1. 5

Headline overpromises on content; subtitle is “The data centers that support the Internet use a huge amount of energy.”

  1.  

  2. 5

    I don’t regret reading this, but I’m disappointed it didn’t tell me the promised fact. What is the environmental toll of a Netflix binge? Before streaming, I would watch two DVDs a week. How much carbon does it cost to ship plastic discs around vs an always on streaming datacenter? I also spend far more time watching TV now, streaming perhaps four hours a day (while doing other things). My local costs have obviously increased, too, though I have a handle on calculating them.

    I think I was vaguely promised that streaming was environmentally friendlier, but is that true? Or out of sight, out of mind?

    1. 3

      I don’t have the numbers on hand to answer this, but I think the long answer is that the environmental cost of technology is very complicated, because you often have high fixed costs and low variable costs. When I read that a Google search consumes as much energy as boiling N cups of coffee, I don’t know if the (amortized) fixed costs are included, and that makes a huge difference. Are we counting the additional energy in the data center, or my proportional share in all of it? Are we including the considerable energy used by ad-related infrastructure that, while I recognize its necessity in funding much of the web, I don’t specifically want? Are we including the amortized cost to build the data center and the fossil fuels involved in making the computers? And how about the daily commutes of tens of thousands of people? Oh, and while we’re at it, are we including the cost to make the undersea cables and launch the satellites that make the internet possible? Finally, do we hold the consumer or the provider morally liable for inefficient code? Not to take this on too much of a tangent, but the argument could be made that “Agile”/Scrum (designed to flood the market with mediocre “commodity programmers” who’ll lead your startup directly to $50,000/month AWS and Heroku bills) is an environmental catastrophe.

      All in, I think that technological solutions are a win, and that’s why we’re using them. Transporting humans is extremely expensive, financially and environmentally, and that’s what business used to rely on. Now, we’ve replaced an army of clerks doing mind-numbingly boring work on-site (and having to burn up fossil fuels in order to get to work) with a data center that doesn’t need to be cooled because it’s in Iceland which rarely breaks 15°C, even in summer, and uses natural cooling. We might consume more search (which used to involve a person looking through filing cabinets) and media, but we’re using far less per item to do so.

      We have other problems– technological unemployment is hitting us hard, and we don’t seem to be socially evolved (at least, not in the US, where even basic amenities like universal healthcare can be decried as “socialist”) enough to handle it– but our GDP-per-joule statistics have been improving for a long time.

    2. 5

      Using Google’s 0.01 percent estimate and electricity-consumption data from the CIA World Factbook, they’re using about as much electricity annually as the entire country of Turkey. (Honestly, that number seems impossibly high considering that in 2011 Google disclosed that it used merely 260 million watts of power, at the time noted for being slightly more than the entire electricity consumption of Salt Lake City.)

      Not to be a pedant, but I think this is important: the Turkey comparison is not quite right. Turkey is #20 with 1% (not 0.01%) of global electricity consumption. 0.01% would be #143 Madagascar. Turkey has about 75 million people and a fairly typical standard of living for Eastern Europe, so the idea that it would clock in at 0.01% of the global electricity budget set off my detector. I imagine this was an honest mistake, but it’s off by 2 orders of magnitude.

      Also, 260 million watts is 2.28 billion killowatt-hours per (non-leap) year, which comes in right between 0.01 and 0.02 percent of cited global consumption.

      1. 5

        The data centers that support the Internet use a huge amount of energy

        Compared to what? I’d be interested to see an aggregate of how much energy was used for activities that we now use the internet; letters, photographic film, newspapers, books & encyclopaedias and so on.

        I’d be willing to wager that things are much more economical and efficient now than before.

        1. 2

          That would only work if those things had disappeared completely. People can still write letters, use film, read newspapers, books, and encyclopedias, so we’re still using all the resources for those, and then additional resources for data centers.

          I’d be willing to wager that things are much more economical and efficient now than before.

          Why? Is there any evidence to support that?

          1. 1

            we’re still using all the resources for those

            I don’t think we’re using anywhere near the amounts that we used to. Rolls of film used per capita, for instance.

            Why?

            On an intuitive level, a smartphone can perform all the functions that once took several devices (phones, radios, cassette players etc) and new functionality can be added thru software instead of manufacturing a new device.

            1. 1

              Why? Is there any evidence to support that?

              Monetary cost tends to correlate (noisily) with environmental cost, and the monetary cost of accessing information via the Internet is much less than doing same without.

            2. 1

              Isn’t the point of the article that we have drastically underestimated these costs because they are hidden? I would have liked more numbers, too, but some of the ones here are pretty crazy.

              Consider the cited stat that Google uses 0.01% of the worlds energy. That puts a rather low bound on the number of googles the planet can sustain.

              1. 1

                Isn’t the point of the article that we have drastically underestimated these costs because they are hidden?

                Yes, but I also think that costs were even more diffuse in the past.

                It would be good to see more research into this!