Picking up on my China article, I pulled a few links together to see how much energy the US is using to power our computing infrastructure. Apparently there was quite a debate in 1999, when Mark Mills kicked off by writing an article entitled "The Internet Begins with Coal", claiming the Internet used 8% of all electricity, and office equipment 13 percent. Turns out it was not so accurate; the Internet was only 1 percent, office equipment maybe 2-3 percent - damning slide deck here. Must have been pretty tough on the publisher, as I can't find the original article anywhere, but there are plenty of I told you sos - here too.So if we are blowing 2-3 percent of our energy on office equipment, where does the rest go? Well, looks like 21 percent gets poured into homes. Doing a little math from this site, that's about 2 percent for air conditioning, 9 percent for space and water heating, and 8 percent to run appliances. Oh, and 2 percent is completely wasted through vampire energy. These are national numbers.
Tack on 28 percent for shuttling ourselves around.
Continuing, about 0.6 percent of our national energy to used to make cement. 15 percent goes to make food to eat. We blow about 3 percent making aluminium, wow. Lets see what we got:
- Transportation - 28
- Food -15
- Household Appliances - 8
- Internet - 3
- Household Air Conditioning - 2
- Household Heating - 9
- Aluminium - 3
- Household Vampire Energy - 2
- Cement - 0.6
5 comments:
More energy is spent for Internet use than heating and air conditioning? That's an eyeopener.
Hi Lynn, More than air conditioning, but less than heating. The "Internet" figure includes all the hardware (servers, pcs, etc) amd all the telecom too (routers, switches, etc.)
thanks for posting!
Actually, to clarify, that's just residential air conditioning; commerical and industrial are not included.
Mark - I think I found the damning article - see the 4th link in the appended note I put together 2 days ago - Peter
---------
Gartner estimates that within five years between twenty and forty percent of IT expenses will be for electricity (http://hardware.silicon.com/servers/0,39024647,39163893,00.htm).
There are competing points of view on this. A 1999 Forbes article asserted that US IT energy usage was about 10% of total energy usage. The Forbes article (https://www.keepmedia.com/Auth.do?extId=10022&uri=/archive/forbes/1999/0531/6311070a.html) requires a subscription, but essentially the same thing is said in a 2001 Salon article (http://archive.salon.com/tech/feature/2001/01/17/net_power/index.html).
However, a 2000 report from the Lawrence Berkeley National Laboratory (http://enduse.lbl.gov/Info/annotatedmillstestimony.pdf) puts the energy percentage closer to 3%.
So, there's quite a spread between the lowest and highest estimates. However, whether the estimate is 3% or 10%, usage will grow. And energy costs per megawatt-hour will continue to rise, too.
Unfortunately, growth estimates range from "minimal" to "60% since 1980" (although it wasn't clear if that 60% was ALL electricity or just IT electicity).
There's been recent movement in Congress that may result in EnergyStar ratings for servers (http://www.itjungle.com/tfh/tfh010807-story03.html, discussing H.R. 5646).
The article below discusses usage, current wastage, and various mitigations, and gives a good background on the issue.
In the light of H.R. 5646, which likely will eventually dictate increased efficiencies, and the increasing percentage of IT cost going to energy in the private sector, we will certainly be addressing this in various ways--virtualization, low-power technology developments, data center layouts, and so forth.
Hi Peter,
Thanks for posting, and for the great research. It's clear that there are a lot of numbers being thrown around out there. At the present date, the 3 percent figure to run the Internet is almost certainly too low, I personally think it's probably 4 or 5 percent now.
As you mention the prices are going to keep going up, and with the recent policies to subsidize liquified coal and drill in the Arctic, it seems we are going to try and have capital trump resources, at least for a while; we are going to pour money in dredging up more energy. Ultimately I think this strategy is bankrupt for, as sridhar pointed out on Charlie Rose, resources always end up trumping capital in the long run. And the run may not be that long.
The alternative, conservation, doesn't seem popular yet. IT folks need to tout it as the best and only way going forward.
mark
Post a Comment