US Internet Power Bill Growing: $2.7 billion in the USA, 45 million kWh in 2005.
Hydrogen TRUTH
Energy bill for data centres hits the roof
By Manek Dubash, Techworld
The energy demands of modern data centres are huge and growing,
according to a new study.
A study by the US Lawrence Berkeley National Laboratory (LBNL),
released today and funded by chipmaker AMD, found that servers in the
US consumed 45 million kWh in 2005.
That's almost one-third of the world's total, and more than the total
for US state of Mississippi or 19 other US states, reckoned LBNL
scientist Jonathan Koomey. The global total was 123 million kWh, and
the annual energy bill was $2.7 billion in the USA, and $7.2 billion
globally.
Koomey, who is also a consulting professor at Stanford University,
said that demand was growing fast, having doubled in the five years to
2005. He said he expected this level of energy consumption to increase
by 2010 by 40 per cent, a figure he calculated by extrapolating from
server sales forecasts provided by research firm IDC. The assumption
was based on server power consumption remaining static -- if it rises
at the same rate as before, by 2010 demand will have risen by 75 per
cent. Most of the growth came from low-end servers, he said.
Koomey said the study included consumption by infrastructure such as
air-conditioning and lighting, but not data processing equipment such
as storage arrays or network switches, which increases the total about
one-third.
"Though we have long known that data centres worldwide consume a
significant amount of energy, AMD believes Dr. Koomey's findings are a
wake-up call not just for the IT industry, but also for global
business, government and policy leaders," said AMD's Randy Allen, vice
president for servers and workstations.
"This study demonstrates that unchecked demand for data centre energy
use can constrain growth and present real business challenges. New
generations of energy-efficient servers are now able to help provide
IT departments with a path to reduce their energy consumption while
still achieving the performance they require."
Koomey's study was subject to peer review by IT industry, government
and energy efficiency policy professionals, said AMD, which plans to
share the study's findings at an industry stakeholder workshop in
Silicon Valley tomorrow, 16 February. Organized by the US Government's
Environmental Protection Agency (EPA), the workshop is designed to
explore technical issues relating to a study mandated by the US
Congress and requiring the EPA, "to study and promote the use of
energy-efficient computer servers in the United States."
AMD didn't say that it expected its processors, which initially led
the way towards lower CPU power consumption between 2003 and 2006, to
make a significant difference to this level of power consumption --
which might lead one to assume that the company didn't believe it
could make that claim with any credibility.
Vendicar Decarian
"Hydrogen TRUTH" <Hydrog...@HydrogenTRUTH.info> wrote
> A study by the US Lawrence Berkeley National Laboratory (LBNL),
> released today and funded by chipmaker AMD, found that servers in the
> US consumed 45 million kWh in 2005.
>
> That's almost one-third of the world's total, and more than the total
> for US state of Mississippi or 19 other US states, reckoned LBNL
> scientist Jonathan Koomey. The global total was 123 million kWh, and
> the annual energy bill was $2.7 billion in the USA, and $7.2 billion
> globally.
The power output from a typical generating station is in the GigaWatt range.
So one hour of operation will produce 1,000 MegaWatt Hours of power, or
roughly 20 times the annual worlds server total stated in the first
paragraph.
In 2005 the U.S consumed close to 4 billion, MegaWatt hours of electrical
energy.
I would estimate about 100 million servers running in the U.S. each
consuming 100 watts of power. This totals 87600 MegaWatt Hours of energy
per year or about .002% of the U.S. total.
The reporter is innumerate.
T. Keating
The reporter got his facts wrong about the level of energy
consumption, it's supposed to be 45 Billion kWh.
here is a link to an article quoting the correct numbers
http://www.physorg.com/news90842269.html
"U.S. Data Centers Consume 45 Billion kWh Annually, Study"
http://www.amd.com/us-en/Corporate/VirtualPressRoom/0,,51_104_543~115850,00.html
http://enterprise.amd.com/us-en/AMD-Business/Technology-Home/Power-Management.aspx
http://enterprise.amd.com/Downloads/svrpwrusecompletefinal.pdf
Vendicar Decarian
"T. Keating" <tkus...@ktcnslt.com> wrote in message
news:c7ta03pa477cvsrge...@4ax.com...
He not only got the Amount of energy used wrong, He made numerous errors
in his calculations to come up with the "factoid" that 1/1000th of the real
consumption is 1/3rd of the worlds total.
What is 45 billion/4 trillion? 0.01125 or 1.125% not 33%
AmeriKKKans are innumerate. I see these kinds of grade school errors
being made all over the place. The AmeriKKKan public are simply too stupid
to catch them.
R.H. Allen
> On Fri, 23 Mar 2007 16:37:35 -0800, "Vendicar Decarian"
> <BushIsA...@hotmail.com> wrote:
>
>> "Hydrogen TRUTH" <Hydrog...@HydrogenTRUTH.info> wrote
>>> A study by the US Lawrence Berkeley National Laboratory (LBNL),
>>> released today and funded by chipmaker AMD, found that servers in the
>>> US consumed 45 million kWh in 2005.
>>>
>>> That's almost one-third of the world's total, and more than the total
>>> for US state of Mississippi or 19 other US states, reckoned LBNL
>>> scientist Jonathan Koomey. The global total was 123 million kWh, and
>>> the annual energy bill was $2.7 billion in the USA, and $7.2 billion
>>> globally.
>> The power output from a typical generating station is in the GigaWatt range.
>> So one hour of operation will produce 1,000 MegaWatt Hours of power, or
>> roughly 20 times the annual worlds server total stated in the first
>> paragraph.
>>
>> In 2005 the U.S consumed close to 4 billion, MegaWatt hours of electrical
>> energy.
>>
>> I would estimate about 100 million servers running in the U.S. each
>> consuming 100 watts of power. This totals 87600 MegaWatt Hours of energy
>> per year or about .002% of the U.S. total.
>
> The reporter got his facts wrong about the level of energy
> consumption, it's supposed to be 45 Billion kWh.
Even so, that's only 1.1% of US electricity consumption. To once again
use Mississippi as a point of comparison, about 1.0% of US electricity
customers are in that state
(http://www.eia.doe.gov/cneaf/electricity/epa/fig7p1.html). It really
doesn't seem like an excessive amount of electricity consumption,
particularly for an industry whose only real consumable is electricity.
It isn't very surprising, though -- for precisely that reason, the IT
industry has always been pretty conscious of its power consumption.
That's not to say that there isn't room for improvement, but there
probably isn't as much room as the article seems to imply.
R.H. Allen
And apparently you're illiterate. The article rather clearly states that
the amount is one-third of the total *used to power servers* worldwide.
The author even states the global total. Using it to replicate the
author's calculation:
45 billion kWh / 123 billion kWh = 0.366
Looks pretty close to a third to me.
Vendicar Decarian
"R.H. Allen" <kka...@hotmail.com> wrote in message news:0-
> And apparently you're illiterate. The article rather clearly states that
> the amount is one-third of the total *used to power servers* worldwide.
> The author even states the global total. Using it to replicate the
> author's calculation:
"A study by the US Lawrence Berkeley National Laboratory (LBNL),
released today and funded by chipmaker AMD, found that servers in the
US consumed 45 million kWh in 2005.
That's almost one-third of the world's total, and more than the total
for US state of Mississippi or 19 other US states, reckoned LBNL
scientist Jonathan Koomey."
Sorry, Rereading it produces the same result. 45million kWh = 1/3rd of
world total energy consumption.
What a liar.
R.H. Allen
So you think 30 US states *each* use more than a third of the world's
total electricity consumption? Why, that would mean that the US uses at
least 10 times the worldwide total, but that's what your reading
implies. But then, you evidently don't know the difference between
energy and electricity, so I suppose I ought to cut you a break....
Vendicar Decarian
> Vendicar Decarian wrote:
>> "R.H. Allen" <kka...@hotmail.com> wrote in message news:0-
>>> And apparently you're illiterate. The article rather clearly states that
>>> the amount is one-third of the total *used to power servers* worldwide.
>>> The author even states the global total. Using it to replicate the
>>> author's calculation:
>>
>> "A study by the US Lawrence Berkeley National Laboratory (LBNL),
>> released today and funded by chipmaker AMD, found that servers in the
>> US consumed 45 million kWh in 2005.
>>
>> That's almost one-third of the world's total, and more than the total
>> for US state of Mississippi or 19 other US states, reckoned LBNL
>> scientist Jonathan Koomey."
>>
>> Sorry, Rereading it produces the same result. 45million kWh = 1/3rd of
>> world total energy consumption.
>>
>> What a liar.
"R.H. Allen" <kka...@hotmail.com> wrote
> So you think 30 US states *each* use more than a third of the world's
> total electricity consumption?
No I think the author is innumerate, and that AmeriKKKans are too stupid
to notice.
I also think that it states that it's false statement of usage is claimed
to be larger than the total for the state of Mississippi. and larger than
the total of 19 other U.S. states.
This is somewhat different than your interpretation that it is smaller
than 30 U.S. states.
In fact the number 30 doesn't appear anywhere in the article at all.
So you have the interpretation backward, have the count of states wrong.
Fucking innumerate AmeriKKKans.
"R.H. Allen" <kka...@hotmail.com> wrote
> But then, you evidently don't know the difference between energy and
> electricity, so I suppose I ought to cut you a break....
"Electricity" is a reasonably generic term that refers to statically
displaced or motile electrons that are either conducted from one place to
another or which are deposited on a surface as a fixed unbalanced electric
charge.
Electrical energy is defined as the amount of energy acquired by an
electron as it passes through a potential difference of 1 volt. This energy
is called (astonishingly) an ElectronVolt. In conventional bulk terms, the
amount of Power that is carried by a current I through a resistance R P =
I**2*R. Alternately P = IE Energy = IET or I**RT
Often electrical energy is transferred in a sinusoidal form, in which
case if I is the peak current then the IET must be integrated over one sine
wave and that causes the inclusion of a constant sqt(2)**-1 to be added as a
coeffiient.
When working with capaitors or inductors where the phase of voltage and
current are not matched, then one computes a "power factor" correction and
applies that as a coefficient.
Computing the power factor requires the use of complex numbers and some
vector addition, which I don't care to go into here. as your inferiority has
been aptly demonstrated here Mr. Allen.