Bringing the Cloud Back Down to Earth
BY BLAIR TALBOT
On Sunday, October 20th I attended the Radical Networks conference and attended two talks: The Carbon Footprint of the Internet with Jasmine Soltani and Everything has a Resonant Frequency: Crystals, Networks, and Crystal Networks with Ingrid Burrington. Both talks covered a lot of ground (or rather, earth) on the sweeping topic of the environmental impact of the Internet and the manufacture of its physical infrastructure by two very broad thinkers whose research has forged ahead in areas where concrete data is hard to come by. For clarity and concision, I will focus on my main takeaways from the first talk, the Carbon Footprint of the Internet.
Soltani began her talk by explaining the bottom-up approach she and other activists have taken to calculating the carbon footprint of the Internet in the absence of definitive, trustworthy sources: identify all the components and processes that make up the Internet, calculate the energy consumption of each, and identify the energy sources of each and convert that energy amount to the CO2 equivalent. To date there still exists no surefire way to calculate the carbon footprint of the Internet, and therefore estimates of the energy intensity of the Internet diverge by a factor of 20,000, which can in part be explained by different definitions of what the Internet is and what it includes (Hilty & Aebischer, 2015).
Current estimates state that the Internet accounts for about 3% of global greenhouse gas emissions (Belkhir & Elmeligi, 2018), which is both significantly less than what I would have predicted and still too abstract to comprehend. For comparison, in Greenpeace’s suspiciously optimistic 2014 report on renewable energy and the Internet, Clicking Clean: How Companies are Creating the Green Internet, the authors state that if “the Cloud” were a country, it would be the 6th largest consumer of electricity on the planet (Cook et al., 2014).
One fact that is generally agreed upon is that the most energy consumptive element of the Internet is the manufacture and maintenance of its physical infrastructure, beginning with client devices (49%), Telecom infrastructure (37%), and data centers (14%; The Climate Group, 2008). For example, client devices, which refers to all of the devices we use to access the Internet (mobile phones, tablets, laptops, desktops, etc.) account for the highest proportion of energy consumption; depending on the study, estimates vary from 40% (Belkhir & Elmeligi, 2018) to 53% of the total energy consumption of the Internet (Raghavan & Ma, 2011). Most of this energy consumption is due to the manufacturing phase of client devices (referred to variously as either Embodied Energy or Grey Energy) and accounts for 45-80% of the total device life cycle energy (Hischier et al., 2015).
There are numerous complications with this measure and others like it because not all of these devices are used to access the Internet 100% of the time—someone can use a laptop to write a paper, for instance, without ever using the Internet, nor are device lifecycles consistent across all users—one person can use a phone for 6 months and another for 6 years. These are just some examples of the nuances that make definitive calculations about the carbon footprint of the Internet very difficult, if not impossible. (Ingrid Burrington touches on the difficulties these metrics pose in her article “The Environmental Toll of a Netflix Binge” in The Atlantic.) I found Soltani’s research commendable because despite the scarcity of data available and the opacity of the data that does exist, she has forged ahead and brought attention to this timely topic.
What was also surprising about her talk—and where she differs substantially from Burrington—is she remains optimistic about the overall positive impact that the Internet could have in reducing net carbon emissions in other sectors despite the Internet’s own significant contributions to global greenhouse gas emissions. While she herself strays away from a techno-optimist stance, she does cite some suspiciously sanguine (and perhaps outmoded) views, such as those from an optimistic 2008 report by The Climate Group: “The scale of emissions reductions that could be enabled by the smart integration of ICT into new ways of operating, living, working, learning and traveling makes the sector a key player in the fight against climate change, despite its own growing carbon footprint,” (The Climate Group, 2008). As an example of this, she mentioned that teleconferencing takes about 7% of the total energy cost that a face-to-face meeting would, taking into account factors like different modes of transportation, etc. (Ong et al., 2014).
She does concede that other examples of the “dematerialization” of information, including the move from traditional modes of music and movie distribution to digital streaming platforms, have less of a positive environment impact: depending on the study, streaming video is either only slightly more efficient than DVD distribution (Shehabi et al., 2014), or has an even higher net energy impact that is still rapidly increasing (The Shift Project, 2019). Whatever the exact figure is, it is highly impactful due to the fact that video streaming accounts for 64% of all internet traffic (Ejembi & Bhatti 2015).
The onus of finding environmentally sustainable solutions to this predicament we are all in should undoubtedly lay with the tech companies and governments with the greatest carbon footprints, and not on individuals. If one trusted the government to oversee the private sector, we could take inspiration from Paul Ford’s proposal to establish a government agency which he calls the Digital Environmental Protection Agency, responsible for protecting citizens in the event of repeated “data spills,” (Ford, 2018). How fitting, then, to also imagine tasking this hypothetical branch of a rotten bureaucracy with the additional task of disciplining the tech industry and cleaning up its messes.
Despite the absence of government legislation and private sector self-regulation, Soltani says that individual consumers can also take action. Her suggestions include extending the life of your client devices, using ad blockers (online advertising is very energy consumptive), protecting your data privacy (the storing of your personal data is also energy consumptive), and “stream lower quality videos, I guess,” (Soltani, 2019). All of these actions, save for electing to stream grainy YouTube videos, are all actions that have manifold benefits for consumers: less money spent on replacing devices and less personal data being collected, stored and sold to marketers and insurance companies at our citizenry’s expense.
While these actions she suggests may benefit individual consumers and make a small environmental impact, they do nothing to challenge the structural logic of late techno-capitalism and its extractivist methodology. Capitalism has always benefited from the dislocation of earth materials from Earth, the dematerialization of commodities and the invisibilization of labor. What is unprecedented at this stage of capitalism is that these existing abstractions of capitalist production have themselves become further abstracted and etherealized in the image of the Cloud. The semiotics of the Cloud further mystifies the terms of commodification and shrouds its mechanics in a blanket of mysticism. Divorcing the Internet from the materiality of the Internet in the image of the Cloud directly benefits the Internet’s profiteers and limits people’s ability to see the ideological machinery at work in their daily lives.
To uncloak this mantle of mysticism surrounding the Cloud, Nathan Ensmenger proposes treating the Cloud as a “type of factory” and interrogating it as such:
[W]hat kind of a factory is it? Who works there, and what kind of work to they do, and how is it different from the type of work previously performed by factory workers? Where does it fit in a larger technological, labor, and environmental history of human industry? And perhaps most importantly, how did it come to be seen as categorically different? (Ensmenger, 2018, p. 20)
This historical-materialist critique of the Cloud is a promising start towards resituating the Internet in its material, political, social, and cultural context. Only by bringing the Cloud back down to Earth can we begin to imagine a more equitable distribution of power in our hyper-networked reality.
This talk raised many questions I am still grappling with weeks later. As libraries and cultural centers move towards digitizing their assets and moving more and more services online, endeavors often hailed as universally beneficial and in line with our coupled missions to make information more accessible and to preserve it for posterity, we do so with little to no heed for how this impacts our shared environmental future. When environmental collapse comes to a head it won’t matter if our books are conserved in print format or preserved digitally. As long as the lifecycles of the hardware we’re using to preserve digital formats continue to physically deteriorate, and the software we use for the same mission continues to rapidly accelerate, information professionals are forced to continually endeavor on the perilous journey of continuous data migration, making us complicit in the whole system of filling the earth with toxic, obsolete electronic equipment and mining the same earth yet again for more rare minerals to inaugurate another terminally obsolete technological lifecycle, ad infinitum. Infinite, that is, until we run out of minerals.
“Looking from the perspective of deep time,” Kate Crawford and Vladan Joler write in their essay on the human labor and planetary resources required to operate an Amazon Echo, “we are extracting Earth’s history to serve a split second of technological time, in order to build devices that are often designed to be used for no more than a few years,” (Crawford & Joler, 2018). All media is an extension of the earth, be it codices made of paper manufactured from trees which took hundreds to thousands of years to grow, or be it a PDF viewed on a laptop composed of lithium, cobalt, and silicon (and the 14 other rare earth minerals necessary to manufacture a single laptop or smartphone) that took billions of years for Earth to produce. Where we now differ from Gutenberg’s time is the dizzying rate of acceleration at which we are moving towards total depletion of the earth materials needed to produce the information communication technologies that are embedded in the infrastructure of every branch of daily life.
Belkhir, L., & Elmeligi, A. (2018). Assessing ICT global emissions footprint: Trends to 2040 & recommendations. Journal of Cleaner Production, 177, 448–463. https://doi.org/10.1016/j.jclepro.2017.12.239
Burrington, I. (2015, December 16). The Environmental Toll of a Netflix Binge. Retrieved from The Atlantic website: https://www.theatlantic.com/technology/archive/2015/12/there-are-no-clean-clouds/420744/
Cook, G., Dowdall, T., & Wang, Y. (2014). Clicking Clean: How Companies are Creating the Green Internet [Greenpeace USA report]. Retrieved from Greenpeace website: https://www.greenpeace.org/usa/wp-content/uploads/ legacy/Global/usa/planet3/PDFs/clickingclean.pdf.
Crawford, K., & Joler, V. (2018, September 7). Anatomy of an AI System. AI Now Institute and Share Lab. Retrieved from http://www.anatomyof.ai
Efoui-Hess, M. (2019). The Unsustainable Use of Online Video. Retrieved from https://theshiftproject.org/en/article/unsustainable-use-online-video/
Ejembi, O., & Bhatti, S. N. (2015). Client-Side Energy Costs of Video Streaming. 2015 IEEE International Conference on Data Science and Data Intensive Systems, 252–259. https://doi.org/10.1109/DSDIS.2015.49
Ensmenger, N. (2018). The environmental history of computing. Technology and Culture, 59(5), S7–S33.
Ford, P. (2018, March 21). Facebook Is Why We Need a Digital Protection Agency. Bloomberg Businessweek. Retrieved from https://www.bloomberg.com/news/articles/2018-03-21/paul-ford-facebook-is-why-we-need-a-digital-protection-agency
Hammond, R. (2009, April). In search of Lithium: The battle for the 3rd element. Retrieved from https://www.dailymail.co.uk/home/moslive/article-1166387/In-search-Lithium-The-battle-3rd-element.html
Hilty, L. M., & Aebischer, B. (2015). ICT for Sustainability: An Emerging Research Field. In L. M. Hilty & B. Aebischer (Eds.), ICT Innovations for Sustainability (pp. 3–36). https://doi.org/10.1007/978-3-319-09228-7
Ong, D., Moors, T., & Sivaraman, V. (2014). Comparison of the energy, carbon and time costs of videoconferencing and in-person meetings. Computer Communications, 50, 86–94. https://doi.org/10.1016/j.comcom.2014.02.009
Raghavan, B., & Ma, J. (2011). The energy and emergy of the Internet. Proceedings of the 10th ACM Workshop on Hot Topics in Networks, 9. ACM. https://doi.org/10.1145/2070562.2070571
Shehabi, A., Walker, B., & Masanet, E. (2014). The energy and greenhouse-gas implications of internet video streaming in the United States. Environmental Research Letters, 9(5), 054007. https://doi.org/10.1088/1748-9326/9/5/054007
Soltani, J. (2019, October). The Carbon Footprint of the Internet. Presented at the Radical Networks Conference 2019, New York, NY.
The Climate Group. (2008, June 30). Smart 2020: Enabling the low carbon economy in the information age (pp. 1–87). https://www.theclimategroup.org/sites/default/files/downloads/annual-report-2007-2008.pdf