Giant tech companies with their power-hungry, football-pitched data centers aren’t the environmental villains they’re sometimes described as being on social media and elsewhere.
Turning off your Zoom camera or limiting your Netflix service to a low-definition display isn’t saving big on energy, as some people have claimed.
Even the predicted environmental impact of Bitcoin, which requires a lot of computing power, has been greatly exaggerated by some researchers.
These are the conclusions of a new analysis by Jonathan Koomey and Eric Masanet, two leading scientists in the field of technology, energy use and the environment. Both are former researchers at the Lawrence Berkeley National Laboratory. Mr. Koomey is now an independent analyst, and Mr. Masanet is a professor at the University of California at Santa Barbara. (Mr. Masanet receives research funding from Amazon.)
They said their analysis, released as a commentary on Thursday article in Joule, a scientific journal, was not necessarily reassuring. Instead, they said, it aims to inject a dose of reality into the public debate about the impact of technology on the environment.
The increase in digital activity spurred by the Covid-19 pandemic, scientists said, has fueled debate and raised serious warnings about environmental damage. They fear that capricious claims, often amplified by social media, could shape behavior and policies.
“We are trying to provide mental tools and guidelines for thinking about our increasingly digital lifestyles and the impact on energy use and the environment,” Mr. Masanet said.
The title of their analysis is “Do not calculate: avoid the pitfalls in evaluating the energy and carbon impacts of the Internet”.
The exaggerated claims, both men said, are often the well-meaning efforts of researchers who make what may appear to be reasonable assumptions. But they are unfamiliar with rapidly changing computer technologies – processing, memory, storage, and networking. When making predictions, they tend to underestimate the pace of innovation in energy conservation and the functioning of systems.
The impact of video streaming on network power consumption is one example. Once a network is operational, the amount of power it uses is about the same whether large amounts of data are flowing or very little. And constant improvements in technology are reducing electricity consumption.
In their analysis, the two authors cite information from two major international network operators, Telefónica and Cogent, which reported data traffic and energy consumption for the Covid year of 2020. Telefónica managed a jump of 45 % of data via its network without increasing energy consumption. Cogent’s electricity consumption fell 21% even as data traffic increased 38%.
“Yes, we are using a lot more data services and putting a lot more data on the networks,” Mr. Koomey said. “But we also become much more efficient very quickly. “
Another pitfall, the authors say, is looking at a high-growth sector of the tech industry and assuming both that electricity consumption is growing proportionally and that it is representative of the industry as a whole.
Computer data centers are a case study. The largest data centers, from which consumers and workers use services and software on the Internet, consume huge amounts of electricity. These cloud data centers are operated by companies such as Alibaba, Amazon, Apple, Facebook, Google and Microsoft.
From 2010 to 2018, data workloads hosted by cloud data centers increased by 2,600% and power consumption increased by 500%. But the power consumption of all data centers has increased by less than 10%.
What happened, the authors explain, was primarily a huge shift in workloads to larger, more efficient cloud data centers – and away from traditional data centers, largely owned and managed by non-tech companies.
In 2010, approximately 79% of data center computing was performed in traditional data centers. In 2018, 89% of data center computing was done in cloud data centers.
“The big cloud providers have moved much less efficient corporate data centers,” Koomey said. “You have to look at the whole system and take into account the substitution effects.
The complexity, dynamism and unpredictability of technological development and markets, say the authors, make projection over more than two or three years suspect. they criticized a Bitcoin energy paper who projected decades, based on what they said was old data and simplified assumptions – an approach Mr. Masanet called “extrapolating to Doomsday.”
But Bitcoin, scientists say, is something different – and a worry. Efficiency trends elsewhere in tech are blunted as Bitcoin’s specialized software goes through more IT cycles as more people attempt to create, buy, and sell digital currency.
“It’s a hot spot that needs to be watched very closely and could be a problem,” Mr. Masanet said.
Much is not known about cryptocurrency mining and its energy consumption. It uses specialized software and hardware, and secrecy surrounds the major centers of crypto mining in China, Russia, and other countries.
Thus, estimates of Bitcoin’s energy footprint vary widely. Researchers from Cambridge University estimate that Bitcoin mining accounts for 0.4% of global electricity consumption.
It may not seem like much. But all of the world’s data centers except those for Bitcoin mining consume around 1% of its electricity.
“I think that’s a very good, high-value use of that 1%,” Mr. Koomey said. “I’m not sure the same is true for Bitcoin’s part.”