How Much Power Do Data Centers Use

Energy Innovation partners with the independent nonprofit Aspen Global Alter Found (AGCI) to provide climate and energy research updates. The research synopsis below comes from Eric Masanet and Nuoa Lei of the McCormick School of Engineering and Applied Scientific discipline at Northwestern University, and a full list of AGCI's quarterly inquiry updates covering recent climate change research on clean energy pathways is available online at https://world wide web.agci.org/solutions/quarterly-research-reviews

Data centers tin be thought of as the "brains" of the internet. Their role is to process, store, and communicate the data behind the myriad data services we rely upon every mean solar day, whether it be streaming video, e-mail, social media, online collaboration, or scientific calculating.

Data centers use dissimilar information engineering science (It) devices to provide these services, all of which are powered by electricity. Servers provide computations and logic in response to information requests, while storage drives firm the files and data needed to meet those requests. Network devices connect the data eye to the internet, enabling incoming and outgoing data flows. The electricity used by these It devices is ultimately converted into rut, which must be removed from the data heart by cooling equipment that likewise runs on electricity.

On average, servers and cooling systems account for the greatest shares of direct electricity use in information centers, followed by storage drives and network devices (Figure one). Some of the world'south largest data centers can each comprise many tens of thousands of Information technology devices and crave more than 100 megawatts (MW) of power chapters—enough to power around eighty,000 U.S. households (U.Due south. DOE 2020).

As the number of global net users has grown, so too has need for data center services, giving rising to concerns about growing information center energy utilisation. Betwixt 2010 and 2018, global IP traffic—the quantity of data traversing the net—increased more than 10-fold, while global information center storage capacity increased by a factor of 25 in parallel (Masanet et al. 2020). Over the same fourth dimension menstruation, the number of compute instances running on the world's servers—a measure of total applications hosted—increased more than than half dozen-fold (see Effigy 3) (Masanet et al. 2020).

These strong growth trends are expected to continue as the world consumes more and more data. And new forms of information services such as bogus intelligence (AI), which are peculiarly computationally-intensive, may accelerate demand growth farther. Therefore, the ability to quantify and project data centre energy utilisation is a central energy and climate policy priority.

Data middle energy use estimates: A tale of two methods

Official statistics are non currently compiled on information center energy employ at national or global levels. Therefore, mathematical models must be used to estimate this energy use. So-chosen "bottom-upwardly" models account for the installed stocks of IT devices in different data centers and their energy use characteristics to get in at an estimate of total energy utilisation. While lesser-up studies offer many insights into the drivers of energy use, they are also very data and time intensive; therefore, they don't appear often. For example, the virtually authoritative lesser-upwards written report in the last decade appeared in 2011 (Koomey 2011), and it estimated that data centers deemed for between 1.1 percent and 1.five percent of global electricity use in 2010.

In contrast, extrapolation-based models approximate total energy use by taking previous lesser-up values and scaling them upwardly on the basis of data center market growth indicators, such as global IP traffic (Andrae and Edler 2015) or information middle investments (Belkhir and Elmeligi 2018). Because extrapolation-based approaches are much simpler, they have been used to fill the temporal gaps left backside by desultory bottom-up studies.

However, such extrapolations tend to gauge large increases in data center energy use, given that the market indicators upon which they rely also grow speedily. For example, some oft-cited extrapolations accept suggested that global information center energy may accept doubled since 2010, and by extending this historical logic, that it will go along ascent swiftly in the future (Andrae and Edler 2015, Andrae 2017, Belkhir and Elmeligi 2018, Bawdy 2016). These estimates have received significant attention (Jones 2018), reinforcing the common belief that speedily growing demand for data equals rapidly growing data center energy utilisation.

Nevertheless, new results from the bottom-upwards perspective bespeak otherwise: Despite rapid growth in demand for information services over the past decade, global data center free energy use likely rose by only 6 per centum between 2010 and 2018 (Masanet et al. 2020). These new results were based on integration of numerous recent datasets, which better characterize the installed stocks, operating characteristics, and energy utilisation of information eye IT devices, also every bit structural shifts in the data centre manufacture, compared to past studies.

The finding that global information centers likely consumed around 205 terawatt-hours (TWh) in 2018, or 1 percent of global electricity use, lies in stark contrast to earlier extrapolation-based estimates that showed rapidly-rising information center energy utilisation over the past decade (Figure two).

3 chief efficiency effects explain this near-plateau in free energy use: Starting time, the free energy efficiency of IT devices—and servers and storage drives in item—has improved substantially due to steady technological progress by Information technology manufacturers. Second, greater use of server virtualization software, which enables multiple applications to run on a single server, has significantly reduced the energy intensity of each hosted awarding. 3rd, near compute instances have migrated to large cloud- and hyperscale-form data centers, which utilise ultra-efficient cooling systems (among other of import efficiency practices) to minimize energy utilisation (Effigy ii).

These efficiency furnishings are not well captured in extrapolation-based approaches, given their lack of technological and structural particular. In other words, while extrapolation-based approaches generally capture the drivers of data middle demand that would button free energy use up (top one-half of Figure iii), they fail to adequately capture strong countervailing efficiency trends (bottom one-half of Figure 3) that continue energy use in check.

What almost CO2 emissions?

The substantial electricity use of data centers also gives rising to concerns over their carbon dioxide (CO₂) emissions. Unfortunately, information technology is not notwithstanding possible to accurately estimate total CO₂ emissions, due to a lack of data on the locations of the vast bulk of global data centers and the emissions intensities (measured in grams CO₂ per kilowatt-hour) of their actual electricity sources. Simply a handful of companies, including Google, Apple, Switch, and Facebook publicly report such data, indicating a growing trend among some of the world's largest data middle operators toward renewable energy procurement.

Knowing the electricity use of global information centers, even so, provides a useful criterion for testing claims about the CO_ii implications of data center services. For instance, i oft-repeated claim is that the earth's data centers emit equally much CO₂ as the global aviation industry (Pearce 2018), which is roughly 900 billion kilograms of CO_ii (Air Transport Activeness Group 2020). Because that global data centers recently consumed around 205 billion kWh, for this claim to be true, their average electricity emissions intensity would accept to be around 4.iv kg CO₂/kWh.

Yet the average coal-fired ability establish, the most carbon-intensive option bachelor, has an emissions intensity that is less than one-fourth of this value—around one kilogram CO₂/kWh (U.S. Eia 2020). And it's articulate that all the world'south information centers are not running on coal, specially in light of renewable energy utilisation by some large data centers that account for increasing shares of global compute instances.

Another contempo claim is that "the emissions generated by watching 30 minutes of Netflix (i.6 kg of CO₂) is the same every bit driving most four miles." This claim is backed up by assumptions that information centers providing Netflix streaming services would consume around 370 TWh per twelvemonth (Kamiya 2020). Yet this value is 1.8 times larger than the 205 TWh estimated for all of the globe's data centers combined, which provide gild with myriad other information services beyond simply streaming Netflix videos. (For a more consummate assessment, meet Kamiya 2020.)

Therefore, the improved clarity that these contempo lesser-upwards estimates accept brought on global data heart use tin can besides enable "reality checks" that betrayal the implausibility of some attention-grabbing and widely-circulated claims virtually data centers' contribution to climatic change.

The path forwards

That said, in the coming decade, a significant hazard exists that rapidly growing demand for information services—and compute-intensive applications like AI in particular—volition brainstorm to outpace the efficiency gains that have historically kept data centre energy utilisation in cheque. Potential all the same remains for substantial efficiency gains but investments in adjacent-generation computing, storage, and heat removal technologies volition be required to avert potentially steep energy utilisation growth later this decade. And parallel investments in renewable power sourcing will be required to minimize the climate implications of unavoidable data center energy use (Masanet et al. 2020).

Better modeling capabilities are required for decision makers to confidently evaluate time to come efficiency and mitigation options, and so developing more robust and predictive methods that increase the frequency of bottom-up insights and overcome the limitations of extrapolation-based forecasts are a key priority for the energy analysis community.

These models will exist needed by policy makers and energy planners for monitoring future information center energy use trends, understanding key energy use drivers, and assessing the effectiveness of various policy interventions for managing possible energy growth. Because information centers are present nearly everywhere, such capabilities volition be required at both national and global levels, and peculiarly for China, where data centre capacity is expanding speedily.

Analysts should consider several key priorities. Outset, the evolution and open sharing of reliable information sources on installed stocks, configurations, and the energy use characteristics of Information technology devices and cooling/power systems will enable more than common and accurate technology representation beyond models. Such technological detail will also ensure proper consideration of important efficiency trends. 2nd, models should be shared and inter-model comparisons should be conducted so that analysts can develop best practices, increasing the conviction in model outputs. Third, analysts must work together to develop methods for modeling emerging trends, such as AI, the rollout of 5G, and increased border computing, giving policy makers early insights into their possible energy use implications. Fourth, more reliable information should be developed and shared openly for the Asia Pacific region, and for China in detail, where data eye demand is growing rapidly. Lastly, better methods are needed for prospective analysis of side by side-generation calculating, storage, and estrus-removal technologies for accelerating investments in technologies that might avert futurity free energy employ growth.

References

Air Transport Action Group (2020). Facts and Figures. https://www.atag.org/facts-figures.html

Andrae, A.Due south. and Edler, T. (2015). On global electricity usage of communication engineering: trends to 2030. Challenges, 6(1), pp.117-157.

Andrae, Anders. (2017). Total Consumer Ability Consumption Forecast. https://www.researchgate.internet/publication/320225452_Total_Consumer_Power_Consumption_Forecast

Bawdy, T. (2016). Global warming: Information centres to consume 3 times every bit much energy in next decade, experts warn. The Independent. Sabbatum 23 January 2016.

Belkhir, L. and Elmeligi, A. (2018). Assessing ICT global emissions footprint: Trends to 2040 & recommendations. Journal of Cleaner Production, 177, pp.448-463.

Jones N. (2018). How to stop data centres from gobbling up the world's electricity. Nature. 2018 Sep;561(7722):163-166. doi: 10.1038/d41586-018-06610-y.

Kamiya, George (2020). Factcheck: What is the carbon footprint of streaming video on Netflix? Carbon Brief, February 25^th. https://www.carbonbrief.org/factcheck-what-is-the-carbon-footprint-of-streaming-video-on-netflix?

Koomey, Jonathan (2011). "Growth in data center electricity utilize 2005 to 2010." A report by Belittling Press, completed at the request of The New York Times 9 (2011): 161.

Masanet, Eric, Arman Shehabi, Nuoa Lei, Sarah Smith, and Jonathan Koomey. "Recalibrating global data center free energy-employ estimates." Science 367, no. 6481 (2020): 984-986.

Pearce, F. (2018). "Energy Hogs: Tin World'south Huge Data Centers Be Made More than Efficient?" Yale Environment 360, Apr 3^rd, https://e360.yale.edu/features/energy-hogs-tin can-huge-data-centers-exist-made-more-efficient

Shehabi, Arman, et al. (2016). United states data center energy usage study. No. LBNL-1005775. Lawrence Berkeley National Lab.(LBNL), Berkeley, CA (U.s.).

U.S. Department of Energy (2020). Annual Free energy Outlook 2020. https://www.eia.gov/outlooks/aeo/

U.Southward. Energy Information Administration (2020). How much carbon dioxide is produced per kilowatthour of U.S. electricity generation? https://world wide web.eia.gov/tools/faqs/faq.php?id=74&t=xi

How Much Power Do Data Centers Use,

Source: https://energyinnovation.org/2020/03/17/how-much-energy-do-data-centers-really-use/

Posted by: wisegion1993.blogspot.com

Share this post