The Environmental Cost of Electronic Collaboration

An ex-Cisco colleague of mine called earlier this week to talk about his concern that distance collaboration technologies were not as ecologically friendly as they could be. His concern was that by constantly increasing the fidelity of technologies over time to approximate in-person interactions, that the end-point and infrastructure equipment required would increase emissions beyond what was necessary to complete the task. One quote of his stuck with me...."For a staff meeting, you don't need immersive telepresence and all the cost and environmental impact. You just need a phone call."

Not to be a neo-Luddite about collaboration, but he has a point.  You don't always need the highest-fidelity technology available to you, you need the appropriate one for the task at hand.  With Internet electricity use near or over 100 terawatt hours in 2009, and the high electricity consumption of new endpoint devices, there should be a 'pause' before we mandate holographic immersive virtual reality conversations with our co-workers about every little thing.

This was a threshold concept for me, having traditionally been an advocate of using the highest-fidelity collaboration technology available for each interaction.  What I hadn't synthesized was my environmentally-friendly views with these collaboration tools, as I had been so focused on eliminating physical travel, that the cost of emissions from electronic collaboration were lost in the noise.  I decided to look around at the research to see what if any the energy costs were. It turns out, there is a cost, and it is very real. 

Here are some hard numbers to put things in perspective:

In 2008, two researchers estimated energy use of the Internet, and discovered that the energy use of the Internet had grown from 39 TWh/year to 85 TWh/year (2.2x), while at the same time data traffic had grown by over 20 times (22.5-22.9x).  What does this tell you?  We are doing more work using less electricity, but overall electricity consumption is up. 

If one were to assume that a portion of the uses of distance collaboration tools were substituting for in-person meetings, then there is a fractional savings in emissions from air-travel and such.  Emissions from air travel are also improving, from 37.9 Passenger Miles per Gallon (equivalent of 3300 BTUs per Passenger Mile (PM)) in 2000 to 45.6 PM per gallon/2740 BTU per PM in 2005.  That's a laudable increase in an industry that replaces it's equipment infrequently at best.

This only captures a portion of the environmental impact of travel from home.  The oft-quoted Environmental Protection Agency, speaking about conference attendance, estimated:

• Each conference attendee typically consumes 8 gallons of jet fuel and consequently contributes 0.50 metric tons of CO2 into the atmosphere by plane travel.
• By staying at a hotel, the convention goer generates 20 pounds of trash per day compared to 4.6 pounds at home
• The typical conference goer is responsible for using 200 gallons of water per day.

This doesn't even take into account the healthcare costs of all of the liver damage sustained by those 'breakout meetings' in the bar.

These are huge impacts, no question.  They make distance collaboration technologies look like shining beacons of light in the harsh emissions fog of physical travel.  However, before we begin patting our avatars or webcams for a job well done, lets look at some more numbers......

Broadband access electricity consumption, for the increasing number of people telecommuting on a daily or weekly basis, ranges from .25TWh/yr for DSL to 2.72 TWh/yr for DSL to Cable Modems respectively.  This doesn't factor in the 90kWh per year your wireless router consumes, or the 190-868 kWh your computer gobbles up. 

As we migrate to more and more Internet access devices, this is a hidden cost.  I can count at least 20 Internet-connected devices in my home, ranging from the Prius-like Chumby in our kitchen (using 3.2 watts, less than my idle microwave!) to the electric Humvee of my household....our PlayStation 3 (which guzzles an astonishing 1200kWh per year given my family's household usage). 

Lets put this into perspective.....If I wanted to participate in a one-hour distance collaboration meeting using Sony's Home environment, it would consume nearly five times as much electricity as using my iMac in Second Life.  This comes from not only the gluttonous energy consumption of the PS3, but also the home theater receiver, speakers, and television that are required to use the service (discounting the broadband equipment as common in either scenario). There are some comparable numbers for other televisions, gaming consoles, etc. here.

Jay Walker walked on stage at TED 2008 and stunned the crowd by reminding them that each megabyte of data they downloaded cost the equivalent of a lump of charcoal to transfer.  Think about your Internet browsing today, the bandwidth and downloads, and then imagine how many bags of charcoal briquettes you consumed. It certainly gave me pause.

To wrap on an optimistic note, we are reaping the benefits of the great work done in energy efficency as well as endpoint functionality at a relatively static energy cost. Between 2000 and 2006, the electricity required to transfer a gigabyte of data across the Internet dropped 10-fold from 126kWh/GB to 12.5KWh/GB.  This doesn't absolve us from the responsibility to remember the hidden costs in infrastructure and devices.  Just because we're collaborating electronically instead of in-person, there is still an environmental cost.  Perhaps before you fire up that avatar or videoconference, take a pause and think about those bags of coal. You may just pick up the phone.

Comments

[Tele3dworld]

Web.alive caches the virtual world on to the hard drive exactly like a web page; so this point is not true for all the virtual world platforms. Probably true for 'Second Life' though.