprise

The Green Grid Compute Sub Work

Strategy Group

Group

Zeydy Ortiz, IBM

The Green Grid Thermal

John Pflueger, Dell

Management Work Group

Nik Simpson, The Burton Group

The Green Grid Power Sub Work

Victor Smith, Dell

Group

PAGE 2

Executive Summary The Green Grid (TGG) is a global consortium of companies, government agencies, and educational institutions dedicated to advancing energy efficiency in data centers and business computing ecosystems. With input from representatives around the world, The Green Grid has developed the Data Center Maturity Model (DCMM), which provides clear goals and direction for improving energy efficiency and sustainability across all aspects of the data center. The DCMM touches upon the major components of the data center, including power, cooling, compute, storage, and network. The levels of the model outline current best practices and a five-year roadmap for the industry. The DCMM provides capability descriptions by data center area such that operators can benchmark their current performance using the Data Center Maturity Model Equalizer, thereby determining their levels of maturity and identifying the ongoing steps and innovations required as part of their data center and IT strategy to achieve greater energy efficiency and sustainability improvements, both today and into the future. The paper discusses the motivations for creating the model, the participants in its development, and its structure, terms, and interdependencies. In addition, the paper offers insight as to the methods and benefits of using the Data Center Maturity Model, and it addresses frequently asked questions. Ultimately, the model aims to help data center owners/operators worldwide not only identify and carry out energy efficiency and sustainability improvements in their own facilities, but also collaborate with others throughout the industry to innovate in order to achieve a common goal.

2011 The Green Grid. All rights reserved. No part of this publication may be used, reproduced, photocopied, transmitted, or stored in any retrieval system of any nature without the written permission of the copyright owner.

PAGE 3

Table of Contents I.

Introduction ......................................................................................................................................................... 3

II.

Development of the Data Center Maturity Model ............................................................................................. 3

III.

How and Why to Use Data Center Maturity Model?.......................................................................................... 8

IV.

Frequently Asked Questions ............................................................................................................................ 10

V.

Next Steps ......................................................................................................................................................... 11

VI.

Conclusion ......................................................................................................................................................... 12

VII. References ........................................................................................................................................................ 12 VIII. About The Green Grid ....................................................................................................................................... 13

I.

Introduction In 2010, The Green Grid (TGG) noted gaps in the worldwide market with regard to effective ways for data centers to enhance energy efficiency and sustainability. TGG identified the need for a comprehensive model of what could and should be done over time throughout the data center to improve overall energy efficiency and sustainability. In response, The Green Grid developed the Data Center Maturity Model (DCMM) and supplied this paper to clarify and support the use of the model. The DCMM provides clear capability descriptions by data center area so that data center owners/operators can benchmark their areas’ current performance, determine their levels of maturity, and identify the ongoing steps necessary to achieve greater energy efficiency and sustainability, both today and into the future.

II.

Development of the Data Center Maturity Model The Green Grid assembled an international task force to lead the effort of crafting a model for use worldwide. The task force solicited input from a variety of subject matter experts throughout The Green Grid. The task force sought guidance, ideas, and improvements from these individuals, asking them to pay particular attention to their areas of expertise. The task force then presented the DCMM to the full TGG member audience for review and input. As a result, the model gained detailed descriptions for multiple content areas, expanding it beyond the industry’s traditional focus on facilities to include Compute, Storage, Network, and others in the realm of IT.


PAGE 4

DATA CENTER MATURITY MODEL ELEMENTS The task force opted to structure the model around set levels with clear definitions in each content area. These definitions give data centers a basis for comparing the energy-efficiency sophistication of their own operations with that of the typical enterprise. The model’s levels reflect the collective experience of the task force and the wider TGG membership, including members’ involvement in other organizations and initiatives around the world.

Figure 1. Data Center Maturity Model Click here to download the DCMM

Levels of Data Center Maturity Level 0: Minimal/No Progress

The initial levels chart the progress of a typical data center that

Level 1: Part Best Practice

has taken no efficiency improvement measures to one

Level 2: Best Practice

exemplifying the state of an average data center to one that

Levels 3 and 4: Reasonable Steps

employs current best practices. Levels 3 through 5 represent

(between current best practices and

future capabilities toward which the industry should collectively

the visionary five-year projection)

move and innovate. An optimistic—but by no means

Level 5: Visionary – Five Years Away

prescriptive—timetable for achieving Level 5 maturity is five years. 2011 The Green Grid. All rights reserved. No part of this publication may be used, reproduced, photocopied, transmitted, or stored in any retrieval system of any nature without the written permission of the copyright owner.

PAGE 5

Over time, The Green Grid expects organizations to progress through the levels of the maturity model. An illustration of progress can be found in Figure 2 below; the expectation is that state-of-the-art data centers (green line) will progress from Level 2 today to Level 5 by 2016, typical data centers (amber line) will move to Level 3, and underperforming data centers (red line) will lag at Level 2. The Green Grid encourages all organizations to move through the levels of the maturity model as soon as feasibly possible, taking into account business and site constraints.

Figure 2. Illustration of progress through the levels of the maturity model for the different types of data centers The DCMM operates within the axes of Data Center Efficiency & Sustainability and Investment (Financial, Time & Resource). These content areas align with the way many organizations are structured, which means that senior managers can easily separate the model into logical sections—Power, Cooling, Compute, Storage, Network, etc.—and send their internal teams the sections that most apply to them (power/cooling issues addressed by the data center team, compute assessments and improvements handled by the server team, storage initiatives passed to the storage team, and so on). These existing teams can use the DCMM to benchmark their operations, inspire improvement, monitor progress, and continually advance. Senior managers can ask teams to report on progress against the levels, thus the DCMM provides a common framework for discussion. Each initiative within the model should be considered alongside other business factors, such as financial, time, resource, and site constraints. Dissecting the model and using its different parts to evaluate the facility and IT aspects of the data center enables managers to better understand which parts of the data center are excelling and which present the largest opportunity for improvement. This helps focus time and resources toward the areas with the greatest potential. Where data centers have already adopted energy efficiency measures, TGG would like the model to inspire them to take the additional steps to move toward the higher levels of the DCMM. 2011 The Green Grid. All rights reserved. No part of this publication may be used, reproduced, photocopied, transmitted, or stored in any retrieval system of any nature without the written permission of the copyright owner.

PAGE 6

The task force deliberately included the IT side of the data center (Compute, Storage, Network, and Other IT) in the model because, although power and cooling tend to get more attention when it comes to tackling energy efficiency, there are also significant improvements to be made in the IT arena. Server teams can use the Compute sections of the model to ask and answer questions such as, ―We have virtualized our servers, but what is our average utilization?‖ The answer may be, ―We are not aiming as high as we could in terms of target utilization.‖ A manager may then acknowledge that the data center could/should be doing better in that area, such as aiming for the 35% server utilization goal in Level 3.

TERMS AND DEFINITIONS The Green Grid’s glossary of terminology can be found at www.dcglossary.com/foswiki/Glossary/WebHome. Specific terms and definitions for the Data Center Maturity Model: Storage Auto-tiering. Auto-tiering is a process by which the system looks at the read and



write characteristics of the application and then, taking into account SLAs, uses a set of heuristics to decide on which tier of storage the data should reside. Compute Energy proportionality. Energy proportionality is a state where energy consumption scales directly



with use of the equipment. Power Additionality. In the context of low-carbon power generation, additionality refers to whether the



customer’s action or decision has made a genuine reduction in emissions above and beyond what would have happened anyway. Critical power path efficiency. Critical power path efficiency examines the efficiency of the power



infrastructure (excluding all mechanical infrastructure) from the entrance of the building to the IT equipment. (See Figure 3 below.)

Building Entrance

Input Switchgear & Distribution

UPS

Transformer (Optional)

PDU

IT Internal Power Supply

Critical Power Path May be placed outside the conditioned space

Figure 3. Example of a data center’s power infrastructure in which the dotted line denotes its critical power path. 2011 The Green Grid. All rights reserved. No part of this publication may be used, reproduced, photocopied, transmitted, or stored in any retrieval system of any nature without the written permission of the copyright owner.

PAGE 7



Low-carbon power generation. Low-carbon power is the term used to describe the energy produced from low-carbon energy sources. These include solar power, wind, wave, nuclear, tidal, solid renewables (wood, straw, and waste), gaseous renewables (landfill gas and sewerage gas), and hydroelectricity.

Cooling 

RCI. This stands for rack cooling index, which is a numerical value for air inlet temperature consistency and can be high or low (if applicable). For example, if every server is receiving air at the desired temperature, the result is an RCI of 100%. If half of the servers receive inlet air that is above the desired temperature, the RCI (high) is 50%.



PUE cooling contribution. This is the contribution made to PUE (power usage effectiveness) by the cooling and all its associated infrastructure. PUE contributions generally come from IT (1.0), cooling, power distribution, other sources such as lighting, or miscellaneous contributors. Everything above 1.0 (just the IT) represents the overhead of the data center infrastructure.



Mechanical/refrigerant cooling reduction. This refers to the amount of time when the site is not using any mechanical (compressors) or refrigerant cooling; instead, the site is generally using airside or water-side economizers or other free-cooling schemes. The cooling system will typically still require fans and pumps during these cooling modes.



Free cooling. Free cooling is any cooling method that uses naturally occurring local cooler fluids to cool without compressors or other processes such as absorption. Examples include the use of outdoor air, evaporative cooling in cooling towers, and using ground-sourced cool water as the final heat sink for energy dissipation from the data center. Free cooling is not completely ―free,‖ as it generally involves fans or pumps as fluid movers. An alternate consideration is that with the exception of fluid transport, the entropy in each subsequent process step increases.

Management 

ERE™. This metric represents the total data center energy minus reuse energy divided by IT equipment energy, the beneficial energy reuse that occurs outside the data center and its support infrastructure (more technically defined as the data center’s control volume).



WUE™. This is the annual site water usage divided by IT equipment energy, a site-based metric that is an assessment of the water used on-site for operation of the data center. This includes water used for humidification and water evaporated on-site for energy production or cooling of the data center and its support systems (similar to carbon Scope 1).



WUEsource. This metric is the annual source energy water usage plus annual site water usage divided by IT equipment energy, a source-based metric that includes water used on-site and water


PAGE 8

used off-site in the production of the energy used on-site. Typically, this adds the water used at the power-generation source to the water used on-site (similar to carbon Scope 2). 

CUE™. This is the total CO2 emissions caused by the total data center energy divided by IT equipment energy. Note that, regardless of the existence of on-site generation, CUE is a sourcebased metric and includes carbon generated in the full energy distribution string.

RELATIONSHIPS AND INTERDEPENDENCIES The nature of the data center is such that some of the goals in the DCMM content areas may not be achievable independently due to interdependencies. Cooling 

Airflow management and increasing temperatures are related. In order to increase temperature as per Level 3 to Level 5 of the Environmental section (―Set point range at inlet conditions to IT equipment‖), a data center would need to have air flow and temperature management in place. This would be achieved by implementing Level 1 and Level 2 of the Operational section— starting with configuring hot/cold aisles, filling floor gaps, optimizing tiles, installing blanking panels in the cabinets, and so on. To get to Level 5 on PUE cooling contribution or Level 5 on hours of mechanical/refrigerant cooling, it is likely that hot/cold aisle segregation, advanced architecture, etc. would be required.



Level 5 includes the ―No mechanical/refrigerant cooling (e.g., economization) for 100% of annual hours – 8,760 hours.‖ However, this would only be possible for data centers in warmer climates if the IT equipment can be operated at higher inlet temperatures, which is outlined as part of Level 5 of the Other IT Procurement section—―All IT equipment for the data center available to be operated continuously and warrantied at air inlet temperatures between 5°C/41°F and 40°C/104°F (and under exceptional conditions up to +45°C/113°F) and 10% to 80% relative humidity, non-condensing respectively.‖ If the inlet temperature maturity level is not met, the suggested hours for mechanical/refrigerant cooling reduction may be difficult to achieve.



Widening temperature and humidity parameters as per Environmental Levels 4 and 5 (―Set point range at inlet conditions to IT equipment‖) is likely to be more relevant for new data centers, as all IT equipment can be procured in alignment with the same temperature and humidity standards outlined as part of the Other IT Procurement section. Legacy sites would have a mix of equipment that is warrantied at different temperature and humidity parameters, making it difficult to implement Environmental Levels 4 and 5.

III.

How and Why to Use Data Center Maturity Model?

Not surprisingly, data center owners/operators who are interested in improving their data centers’ energy efficiency and sustainability will gain the most from the DCMM. Adopting it can help them use data center 2011 The Green Grid. All rights reserved. No part of this publication may be used, reproduced, photocopied, transmitted, or stored in any retrieval system of any nature without the written permission of the copyright owner.

PAGE 9

assets more effectively, thereby reducing both their capital and operational expenditures, as well as minimizing their environmental impact and improving their corporate/social responsibility standings. Data center owners/operators and senior managers can use the DCMM as a tool for direction, self-assessment, and informed decision making. They can determine where they stand in relation to current best practices and understand the roadmap of the industry. Users should involve specific teams from across their organization to ensure they have a clear idea where each content area (e.g., Power, Cooling, Compute, Storage, Network) currently resides, and then set appropriate goals for each team to progress through the model. Users should map their data centers and IT portfolios using the Data Center Maturity Model Equalizer. Figure 4 below outlines the Equalizer for four content areas; users should map every part of the maturity model in a similar manner). The Data Center Maturity Model Equalizer outlines where the organization is today (green), identifies the organizational target (yellow), and shows the maximum levels of the Data Center Maturity Model. The organizational target will vary depending on business and site constraints.

Figure 4. The Data Center Maturity Model Equalizer showing four content areas—users should map every part of the maturity model The model may also be helpful to decision makers within any government organization that seeks to encourage industry-wide data center improvements and/or standards for efficiency/sustainability. It should be noted that The Green Grid made a concerted effort to avoid geographical biases toward any one country or region, intending the DCMM to serve as a useful model for organizations around the globe. The DCMM is also a tool that enables data center managers and owners/operators to benchmark their data centers with the performance, operations, and common best practices of a typical data center.


PAGE 10

Adopting a data center–specific maturity model makes it easier for owners/operators to identify and benchmark their data centers’ current energy efficiency and sustainability maturity state. Because the DCMM provides a comprehensive dashboard for understanding the efficiency and health of the data center and IT portfolio, users are encouraged to share the DCMM alongside the Data Center Maturity Model Equalizer with ―C‖-level executives and senior management—together, the two tools outline a roadmap for the organization. Managers can then share that information with colleagues across the business, furthering efforts to innovate and promote ideas for progressing to higher levels of efficiency. A key aspect of TGG’s particular model is that, unlike some others, the DCMM does not promote certain technologies or even methods for attaining each level’s goals. The different content areas describe the target circumstance but do not provide specifics as to how to reach that goal. Instead, the DCMM provides a framework for individuals to use for innovation and for data center peers to use for collaboration to achieve higher levels of energy efficiency. For instance, an operations manager may ask colleagues from other data centers how they reached Level 3 for cooling, and the answers could be specific economizers, liquid cooling, or another means. In addition, defining a common vision for a Level 5 data center will help move the overall industry forward. The Green Grid would like the model to stimulate innovation across the world to help organizations reach some of the higher levels of efficiency and sustainability across different geographies, regardless of their constraints (e.g., power voltages, local climatic conditions, etc.). For vendors and manufacturers of data center technology, the DCMM provides a roadmap toward which they can plan their future innovation and products. Data centers conducting their efforts based on the DCMM can give direction to and inspire related organizations in the greater data center community as to what new methods, products, services, and so on could/should be developed in the future to further enhance data center energy efficiency. Another example is the Storage section’s goals regarding variable speed components; the DCMM users’ identification of the need for improved energy efficiency through something like variable speed drives should inform the industry’s overall product roadmap for what is available in two to three years.

IV.

Frequently Asked Questions How do I use the model? First, benchmark your current status, and continuously set goals for improvement. It’s critical to gain an understanding of where your organization stands versus current best practices so you can appropriately set those goals. You also should identify innovative ideas to move up the levels. Targets from the DCMM should be assessed alongside other business requirements and constraints. How do I move through the DCMM’s stages? To advance through the model, organizations must implement the initiatives that are outlined in each section’s relevant level. TGG encourages organizations to benchmark themselves with the aim to continually improve and move up the levels, resulting in more efficient and sustainable data centers. 2011 The Green Grid. All rights reserved. No part of this publication may be used, reproduced, photocopied, transmitted, or stored in any retrieval system of any nature without the written permission of the copyright owner.

PAGE 11

How much of one level do I need to accomplish before I can claim I am at that level? You need to complete all of the initiatives as far as they are feasible and your business permits. When claiming a level, however, you should be clear about the elements that you have achieved and those that you have yet to or cannot achieve. What if I am at multiple levels? TGG expects organizations to be at different levels across the model. The model provides a basis for you to benchmark your operations today, monitor progress, and focus future resources. Do these initiatives need to be undertaken at all sites or just some? The Data Center Maturity Model is expected to be used at a site level, although TGG encourages adoption across all your data centers. Organizations should strongly consider creating a dashboard that brings together all of the individual data centers to gain a more comprehensive picture of overall data center maturity. Is the DCMM designed for use only for new builds, or should retrofits and legacy sites use it, too? The model should be utilized for legacy, retrofits, and new data centers. TGG encourages organizations to move their legacy data centers through the maturity levels as far as they can, although some of the initiatives may only be possible in newly built data centers. For instance, significantly increasing the environmental range of the data center is predicated on having servers that run in those parameters and is not something that may be possible in legacy sites. The higher levels of the model should inform an organization’s data center strategy, with organizations aiming as high as possible. Who will review my data center’s level? Is it self-assessment? Yes, an organization assesses its own data centers. Should I use the DCMM to compare with data centers from other organizations? Currently, the DCMM should not be used to compare facilities from different organizations. TGG task force believes the best use of the DCMM is the enterprise comparing its own data centers over time. Is this a rating system or certification? The DCMM has not been created as a rating system or for certification. If the model were to evolve into a rating system, further iterations would be necessary.

V.

Next Steps Rather than a static, guaranteed roadmap, TGG intends the Data Center Maturity Model to be a living document—one that undergoes regular updates and improvement as data centers learn, share, and progress. The expectation is that the DCMM will need to change as the whole industry advances. To that end, The Green Grid plans to engage the industry to garner feedback on the DCMM’s different levels, which it will use to revise and update the model’s benchmarks and achievable goals.


PAGE 12

VI.

Conclusion The Data Center Maturity Model shows the relative steps that data centers can take to enhance their energy efficiency, and it does so while remaining technology agnostic. Other models, while valuable, often focus on limited aspects of the data center and promote particular products and solutions, but The Green Grid lays out the journey to take within numerous areas of the data center for users as well as for the industry as a whole. The DCMM promotes collaboration, innovation, and change, pushing the industry forward without constrained thinking because it emphasizes resulting efficiencies rather than prescriptive steps. The ultimate goal is for data centers around the world to be able to do more with less, minimize their environmental impact, and take advantage of each others’ insights to reach efficiencies faster using the common language of the Data Center Maturity Model.

VII.

References Power 

NEMA http://www.nema.org



ASHRAE http://www.ashrae.org



ASHRAE Technical Committee 9.9 http://tc99.ashraetcs.org



ASHRAE Datacom Series http://www.ashrae.org/publications/page/1900, particularly Thermal

Cooling

Guidelines for Data Processing Environments, Second Edition, ASHRAE, 2009 

Herrlin, M. K., 2005, Rack Cooling Effectiveness in Data Centers and Telecom Central Offices: The Rack Cooling Index (RCI), ASHRAE Transactions, Volume 111, Part 2



The Green Grid, Data Center Infrastructure Efficiency (DCiE) Detailed Analysis (2008) http://www.thegreengrid.org/en/Global/Content/white-papers/DCiE-Detailed-Analysis

Management 

The Green Grid, ERE: A Metric for Measuring the Benefit of Reuse Energy from a Data Center (2010) http://www.thegreengrid.org/en/Global/Content/white-papers/ERE



The Green Grid, Carbon Usage Effectiveness (CUE): A Green Grid Data Center Sustainability Metric (2010) http://www.thegreengrid.org/en/Global/Content/whitepapers/Carbon_Usage_Effectiveness_White_Paper



The Green Grid, Water Usage Effectiveness (WUE™): A Green Grid Data Center Sustainability Metric (2011)

Other Facility 

LEED: http://www.usgbc.org/DisplayPage.aspx?CategoryID=19


PAGE 13



BREEAM: The Environmental Assessment Method for Buildings Around the World – http://www.breeam.org

Other IT

VIII.



CSCI: Climate Savers Computing Initiative http://www.climatesaverscomputing.org



RoHS: http://www.rohs.eu/english/index.html



WEEE: http://ec.europa.eu/environment/waste/weee/index_en.htm

About The Green Grid

The Green Grid is a global consortium of companies, government agencies, and educational institutions dedicated to advancing energy efficiency in data centers and business computing ecosystems. The Green Grid does not endorse vendor-specific products or solutions, and instead seeks to provide industry-wide recommendations on best practices, metrics, and technologies that will improve overall data center energy efficiencies. Membership is open to organizations interested in data center operational efficiency at the Contributor, General, or Associate member level. Additional information is available at www.thegreengrid.org.
