There’s also a parable about predicting the future: the contrast of using extrapolation versus “bottom-up” analysis.
Data, Data, and More Data. In the last decade, there has been immense growth in data center activity around the world. These banks of computers and their associated infrastructure have been consuming around one percent of worldwide electrical use.
It’s more than just moving around digital bits. For example, as noted by Steve Lohr in “Cloud Computing is Not the Energy Hog That Had Been Feared,”The New York Times, February 27, 2020, “Each of these sprawling digital factories, housing hundreds of thousands of computers, rack upon rack, is an energy-hungry behemoth. Some have been built near the Arctic for natural cooling and others beside huge hydroelectric plants in the Pacific Northwest.”
Image from The New York Times, February 27, 2020.
Masanet and his colleagues observe, “Several oft-cited simplistic analyses claim that the energy used by the world’s data centers has doubled over the past decade and their energy use will triple or even quadruple within the next decade.” Thus, the fear of an electrically browned-out future.
Extrapolation vs Bottom-Up Analysis. However, the researchers note that such predictions have been based on “top-down” assessments depending primarily on extrapolation. By contrast, “bottom-up” assessments look at individual factors and better reflect the dynamics of such quickly evolving technology.
For example, every measure of data center utilization has grown over the period 2010–2018: Global storage capacity is 26 times larger. Global traffic has grown by a factor of 11. Work loads are 6.5 times greater. And the number of servers is greater by a factor of 1.3. The industry is clearly a bullish one.
Image from Science, February 28, 2020.
However, it has also used energy more efficiently. For example, Masanet and his colleagues note “the combination of increased storage-drive efficiencies and densities has enabled a 25-fold increase in storage capacity with only a threefold increase in global storage energy use.”
The sixfold growth in “compute instances,” i.e., episodes of in/execute/out, came with only a 25-percent increase in global server energy use.
The researchers observe, “Shifts to faster and more energy-efficient port [in/out] technologies have enabled a 10-fold increase in data center IP [Internet Protocol] traffic with only modest increases in network device energy use.”
Enhanced technologies of power supply and cooling systems have steadily increased the Power Usage Effectiveness of data centers. PUE measures the total amount of energy used by a data center divided by the energy used by its information-technology equipment.
Researchers say these trends represent “notable improvement compared with recent annual efficiency gains in other major demand sectors (e.g., aviation and industry), which are an order of magnitude lower.”
The Cloud is Green.“The Cloud,” as it has come to be known, replaces personal computing and local servers with often huge remote servers. Masanet and colleagues note that many improvements are “explainable by ongoing shifts in servers away from smaller traditional data centers (79 percent of compute instances in 2010) and toward larger and more energy-efficient cloud (and hyperscale) data centers (89 percent of computer instances in 2018).”
“On the infrastructure side,” researchers continue, “world-class hyperscale data centers are already operating with PUEs of 1.1 or lower, which is close to the practical minimum value.”
How to Keep It Green. Masanet and his colleagues recommend three strategies for policy makers: First, help data centers to “seize the remaining efficiency potential of current technology and structural trends.” One way is promoting Energy Star standards of electrical efficiency. Second, “investment in new technologies is needed… once current efficiency trends reach their feasible limits.” Examples include artificial intelligence applied to managing resources and infrastructures as well as the evolving state of quantum computing. Third, national policy makers should get involved “in much the same way as has been done historically for other demand sectors…. Global data center energy use is entering a critical transition phase; to ensure a low-carbon and energy-efficient future, we cannot wait another decade for the next reliable bottom-up estimate.” ds