April 15, 2010

Technical scribblings RE Harmonizing Global Metrics for Data Center Energy Efficiency

Posted by John Stanley

A couple of weeks ago, several industry and government groups from the US, Europe, and Japan announced that they had come to some basic agreements on “the guiding principles of data center energy efficiency metrics.” There are some very encouraging technical aspects to what was decided, which I’ll describe in this post.

(Two weeks is an eternity in blog time, so I won’t bother re-hashing the press releases yet again. If you want the basics, please read about them via The Green GridIT World, and Environmental Leader. Even better, read the different documents from early February stating what was agreed to, which can be found via EPA ENERGY STAR, another Green Grid post, or Rich Miller’s recent post.)

Here are a few encouraging technical points:

1) The first encouraging point is that the agreed-upon guidelines  for PUE define “total energy” in the numerator to include all forms of energy, not just electricity. For example, if you run a natural gas-fired absorption chiller in your data center, you might use much less electricity than your neighbor with an electric compressor chiller, but you should still count that gas energy as overhead. You’re paying for it, and it’s sure not going to your IT equipment or directly serving your IT customers.

2) The second encouraging point is similar: For data centers that are part of larger buildings, total energy in the PUE numerator should count all data center cooling, lighting, and other support infrastructure. Some data centers in larger buildings forget that their cooling systems consume lots of chilled water from the central building chiller, driving up chiller electricity use. If you forget this chiller energy, and just count electricity on the circuits directly powering the server room, then your PUE will look better than it really is.

Being diligent about counting all relevant energy is something Uptime Institute has long advocated, so I’m glad to see that #1 and #2 made it into the global guidelines.

3) The third encouraging piece was that total energy in PUE numerator should be calculated using source energy. When we talk about electricity, we often talk about source energy as opposed to site energy. Site energy is the amount of energy consumed at the site where it’s used, in the form of electricity. Source energy is the amount of fuel energy needed back at the power plant to get that electricity to the site. Source energy is much bigger. Of all the fuel energy put into a power plant, only about 33% of it is turned into electricity–the other 2/3 turns into waste heat. Then, ~6-7% of the generated electricity is lost as it makes its way down the power lines to its destination. So, for every 10 kBTU (~3 kWh) of electric energy used by a data center, you need to burn about 10 / (1-.07) / .33 = 32 kBTU worth of fuel.

Source energy is the correct way to compare electric and non-electric energy consumption. Image that you just put 10 kBTU of natural gas energy into your gas-fired chiller, while your neighbor just used 10 kBTU of electricity in his/her electric compressor chiller. The site energy is the same, but your neighbor has just consumed about 3x as much natural gas if you trace the electricity back to the natural gas-fired power plant.

Source energy is what we care about from an  environmental perspective, since it’s BTUs of coal and gas that cause pollution. Electricity is just the middle-man. Now we’re getting somewhere.

-

So, good progress by the industry, and I’m glad we’re getting buy-in for universally accepted practices across the globe.  I only want to have to write blog posts about one set of standards.

April 6, 2010

The SEC ruling and your data center, Part 2

Posted by John Stanley

This is Part 2 of a multi-part post on the SEC’s recent ruling on climate change. The first post provided a brief summary of the ruling itself. This post drills down into the possible effects on data centers.

Q: So climate change is material. Why should my data center care?

A: Fear of rising electricity costs will likely change how data center decisions are made

With the SEC’s ruling, companies will have to report possible risks (or benefits) that may result from climate legislation, climate-related market or technology changes, or physical climate impacts. The question that data center managers should ask themselves is “Will my company’s cognizance of these risks cause upper management to make different decisions about the way our data centers are run?”

There are three points I will make in response to this question, and I’ll expand each one below.

  • Yes, reporting of risks will make upper managers more observant of data center energy use.
  • Ironically, the risk may not be “material,” except in a few cases.
  • Even if the risks of higher electricity cost aren’t material to the company overall, they may be material to the CIO.

Yes, reporting of risks will make upper managers more observant of data center energy use

Having to report climate-related risks to the SEC will almost certainly make upper managers think about what would happen to energy prices under climate legislation, and that will make them think about energy consumption. And once they’re thinking about consumption, data centers will stick out like a sore thumb.

According to Uptime Institute data, at one major financial institution, just 3 data centers out of 3,400 street addresses accounted for 12% of total corporate energy consumption. For three other financial institutions, data centers consumed 15%, 25%, and 33% of total company energy. (This is from the slide deck “Three Imperatives for Making Data Center Efficiency a C-Suite Priority”, presented by Ken Brill at Uptime Institute’s Symposium 2009. The deck is available on Uptime’s website.)

The image of data centers as energy hogs will likely make upper management much more interested in efficiency. The key is for data center managers to eliminate energy use that is genuinely “waste,” while vigorously defending energy use that only looks like waste to higher-ups not intimately familiar with data center operations. Even though a UPS throws away 10% of its energy in conversion losses, you probably can’t meet your SLA without it. That example is absurdly obvious, but there will be subtler situations. We as an industry will come under more pressure to cut energy use, and we’ll be under pressure to justify the energy use we can’t cut by tying it directly to an SLA or other business requirement.

Ironically, the risk may not be “material,” except in a few cases

Interestingly, even though the fear of higher electricity prices will drive increased scrutiny of data centers, I would argue that the overall risk to most companies from climate legislation raising data center-related electricity costs is actually fairly small, maybe too small to be “material” by SEC standards.

For most organizations, the IT budget is less than 10% of revenues. (Thanks to an old article in CIO magazine and their 2009 State of the CIO survey for this.)

Of the IT budget, around 25% goes to data centers in a typical company. (This is according to the Revolutionizing Data Center Efficiency report by McKinsey & Co.  and Uptime Institute. The slide deck is available on Uptime Institute’s website, and the full McKinsey paper is available on McKinsey’s website. These reports are well worth reading, to give you the big picture of how a “typical” IT budget breaks down into line items.)

Next, electricity costs account for around 10% of the annualized TCO of a data center and the IT equipment inside it. Around 45% is IT capital costs, and around 31% is facility infrastructure CapEx.  (These results come from Jonathan Koomey’s paper “A Simple Model for Determining True Total Cost of Ownership for Data Centers”, which was also written with Uptime Institute.  The paper is available via Uptime’s website.)

Overall, then, data center-related electricity bills likely account for only about 2.5% of the overall IT budget in a typical organization, and likely less than 0.25% of company revenues.

Even if climate legislation causes electricity bills to double, it doesn’t make much of a dent in a company’s overall cost/revenue structure. So, this might not present a material risk to the company by SEC standards.

There are a couple of exceptions, though, in which case the risk could be material:

  • If your company is in an industry where margins are razor-thin, then 0.25% of revenues might actually represent a significant chunk of profits.
  • If your company’s core business is focused on data centers (web companies, co-location providers, etc.), then obviously the percentages above don’t apply. Your data center electricity bills are likely a significant business input cost.

Even if the risk of higher electricity costs aren’t material to the company overall, they may be material to the CIO

This last point has nothing to do with the SEC ruling per se, but it is yet another reason why data center operators will see increased attention to efficiency as climate legislation looms.

Even in a “typical” company where data center electricity use is only 2.5% of the IT budget, keep in mind that the average annual increase in IT budgets is only 6% per year. (Again, this is from the McKinsey reports.) This means that next year’s IT budget could absorb a doubling of electricity prices in its standard annual budget increase, but that’s a huge part of the budget increase chewed up by new electricity costs. For every $100 of IT budget, a doubling of electricity prices adds $2.5 in new electricity costs, on top of the $2.5 that was already there. The overall budget increase is only $6, meaning that only 6 – 2.5 = $3.5 in new budget money is actually left for application roll-out that year. (In fairness, such a price doubling would likely be spread out over multiple years, but each year’s installment would still be a significant chunk of each year’s annual budget increase.)

So, even if the risks of electricity cost hikes driven by climate change legislation aren’t material to your company by SEC standards, they may well be material to the CIO.

April 1, 2010

The SEC ruling and your data center, Part 1

Posted by John Stanley

This is Part 1 of a multi-part post on the SEC’s recent ruling on climate change. This post provides a brief summary of the ruling itself. One or more subsequent posts will drill down into the possible effects on data centers.

SEC: Climate change is material to your business

In late January, the Securities and Exchange Commission (SEC) ruled that public companies must disclose to investors the risks they face related to climate change.  The SEC ruling does not create new rules or modify existing ones–it merely provides interpretive guidance regarding the risk disclosure rules already on the books.

The ruling describes four particular areas where climate change risks (and opportunities) may trigger disclosure requirements:

  • Impact of legislation and regulation – Risks may include costs to purchase allowances under a “cap and trade” scheme, cost to retrofit facilities in compliance with new standards, or reduced demand for carbon-intensive goods/services sold by the company. Interestingly, the SEC states explicitly that “a registrant should not limit its evaluation of disclosure of a proposed law only to negative consequences” (p.23). For example, there may be opportunities to profit from selling allowances, or from an increase in the demand for a company’s (low carbon) products.
  • International accords – Similar to the impact of domestic legislation, international accords may affect business.
  • Indirect consequences of regulation or business trends – Legal, political, scientific, and technological developments associated with climate change can also be material to a company. Again, the SEC asks companies to discuss opportunities as well as risks. New developments may affect price and demand for certain goods and services sold by a company, partners, suppliers, or competitors.  Reputational impacts should also be considered, for companies whose business operations are sensitive to public opinion. (For example, companies to whom branding and image are extremely significant to sales.)
  • Physical impacts of climate change – Finally, companies must disclose risks associated with the physical impacts of climate change, such as severe weather, water availability, and the effects of sea-level rise on coastal operations. Higher insurance premiums or an increase in insurance claims are another potential risk.

The text of SEC Interpretive release 33-9106 is available on the SEC’s website, as is their press release regarding the ruling. The Climate Change and Clean Technology Blog also has a post about the ruling with more details.

Overall, the ruling provides businesses–including those in the IT and data center industries–yet another reason to keep energy use and environmental performance on their radar at the highest levels of management.

March 29, 2010

Sustainable Silicon Valley releases handy reference paper describing CSR frameworks

Posted by John Stanley

Companies getting started with Corporate Social Responsibility (CSR) or environmental tracking seem to face a dizzying array of options for how to count, quantify, and report their progress. Which standard do you use for counting carbon? Which third-party should you report it to? What are the standards for the non-environmental aspects of CSR, like labor rights or corporate transparency?

Last week, Sustainable Silicon Valley (SSV) posted a white paper that provides a handy reference guide for answering these questions.  (White Paper on Sustainability Frameworks (Programs, Guidelines, Protocols, Registries). It should be useful  for any company that’s getting started with CSR, or that wishes to adapt its current initiatives to an accepted standard.

The paper covers four major areas:

  1. Multiple-attribute programs – Programs, protocols, guidelines, or standards for reporting on multiple aspects of CSR. For example: energy, waste, water, and carbon, or “people, planet, profit”
  2. Single-attribute programs – Programs, protocols, guidelines, or standards for reporting on a single aspect of CSR, such as energy use or GHG emissions
  3. Registries, indexes, eco-labels, and seals – Third parties to which companies can report CSR information. These third parties typically verify and certify the reported information, and may provide a label or seal of approval.
  4. Legislation, mandates, and other government initiatives

Of course, many of these areas overlap. For example, the Global Reporting Initiative (GRI) is both a framework for reporting and a registry to which companies can submit their results. The reference table in Appendix 1 of the paper contains a summary list of all the frameworks, registries, and legislation described, showing what each one includes.

The paper also notes that the GHG Protocol–developed by the World Resources Institute and the World Business Council for Sustainable Development–is the clear winner where greenhouse gas reporting standards are concerned:

GHG Protocol provides the accounting framework for nearly every GHG standard, program and registry in the world – ISO (ISO 14064), Carbon Disclosure Project, The Climate Registry, California Climate Action Registry, California Air Resource Board, Global Greenhouse Gas Registry, EPA’s Climate Leaders etc., as well as hundreds of GHG inventories prepared by individual companies. –p.21

The GHG Protocol is also used in the Global Reporting Intiative (GRI), another popular framework for multiple aspects of CSR.

The paper does not attempt to cover organizations primarily concerned with carbon trading, credits, or emissions reduction projects, such as:

  • Verifiers / Auditors / Certifiers
  • Software / Service providers for sustainability tracking
  • Carbon Trading / Offsets

The paper doesn’t go much beyond what diligent individuals can find on their own via some online searching–the paper says that most of the program descriptions are taken verbatim from the relevant organizations’ websites. However, Sustainable Silicon Valley has provided a valuable service by doing that work for us, and making the results freely available.

March 22, 2010

PricewaterhouseCoopers’s template for GHG reporting

Posted by John Stanley

For those of you who wonder what a company’s annual public Greenhouse Gas Emissions Report might look like in the future, PricewaterhouseCooper’s has released a publicly available sample statement. The report, created for the fictitious global company “Typico plc,” gives an example of how a company might arrange such a document,  connect financial and non-financial data, and inform stakeholders about the risks and opportunities the company faces when trying to manage its carbon footprint.

If and when carbon regulations take effect in the US, this template report could be helpful for those in the IT and data center industry who wonder “Where do we begin?”

As I skimmed the report, a few things jumped out at me:

  • PWC’s recommendation is that the company’s standard Annual Report allude to the GHG report, but not include the entire text. Conversely, the GHG report may mention key financial data and direct readers to the Annual Report for details.
  • Typico’s has chosen to include climate change, environment, and other focus areas in its calculation of executive bonuses. This is a good way to make sure busy executives keep the environment on their list of critical priorities.
  • The GHG statement contains an Independent Assurance Report conducted by a third party, similar to the way financial statements are done. As carbon becomes more material to a company’s operations, independent audits become a requirement.

The PWC report itself is a pretty quick read, so I won’t delve into further detail on this post. The document is not intended to be an encyclopedia of everything a manager would need to know to create a report like this, I definitely think it will serve as a handy “getting started” manual for many firms, including many in the IT and data center industry.

March 15, 2010

UK CRC – Policy challenges in dealing with the critiques

Posted by John Stanley

As I mentioned in a previous post, the concerns that the UK’s IT industry has raised over the CRC present several public policy challenges. In this post, I’ll make a few points about each of the critiques.

(1) “Under CRC rules, you can’t claim credit for renewable energy for which you receive subsidies”

An article last year in eWeek Europe, provocatively entitled “Companies Threaten to Kill Green Energy Over Government Rules,” points out that companies like Sun and BT are rethinking planned renewable energy projects because the government rules undermine the business case. The rules force data centers with on-site renewable generation to choose between counting the energy as renewable under their CRC tally, or receiving payments from selling Renewable Obligation Certificates (ROCs) to utilities. Many projects need the ROC payment subsidies to be economical, and it’s confusing to project owners why their green energy wouldn’t then count as green when they report to CRC.

However, allowing data centers to count a single renewable energy project under CRC and simultaneously sell ROCs for it would be double-counting the renewables benefit, and that’s not reasonable.

The UK government has created two separate requirements, so that different groups can each do their part. (A) Utilities are required to do their part by procuring more renewables, and (B) data centers are required to do their part by cutting their fossil energy use. You can’t count a single renewables project toward both of these requirements; it messes up the accounting and reduces the overall number of projects that get done. Allowing this would be like letting your company’s sales team take a single deal closed over the weekend and count it towards sales quotas for both this week and next week. This throws your revenue numbers off, and it allows a slack sales team to get away with one deal instead of two.

So, I don’t consider it a flaw that the CRC considers a renewable energy project that sells ROCs to be non-green. That project used to be green, but the owner took the “greeness,” bundled it into a ROC, and sold it to a utility. The utility now owns the “greeness,” not the project operator.

However, I do support the data center industry’s sentiment that the government should aggressively encourage renewable energy. But let the government do it with additional incentives, not obfuscated accounting tricks.

(2) “The CRC encourages companies to outsource / offshore their carbon, rather than actually reducing it”

Another criticism of the CRC is that it might just cause companies to outsource energy-intensive data center operations, rather than improving efficiency.

This criticism presents much more of a public policy conundrum. It would indeed be tragic if UK companies simply moved their data centers to somewhere else in Europe, or even to China, where the electricity could be even more carbon-intensive. This wouldn’t reduce emissions, and it would export the jobs and economic benefits associated with the data centers out of the UK.

One possible correction might be something like an import tariff on data center services from countries that don’t regulate their carbon emissions. That way, companies would have no cost incentive to cheat the system by shipping their energy-intensive data centers to places with lax rules. However, doing this would be a major challenge. First, a data center service is not like a shipping container full of bananas; it’s much more difficult to quantify data center services, stop them at the border, and assess the appropriate toll. Second, even if you could do that, free trade laws in the EU prohibit many kinds of import tariffs.

On the other hand, the prospect of waiting until all countries regulate their carbon emissions, so that there’s no incentive for offshoring, is equally discouraging. If countries can’t move together, and everyone is afraid to move first, how is anything ever going to get done?

Hence the conundrum.

(3) “The CRC penalizes growth by focusing too much on absolute emissions reductions rather than relative intensity metrics”

This is another big challenge. What if a data center company doubled its actual CO2 emissions, but quadrupled its revenues and services provided? The company has clearly gotten a lot more efficient in terms of reducing CO2 per dollar of economic benefit, and it should be rewarded for its success.

On the other hand, if CO2 emissions go up by 10x, and Gross World Product goes up by 100x, we’d still face potentially catastrophic climate change. The atmosphere doesn’t give us credit for our factor-10 improvement in CO2 per dollar. (To put it another way: We still cook. We’re just rich enough that we can cook in a golden pot.)

If we want to tackle global warming, we need to find a good way to reward efficiency while meeting calls for a global reduction in absolute emissions.

No easy answers

With the possible exception of #1, these three criticisms of the CRC are very reasonable, but it’s also very hard to simply give the critics what they want. I hope lawmakers and we in the industry can work together to find a more creative solution.

March 9, 2010

UK Climate Legislation Alphabet Soup

Posted by John Stanley

After my last post about the CRC developments in the UK, I realized that it might be helpful to provide a quick guide to the alphabet soup of acronyms that describe the UK’s climate change legislation. At the very least, making this guide was helpful for me.

The most important feature of this guide is that I’ve tried to include information links to the official government sites wherever possible, rather than simply reporting information I’ve read in secondary articles. This way, readers who want to know more can quickly find and browse the official UK government info.

.

Climate Change Act of 2008

(The Act, online)

This Act is the centerpiece of the UK’s climate legislation. It sets a carbon reduction target of 80% below 1990 levels by 2050, and 34% below 1990 levels by 2020. The Act also creates the Committee on Climate Change (CCC) and establishes five-year carbon budgets.

(Note that the original text of the Act said the 2020 target was 26% below 1990 levels, but this was amended to 34% in 2009. )

More source links: Sections of the Act on 2050 target, 2020 target, carbon budgets, and CCC. Amendment to 34% here. DECC’s summary of the Act here.

.

Committee on Climate Change (CCC)

(CCC home page)

The Committee on Climate Change (CCC) is an independent body established under the Climate Change Act to advise the UK Government on setting carbon budgets, and to report to Parliament on the progress made in reducing greenhouse gas emissions. –quoted from the CCC website

.

Department of Energy and Climate Change (DECC)

(DECC’s “About us” page)

DECC was created in late 2008 to bring together the UK’s energy and climate policies, which had previously been handled by separate agencies. The department is responsible for all aspects of UK energy policy. DECC divides its work into four key areas: (a) global climate change and energy, (b) UK energy supply, (c) supporting consumers, and (d) working towards a low-carbon UK.

More source links: DECC What we do page

.

CRC Energy Efficiency Scheme

(formerly know as the Carbon Reduction Commitment, or CRC)

(DECC’s page on the CRC)

The CRC Energy Efficiency Scheme is the UK’s mandatory energy savings and climate change scheme, which will start in April 2010. CRC will operate as a cap-and-trade system and is designed to address CO2 emissions not already covered under Climate Change Agreements and the EU Emissions Trading System. The CRC covers large public and private sector organizations that use more than 6,000 MWh of electricity annually. The government estimates that entities covered under the CRC are responsible for about 10% of the UK’s carbon total emissions.

.

Renewables Obligation (RO)

Renewables Obligation Certificate (ROC)

(OFGEM’s page about the RO)

The RO is the main support scheme for renewable electricity projects in the UK, which came into effect in 2002. It requires electricity suppliers to source an increasing fraction of their electricity from renewable sources. A supplier’s compliance is tallied via Renewables Obligation Certificates (ROCs)–one ROC is issued for each MWh of renewable electricity generated. Each year, a company must present enough ROCs to demonstrate that it has met its obligation, or pay a buy-out price for each ROC by which it falls short.

The RO program is administrated by the Gas and Electricity Markets Authority, whose day-to-day functions are handled by the Office of the Gas and Electricity Markets (OFGEM).

More source links: OFGEM About us page

.

Climate Change Levy (CCL)

Climate Change Agreements (CCA)

(DECC’s page on CCL)

The CCL is a tax on energy use in industry, commerce, and the public sector, designed to encourage energy efficiency and reduced GHG emissions. For electricity, the Levy is 0.0047 £/kWh. The CCL took effect in 2001.

Climate Change Agreements (CCA’s) allow certian energy-intesive businesses to receive a discount on the CCL of up to 80%, in return for meeting energy efficiency or carbon reduction targets.

Other source links: DECC’s page on CCA, CCL introduction page at HM Revenue & Customs, 2009 CCL rates, Carbon Trust’s CCL page

.

March 6, 2010

Worries about flaws in the UK’s Carbon Reduction Commitment (CRC)

Posted by John Stanley

The UK’s data center industry has shown great concern regarding its pending regulation under the government’s Carbon Reduction Commitment (CRC) legislation. Following a March 1 meeting by two major industry groups, the IT industry may instead ask the government to regulate data centers’ carbon footprint under an alternate set of rules called Climate Change Agreements, which apply to certain energy intensive industries. (For readers interested in more details, this is discussed at length in a recent report by The 451 Group.)

While, I am generally wary any time any industry begs the government for special exemptions from environmental rules, the critiques of the CRC raised by UK data center operators make genuine sense. There are significant problems with the CRC’s design that create perverse incentives and could fail to drive genuine emissions reductions.

UK IT industry group Intellect released a document airing some of its concerns last summer (available here), and industry groups like The Green Grid and the British Computer Society (BCS) have also raised critiques. (Several other good posts echo CRC critiques as well–see posts from BusinessGreen, Computer Business Review, and a subsequent post by Intellect.)

There are at least three problems with the rules as they stand:

First, the CRC scheme encourages companies simply to “dump” their carbon footprint elsewhere by outsourcing data centers and other energy intensive operations to places with lax rules. This doesn’t do much for global emissions, and may even worsen emissions if operations move from places with cleaner power to coal-heavy areas like China and India.

Second, the CRC penalizes business growth, even if that growth comes with vast efficiency improvements. So, if you implement efficiency improvements to cut your energy per customer by 1/3, but your customer base doubles, then you still get dinged because your total emissions have gone up.

In addition, the UK’s Climate Change Act rules also disallow data center operators from getting credit for on-site renewable energy, unless they forgo the government subsidies that renewables can often get. (See article in eWeek Europe.)

All of these are legitimate concerns, from an IT industry that appears to have genuine interest in cutting its environmental footprint. However, these concerns also present a public policy conundrum, for reasons I will discuss in a subsequent post.

In the mean time, we’ll see if the IT industry is successful in lobbying the government to be regulated under via the scheme of Climate Change Agreements, rather than the CRC rules.

March 2, 2010

Greening of IT – Closing the Communication Gap Between CIOs and CFO

Guest post by Deborah Grove

On February 25, a crowd of 100 real estate executives gathered in Dallas, TX, to hear Ken Brill speak about the true cost of data center operations. The sponsors of the event, Jones Lang LaSalle, one of the world’s largest property management companies (750 locations in 60 countries) with significant square footage of mission critical facilities under management, wanted their clients to hear about the issues facing CIOs today.

The event was named “Greening of IT – closing the communication gap between CIOs and CFOs.” Along with Ken Brill, Lauralee Martin, CFO/COO of Jones Lang LaSalle Incorporated, spoke to the audience about her drive to focus the company on sustainability as a core value for JLL clients because as a discipline, it saves money as well as green house gases (GHG). Lauralee Martin emphasized the need for CIOs and CFOs to ask tough questions from their data center managers, and advised data center managers to be ready for those questions over the coming year as the C-Suite comes to realize how much of their investment in data centers is largely inefficient, as asset management best practices go.

Ken reminded the audience that servers can produce more than 4 tons of GHG per annum and while a single server might cost $400 – $500 per year in utility bills, the CapEx investment is likely to be $8000 for every $2000 server because the power and cooling infrastructure demands are stringent to deliver uptime levels.

After Brill and Martin spoke, a panel including Chris Crosby of Digital Realty Trust, Chris Page of Yahoo, Bob Morris of Corgan Associates, and Michael Richard of Oracle offered insights on data center industry trends, best practices, and C-level engagement on urgent topics of financing data centers.

Deborah Grove of Grove Associates provided this guest blog as she observed the event. Deborah originally discovered JLL’s industry leadership on sustainable commercial and industrial buildings through the report The Role of Finance in Environmental Sustainability Efforts and connected JLL executives with the Uptime Institute because she saw that collaborating on best practices would benefit JLL as well as Uptime clients.

March 1, 2010

Should you replace old servers with efficient ones? Embodied energy vs use-phase energy

Posted by John Stanley

Andy’s post above brings up an interesting point that’s being debated in the data center industry right now: Should you keep the servers you have as long as you can, so you can avoid the energy and CO2 impact of manufacturing new servers, or should you get rid of your old servers so you can buy more energy efficient ones instead?

In the field of lifecycle assessments, we often talk about three phases of a product’s life:

(1) Production – usually defined to include materials extraction, manufacturing, and distribution

(2) Use

(3) Disposal – which can include throwing something in a landfill, or preferably recycling it

So, the debate comes down to this: Is the use-phase energy/carbon of a server more important than the production and disposal energy/carbon?

(Hint: yes)

As Andy and others point out, the UK government’s Cabinet Office has a green ICT strategy (available here), one provision of which is using equipment in data centers as long as possible, since it it takes energy to make new equipment.

But, let’s look at data for an office PC. According to an IVF 2007 report, about 80% of the energy and carbon impact is from the use phase, and only 20% is from the manufacturing and disposal. (The report in question is succinctly entitled “European Commission DG TREN Preparatory studies for Eco-design Requirements of EUPs. Lot 3: Personal Computers (desktops and laptops) and Computer Monitors. Final Report (Task 1-8)”.)

Or, look at Carnegie Mellon’s EIO-LCA database, which covers all sectors of the economy, and look at the sector “Electronic computer manufacturing.” It takes about 11 GJ of primary enegy and 1 metric ton of CO2 to make and dispose of a $2,500 server, but by my calculation that server will use about 120 GJ of primary energy to make the necessary electricity, which will emit 7 metric tons of carbon at US CO2/MWh rates. So, the use phase is over 90% of the energy and over 85% of the carbon.

These numbers aren’t for servers specifically, but they’re probably reasonable, and they show the use phase is much more important.

BUT, you should only get rid of your old servers if the new server’s actual efficiency improvement is enough to make up for having to build a new one.

Here’s a scenario:

(a) A server you already have took 20 units of energy to make (which you’ve already “paid”), and it will take 80 units of energy to run for another 5 years.

(b) A new, more efficient server server that does the same amount of IT work would require 20 new units of energy to make, and it would take 59 units of energy to run over the next 5 years.

So, buying a new server which is 1-59/80 = 26% more efficient would require a total of 79 units of energy, while continuing to run your old server would be just slightly worse, at 80 units of energy. But if the new server was only 25% more efficient, you’d need 64+20 = 84 units of energy to manufacture and run it, which is worse than just keeping the old server.

So, my recommendation is that the UK government and others not create a blanket policy of keeping all old servers as long as possible. Instead they should do a quick “back of the envelope” calculation to determine if the efficiency benefits of a new model are big enough to beat the energy/carbon penalty of building something new. What they’ll probably find is that it’s well worth replacing very old, very inefficient equipment, but that it’s not worth replacing relatively recent equipment whose efficiency is just a step or two behind today’s models.