Beta
×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

Using the Sea To Cool Your Data Center

ScuttleMonkey posted more than 4 years ago | from the bending-mother-nature-to-your-will dept.

Data Storage 194

1sockchuck writes "We haven't yet seen signs of the Google Navy of seagoing data centers that use the ocean for power and cooling. But data center developers are planning to use sea water air conditioning in a new project on the island nation of Mauritius in the Indian Ocean. Cold water from deep-sea currents would be piped ashore to be used in a heat exchanger for the data center facility. A similar system has been used to replace the chillers at Cornell University, which draws cold water from Lake Cayuga. The Cornell system cost $50 million, but has slashed cooling-related energy usage by 86 percent."

cancel ×

194 comments

Sorry! There are no comments related to the filter you selected.

interest prospect (3, Insightful)

Icegryphon (715550) | more than 4 years ago | (#29496747)

But what are maintenance costs and lifespan of such a piece of equipment,
I can't image Saltwater not eating the hell out of all the piping.

Re:interest prospect (5, Funny)

sopssa (1498795) | more than 4 years ago | (#29496811)

I can't image Saltwater not eating the hell out of all the piping.

Yeah, thats the real problem. I hope we discover such metal soon so we can get boats and ships in the oceans too.

Re:interest prospect (1)

zach_the_lizard (1317619) | more than 4 years ago | (#29497081)

I sure hope they get that fixed quickly, I'm tired of the tubes that connect to my modem rusting out so often. It's very frustrating.

Re:interest prospect (2, Informative)

digitalunity (19107) | more than 4 years ago | (#29497119)

Most steel ships are painted to prevent corrosion. Paint is a thermal insulator. Coating the inside of your heat transfer pipes with a thermal insulator is like masturbating with sandpaper - it might work, but it doesn't work well.

Aluminum is a great thermal conductor and is saltwater resistant with the 6061 and 6063 alloys. Galvanic corrosive action does occur though, but this can be avoided with careful attention to construction methods and avoiding direct metal to aluminum contact.

Re:interest prospect (1)

saforrest (184929) | more than 4 years ago | (#29497367)

Most steel ships are painted to prevent corrosion. Paint is a thermal insulator. Coating the inside of your heat transfer pipes with a thermal insulator is like masturbating with sandpaper - it might work, but it doesn't work well.

My kingdom for mod points...

not a thermal insulator and heat tax (5, Insightful)

goombah99 (560566) | more than 4 years ago | (#29497545)

Other than a set up for your gag, I don't see why you call paint a thermal insulator. It does not have to be so. many kinds of coating promote thermal coupling.

One thing that does bother me is dumping waste heat in someone elses backyard for free promotes the inefficient use of energy. that is, I can decrease my cooling costs by using more efficient but more expensive computers which incidentally produce less waste heat, or I could use less expensive inefficient computers and take advantage of public domain cooling, like cayuga lake.

Is Cornell paying a tax to use Cayuga lake as a heat dump? that would help internalize the economic externalities that drive them to consume more electricity because the cooling is free.

likewise for sea water cooling.

This might seem like worry much about a small thing: isn't the cooling resevoir comparatively infinite? the answer is surprising no, not only is it not infinite, it's never going to grow, and we have already saturated it in much or the US and Europe. For example the big limit on Nuclear power plant growth is now availability of cooling. SOme rivers in Tenesee are known to heat up to 80 degrees when the power plants operate a full power in summer.

thus this needs to be publicly regulated now.

Re:not a thermal insulator and heat tax (0)

Anonymous Coward | more than 4 years ago | (#29497633)

Mod parent up.

Re:not a thermal insulator and heat tax (1)

digitalunity (19107) | more than 4 years ago | (#29497651)

Paint doesn't have to be a thermal insulator, but many are. Like adhesives, there is a paint for everything.

See my other comments re: ecological impact. You're right - the impact is non-negligible, even in seawater.

Re:not a thermal insulator and heat tax (0)

Anonymous Coward | more than 4 years ago | (#29498099)

You say that like organisms living in the water might object to a drastic change in the temperature of their environment.

They can't object after we cook them in place.

Re:not a thermal insulator and heat tax (1)

registrar (1220876) | more than 4 years ago | (#29498309)

One thing that does bother me is dumping waste heat in someone elses backyard for free promotes the inefficient use of energy.

Ok, but for a given amount of heat generated, you may as well get rid of it in the most efficient way possible. "Free" i.e. cheap generally means less energy is expended in the dumping of the heat, and that's a good thing for everyone.

Re:interest prospect (2, Informative)

MicktheMech (697533) | more than 4 years ago | (#29497595)

Aluminum is a great thermal conductor and is saltwater resistant with the 6061 and 6063 alloys.

The 6000 series alloys are also extremely expensive compared to steel and more importantly difficult to weld, even compared to stainless.

Re:interest prospect (1)

fightinfilipino (1449273) | more than 4 years ago | (#29497675)

Coating the inside of your heat transfer pipes with a thermal insulator is like masturbating with sandpaper - it might work, but it doesn't work well.

so that's what i've been doing wrong the entire time!

Re:interest prospect (1)

fireball84513 (1632561) | more than 4 years ago | (#29498095)

Paint is a thermal insulator. Coating the inside of your heat transfer pipes with a thermal insulator is like masturbating with sandpaper - it might work, but it doesn't work well.

i can just imagine some retard reading your comment and then taking a piece of sandpaper to his dick thinking "it might work"

Re:interest prospect (2, Interesting)

FooAtWFU (699187) | more than 4 years ago | (#29496813)

You keep the saltwater on one side of a heat exchanger; it minimizes the vulnerable piping, and helps a lot. Heck, you could build the big seawater pipes out of concrete.

Re:interest prospect (1)

Furmy (854336) | more than 4 years ago | (#29497747)

Why not move the heat-exchanger (thin walls and all) into the ocean. Have a closed-loop of FRESH water that you pump around. It cools down in the ocean and then is pumped back in, through the data centre, heats up. Now you've got internals (and your pump) free of salt. Yes, you still have to maintain the outside of all the pipes AND the ocean-side of the heat exchanger but that seems easier than moving salt around.

Re:interest prospect (1)

Firethorn (177587) | more than 4 years ago | (#29498275)

Not necessarily a bad idea. You can even use plastics and concrete instead - sure, they're poor conductors, but water's so good that you just build it bigger.

I watched a special on the live kelp aquarium, they pump LOTS of seawater and it's very maintenance intensive - they need to run special pigs through the pipes to keep them clean and hire divers to scrub the outsides.

In something like this cheap construction cost and corrosion resistance is probably a bigger concern than thermal coupling efficiency.

Re:interest prospect (4, Informative)

Useful Wheat (1488675) | more than 4 years ago | (#29496861)

A low carbon stainless steel such as the 316 series should be more than sufficient for any piping. Moving parts such as pumps and impellers would be made of titanium for optimum durability and minimum downtime. Lifetime of the pipes is assured by simply adding a small corrosion allowance to the wall thickness (maybe 1/4"), and checking for corrosion once in a while to make sure its not being destroyed faster than you predict. Although that may sound ridiculous, I promise you it is both fairly common and not that hard. Seawater is the lifeblood of many power plants, and it doesn't take a miracle to handle it.

Re:interest prospect (5, Insightful)

initdeep (1073290) | more than 4 years ago | (#29496943)

seawater is the lifeblood of every naval nuclear power plant, and as someone who was in the navy and in charge of the heat exchangers attached to a naval nuclear power plant, i can assure you it is a big deal and a LOT of time and maintenance is put into preventing corrosion and the associated leakage in piping that a heat exchanger utilizes.
In order to have efficient heat exchange between two moving fluids, you need a very thin wall and you need it to be clear of any and all corrosion. This means a lot of time and effort, not too mention chemicals are used.
For a mobile naval vessel, there is no other option, so the cost isn't an issue.

For a land based cooling system, it is an issue because there very well may be less expensive alternatives.

Not too mention the possible ramifications (good and bad) of discharging all of the heated water back into the marine ecology.

Re:interest prospect (1)

digitalunity (19107) | more than 4 years ago | (#29497191)

The ecological impact was my first reaction. Rising ocean temperatures has been shown to increase toxic cyanobacterial algae bloom production. Heating water with megawatts of power and pumping it back into the ocean could have negative localized ecological effects.

As far as stainless steel, sure 316 is resistant, but it's a lot more expensive than 6061 aluminum alloy.

Re:interest prospect (1)

CodeBuster (516420) | more than 4 years ago | (#29497271)

Maybe this is a stupid question, but why can't the surfaces that are exposed to the corrosive fluid (i.e. saltwater) simply be painted with a corrosion resistant paint?

Re:interest prospect (0)

Anonymous Coward | more than 4 years ago | (#29497369)

Maybe this is a stupid question, but why can't the surfaces that are exposed to the corrosive fluid (i.e. saltwater) simply be painted with a corrosion resistant paint?

Any such paint would interfere with the heat exchange, which is the biggest area of concern.

Re:interest prospect (0)

Anonymous Coward | more than 4 years ago | (#29497391)

See the first paragraph of this post [slashdot.org] for an explanation.

Re:interest prospect (0)

Anonymous Coward | more than 4 years ago | (#29497455)

Paint would reduce the effectiveness of the heat transfer. It ends being an extra insulator.

Re:interest prospect (0)

Anonymous Coward | more than 4 years ago | (#29497499)

it would reduce efficiency, see above post.

Re:interest prospect (1)

MicktheMech (697533) | more than 4 years ago | (#29497523)

First, the paint will likely reduce the conductivity of the exchanger tubes and reduce the efficiency. Really though, the issue isn't so much finding a corrosion resistant material (there are austenitics or even duplex stainless steels that hold up pretty well), the issue is that these things are running 24/7 for a long time and salt water is going to eventually eat away at whatever you run it through. Normally you can just use pipes thick enough to last a while, but when designing a heat exchanger you want to use tubes as thing as possible to minimize the thermal resistance. I used to work in oil and gas, we were on a lake, so I don't have direct experience with salt water, but with brackish water we were still using tube thicknesses around 1/10th of an inch in our heat exchangers.

Re:interest prospect (0)

Anonymous Coward | more than 4 years ago | (#29497165)

304 or 316 stainless steel is about the worst selection you could make for a seawater environment. Chlorides (stress corrosion cracking) will eat through the 304 or 316 very quickly. You could try using a higher alloy such as 310 or 314, but you're probably better off staying away from austenitic stainless steel.

Re:interest prospect (1)

mano the shark (1641865) | more than 4 years ago | (#29497669)

Using 316 stainless steel would not be a wise decision in a seawater environment. Chlorides lead to stress corrosion cracking (SCC) http://en.wikipedia.org/wiki/Stress_corrosion_cracking [wikipedia.org] in austenitic stainless steel and types 304 and 316 are especially susceptible to SCC. You could try a higher alloy such as 310 or 314, but are better off staying away from austenitic stainless steel.

Re:interest prospect (4, Informative)

david.given (6740) | more than 4 years ago | (#29497775)

A low carbon stainless steel such as the 316 series should be more than sufficient for any piping.

Stainless steel is prone to pitting corrosion when exposed to water containing chlorides. 316 series stainless steel is significantly corroded by concentrations of chlorides above 1000ppm (ref [hghouston.com] ). Standard sea water at 3.5% salinity has a chloride concentration of about 20000ppm (ref [seafriends.org.nz] ).

Stainless steel works rather like aluminium when it comes to preventing corrosion; the surface oxidises very rapidly to form a passive coating, protecting the bulk of the metal from oxygen. In water, this only works if (a) the water contains enough oxygen to passivate the metal, and (b) the water won't then dissolve the coating as soon as it forms. In particular, this means that stainless steel is not suitable for things like marine bolts, because under the bolt head the water will quickly lose all its oxygen and you'll get corrosion. It also means you have to be very careful in sea water as the salts can strip off the chromium oxy passive layer.

316 stainless is considered 'marine grade', but only just. In particular, it's unsuitable for warm sea water, as this makes the water vastly more corrosive. So you probably don't want to use it for coolant pipes.

And I haven't even mentioned electrolytic corrosion yet. Sea water is one of the most corrosive environments on the planet, and dealing with corrosion is one of the biggest problems when working with it.

Re:interest prospect (1)

Locke2005 (849178) | more than 4 years ago | (#29496947)

Not to mention mussels and other sea life growing inside the pipe. It would be interesting to see a long-term economic study of this to see if the energy savings more than make up for the increased maintenance costs. Heating seawater has long-term ecological consequences. Also, wouldn't this necessitate locating data centers on prime ocean-front real estate, which is some of the most expensive real estate in the world? (Although it would be great for attracting IT professionals!) From a Thermodynamics standpoint this is a big win; not such a big win from a total cost of ownership standpoint.

Re:interest prospect (1)

ShadowRangerRIT (1301549) | more than 4 years ago | (#29497049)

People and data centers have a different idea of what the ideal coastline is. Most people want it shallow and sandy: all the better to swim and surf in. A data center wants it deep and rocky, so it can access extremely cold water without picking up too much debris. The lack of overlap would contribute to lower prices for real estate.

Re:interest prospect (2, Interesting)

Gonoff (88518) | more than 4 years ago | (#29497265)

They manage well in the Sydey Opera House. They keep the salt water out of their system and heat exchange to fresh water which they circulate.

To keep the corrosion low, they use sacricicial anodes http://en.wikipedia.org/wiki/Sacrificial_anode [wikipedia.org] . These are also used on ships, oil rigs and pipelines - probably more things too. This is nothing new. I believe the opera house was finished in 1974 so they are using well tested technology here.

Re:interest prospect (1)

sexconker (1179573) | more than 4 years ago | (#29497723)

Mike Rowe had to replace a bunch of those bad boys in a hurricane barrier.

Re:interest prospect (1)

pearl298 (1585049) | more than 4 years ago | (#29497575)

Obviously this is a problem with cooling the engine of an ocean going vessel too.

The traditional solution is to use corrosion resistant metals and to add sacrificial anodes of a metal such as zinc which make electrical contact with the metal parts to be protected.

It is an added expense, but quite manageable.

For many years I lived on a sailboat with a 30 year old diesel engine which worked just fine, but you had to replace the zinc anodes every couple of months.

They cost about US$2 and took about 10 minutes to replace.

Re:interest prospect (1)

Dayze!Confused (717774) | more than 4 years ago | (#29497773)

Most likely it is a closed circuit circulation of fresh water inside the pipes that works much like a car radiator system, by pumping water through their data center and then into the deep sea radiator that they build. The only part that needs to be monitored and aware of corrosion and such is the outside of these pipes, which someone else mentioned could be dealt with by adding a corrosion layer into the pipes and monitoring every once in a while.

Re:interest prospect (1)

Medievalist (16032) | more than 4 years ago | (#29497981)

But what are maintenance costs and lifespan of such a piece of equipment,
I can't image Saltwater not eating the hell out of all the piping.

Google "sacrificial anode".

I'm more worried about heating up the oceans. Heat pollution is already a problem in rivers and streams, and it's not like we aren't already stressing the hell out of the oceans...

Re:interest prospect (0)

Anonymous Coward | more than 4 years ago | (#29497987)

every (sea water) ship, and powerplant near seawater, manages to use seawater for cooling so it is obviously not an unsolvable problem

So could... (0, Offtopic)

Darkness404 (1287218) | more than 4 years ago | (#29496749)

All the talk about using the ocean makes one wonder if Google or another company could build a data center that was self-sufficient in the middle of the ocean, would it be under any jurisdiction when it came to copyright?

Re:So could... (2, Funny)

aicrules (819392) | more than 4 years ago | (#29496807)

It would then be under the jurisdiction of pirates of the real kind.

Re:So could... (3, Interesting)

Ethanol-fueled (1125189) | more than 4 years ago | (#29497003)

Then the RIAA would pay the U.S. Navy to "accidentally" drop anchor [cnn.com] on the pipe before sniping [go.com] the datacenter's operators.

Then the American public would be told some daring tale about how the heroic navy again thwarted those evil "pirates" [wikipedia.org] and they wouldn't know the difference!

Re:So could... (1)

drinkypoo (153816) | more than 4 years ago | (#29497115)

It would then be under the jurisdiction of pirates of the real kind.

Sounds like it's time for google to go all Ng on their ass. I knew they'd listen to Google Reason 2.0b!

Re:So could... (1)

sopssa (1498795) | more than 4 years ago | (#29496819)

That is what the Google talk in the summary was about, they we're going to use ships as datacenters that float in the ocean.

Re:So could... (1)

maharb (1534501) | more than 4 years ago | (#29496823)

If the company has legal entities in a country I am sure they will be subject to those laws.

Re:So could... (2, Insightful)

aicrules (819392) | more than 4 years ago | (#29496873)

That really depends on what treaties the ....zzzz put myself to sleep even trying to explain it.sorry

Re:So could... (2, Insightful)

digitalunity (19107) | more than 4 years ago | (#29497303)

Not really. There is no real escape from national laws with respect to the internet. The reason is simple - everyone is connected! Google has a physical presence in the US, so the company can be penalized in the US for actions of the company abroad.

Hypothetically, say you have a ship in the middle of the ocean. You, nor your company have no physical presence anywhere other than the ship. You still need peering from someone on the internet. Whether that be joe blow or AT&T, you need peering. So you decide to host the pirate bay on your ship. RIAA sues your peering provider to terminate your internet access. Your peering provider is in the US. Your peering provider loses the copyright battle(assuming, in this fairy tale land they actually fought it) and shuts off your internet access. Ok, that's fine and good, you can just move the ship and find another peering provider. Repeat ad nauseum. Eventaully, you run out of people willing to peer with you.

You need to get peering from someone with sufficient political clout that THEIR peering provider isn't willing to cut them off, AND is willing to stand up to international pressure to terminate your access. I think there's a reason a lot of nefarious activity on the internet comes from Russia. Nobody has the clout to take russia offline and russia doesn't mind having all the crooks using their tubes.

Although it uses less electricity, not "green" (5, Informative)

Useful Wheat (1488675) | more than 4 years ago | (#29496757)

Although this solution is certainly "low power" by no means should it be considered to be entirely green. I work as an engineer on many projects that involve sea water, and when you're using it for a cooling source you typically need to inject some sort of chemical to sterilize the water to keep growths off your heat exchangers (barnacles are sort of a pain in the ass in your exchangers). As a result, using sea water for large scale cooling operations is prohibited in large regions of the United States (specifically the gulf coast) mostly over concerns that the large amounts of warm bleached water will damage the ecosystem. Although, that issue aside, using the ocean as a cooling medium is a great idea, and has been used reliably by power plants for many years.

Re:Although it uses less electricity, not "green" (1)

RingDev (879105) | more than 4 years ago | (#29496887)

I'm surprised that a mechanical sleeve would not suffice. A cheap easily replaceable heat conductive barrier between the internal system coolant and the salt water. Sure, it would raise the design and operational costs slightly, but if it allows for a more ecological solution in areas that currently forbid such activity, it might still be well worth it.

-Rick

Re:Although it uses less electricity, not "green" (1)

Nethead (1563) | more than 4 years ago | (#29496949)

I was wondering about that when I didn't RTFA. I live about 100-500 yards from salt water (depending on the tide) and it's a nasty liquid. I've seen the bilge pumps on some of the local fishing boats.. e-gads! Salt water is always a fight that you can never win, just hold off for a while.

Re:Although it uses less electricity, not "green" (2, Insightful)

jollyreaper (513215) | more than 4 years ago | (#29496963)

Although this solution is certainly "low power" by no means should it be considered to be entirely green. I work as an engineer on many projects that involve sea water, and when you're using it for a cooling source you typically need to inject some sort of chemical to sterilize the water to keep growths off your heat exchangers (barnacles are sort of a pain in the ass in your exchangers). As a result, using sea water for large scale cooling operations is prohibited in large regions of the United States (specifically the gulf coast) mostly over concerns that the large amounts of warm bleached water will damage the ecosystem. Although, that issue aside, using the ocean as a cooling medium is a great idea, and has been used reliably by power plants for many years.

So maybe it would be more environmentally sound to run a closed loop out to the current to cool the water and bring it back? Salt water is nasty, evil shit.

Re:Although it uses less electricity, not "green" (3, Funny)

Spy der Mann (805235) | more than 4 years ago | (#29497111)

Salt water is nasty, evil shit.

You should've seen it when it was filled with primordial soup a thousand million years ago, and then came the primitive lifeforms, eeew!

Dealing with growth (4, Funny)

supernes (1560323) | more than 4 years ago | (#29497329)

Can't you just heat it up to sterilize it?

Re:Dealing with growth (1)

Obfuscant (592200) | more than 4 years ago | (#29497561)

Can't you just heat it up to sterilize it?

Yeah, and then you can bring deep, cold seawater up from the depths of the ocean to cool it off, and bingo, cold sterile water!

Re:Dealing with growth (1, Insightful)

TheGreenNuke (1612943) | more than 4 years ago | (#29497647)

You could... and thus defeat the purpose of using it as a coolant.

Re:Although it uses less electricity, not "green" (1)

skine (1524819) | more than 4 years ago | (#29497889)

"Green" does not mean perfect, good, or even revolutionary. It means better; even if the improvement is only slight.

Cars that burn gasoline are being called "green." Coal power plants are talking about being/going "green." etc.

Great Idea (0)

Anonymous Coward | more than 4 years ago | (#29496789)

This would be a good idea for the Great lakes, where the water stays very cold all year 'round.

It could be a good way to attract business to the Midwest.

Cold water cooling (4, Interesting)

diodeus (96408) | more than 4 years ago | (#29496793)

Toronto already uses cold water cooling for air-conditioning many of its office towers in the downtown core and has for many years. (see: http://www.enwave.com/dlwc.php [enwave.com] ). Unless winter never visits Canada again, this is cold body is self-replenishing.

Re:Cold water cooling (1)

pearl298 (1585049) | more than 4 years ago | (#29497631)

Of course Lake Ontario is one of the largest FRESH WATER lakes in the world ...

The only real problem there is zebra mussels which clog pipes and eat many organic pollutants (you can't win them all!)

Otec. (1)

LWATCDR (28044) | more than 4 years ago | (#29496795)

If you could set up an OTEC system as well you could also power the data center as well as cool it.

Re:Otec. (1)

Nefarious Wheel (628136) | more than 4 years ago | (#29496917)

Yes - check out Whispergen (jfgi) in New Zealand. They make a nice quiet home sized cogenerator based on a Stirling engine. Any decent temperature differential will do.

Environmental impact, anyone? (1, Insightful)

Anonymous Coward | more than 4 years ago | (#29496797)

So Cornell transfers heat from its datacenters into Lake Cayuga. This is understandably good for the datacenters, but what's the impact on Lake Cayuga?

The Mauritius system sounds interesting, though, because the heat gain in the water seems less likely to have an impact on ocean temperatures, even on a very localized scale. For minimal environmental impact, use large, deep bodies of water with good currents. Take note, Cornell!

Re:Environmental impact, anyone? (4, Informative)

caerwyn (38056) | more than 4 years ago | (#29496867)

The Cornell project was actually incredibly controversial prior to beginning operation for exactly that reason. Studies since have shown that any detrimental effects are negligible, though, so the controversy has died down in recent years. (I was at Cornell when the system went into operation and for a few years afterward)

Re:Environmental impact, anyone? (3, Informative)

Trepidity (597) | more than 4 years ago | (#29496983)

Not a subjective judgment in either direction, but for what it's worth, this paper abstract [inist.fr] quantifies the heat imparted to Lake Cayuga as "equivalent to an additional two hours of sunlight each year".

A better headline (1)

smartaleq (905491) | more than 4 years ago | (#29497285)

Using Your Data Center to Warm the Sea?

Warm Water Discharge (2, Informative)

guyfawkes-11-5 (1583613) | more than 4 years ago | (#29496859)

Powerplants use this frequently, it's a great idea until the amount of warm water discharged begins affecting the discharge site. I can't imagine a data center requiring the amount of cooling that a powerplant would need.
The EPA required some modifications to a similar system for a powerplant in PR a few weeks ago.
http://www.waterworld.com/index/display/article-display/1830526029/s-articles/s-waterworld/s-industrial-water/s-wastewater/s-2009/s-08/s-epa-requires_new_pipe.html [waterworld.com]

Re:Warm Water Discharge (1)

vlm (69642) | more than 4 years ago | (#29497131)

Powerplants use this frequently, it's a great idea until the amount of warm water discharged begins affecting the discharge site. I can't imagine a data center requiring the amount of cooling that a powerplant would need.

Typical coal plants run around 40% efficient, top of the line natural gas plants run around 60% efficient. Within some rounding errors, the data center will dump about as much heat as its fractional share of the power plant that feeds it.

http://en.wikipedia.org/wiki/Fossil_fuel_power_plant [wikipedia.org]

Re:Warm Water Discharge (1)

gnieboer (1272482) | more than 4 years ago | (#29497277)

Yep, nuclear reactors in particular have been noted for this problem, especially a couple in the Great Lakes area. Can't seem to find a decent-quality link though.

And at the other end of the spectrum, using ground-water for cooling is not a radical idea for housing, that technology is available at most local contractors, is widely used in Florida for cooling, and is spreading across the country. Higher upfront costs, though, than a freon-based system.

So assume your data center is a) bigger than a single family home, and b) smaller than a nuclear reactor, there is probably a solution out there.

But the water quality police will have to argue with the air quality police about which is better for the planet.

Refrigerating web servers (1)

gmuslera (3436) | more than 4 years ago | (#29496863)

With a bit of luck, slashdotting them could get Global Warming as side effect. Looks like a good terrorist/supervillain/evil scientist plot.

Re:Refrigerating web servers (1)

irondonkey (1137243) | more than 4 years ago | (#29496977)

I thought about this too, and I am definitely not an expert, but the datacenter is probably going to be generating the same amount of waste heat. Isn't it six of one, half-dozen of another whether that waste heat goes into the air or the water?

Re:Refrigerating web servers (1)

Spy der Mann (805235) | more than 4 years ago | (#29497127)

It's going to generate waste heat, anyway. The key is that it's not producing CO2, which adds greenhouse effect to the heat produced.

Re:Refrigerating web servers (1)

stevelinton (4044) | more than 4 years ago | (#29497707)

Air is poor at both carrying and conducting heat, compared to water. Losing that much heat to the air would involve blowing a vast amount of air through the building, especially in warmer climates where the ambient air might be at 40C to start with. Water is much more compact and it costs less to pump enough water past the computers to carry the heat. Especially if you have a source of free chilled water.

Old sawmill (1, Informative)

Anonymous Coward | more than 4 years ago | (#29496871)

Think they are already doing the sea water cooling thing at the old sawmill in Hamina that Google bought.
http://www.google.com/intl/en/datacenter/hamina/index.html

global warming anyone? (0)

Anonymous Coward | more than 4 years ago | (#29496885)

seriously this is pretty direct

Re:global warming anyone? (2, Insightful)

ShadowRangerRIT (1301549) | more than 4 years ago | (#29496959)

Generally speaking, humans are not believed to have the capacity to directly affect the global climate on any significant scale. We can only affect it indirectly by altering the amount of heat received from the Sun and the amount emitted back into space. While the heat emitted by these projects may have localized effects, it's highly unlikely to produce any global climate effects. I suppose there is a tiny chance of disrupting ocean currents, but that's indirect, and only redistributes heat, it doesn't affect the global average. Global warming also redistributes heat, but the global average also climbs, it's not zero sum (within the Earth's atmosphere; everything is zero sum eventually).

Re:global warming anyone? (0)

Anonymous Coward | more than 4 years ago | (#29497617)

if you're taking anti-matter into account even god can be zero-summed. wow, divine kryptonite :O

Computer cause global warming! (2, Funny)

ipoverscsi (523760) | more than 4 years ago | (#29496925)

I'm not too sure about the anthropogenic global warming, but I'm starting to come around to it. Earlier my contention was that global warming scientists are causing global warming [dailymail.co.uk] , but I'm beginning to think that maybe -- just maybe -- computers in general might be the cause. I mean, if computers are having to pump cold water from the ocean depths to cool computers, that's gotta be dumping a lot of heat back into the ocean, right? Right...?

Re:Computer cause global warming! (1)

lenehey (920580) | more than 4 years ago | (#29497513)

How is it the parent is modded insightful? Funny, maybe. Your finger slip?

Re:Computer cause global warming! (1)

bill_mcgonigle (4333) | more than 4 years ago | (#29497763)

that's gotta be dumping a lot of heat back into the ocean, right

yeah, but since nobody drives to a library anymore, it's a wash.

Most of downtown Toronto is cooled by lakewater (3, Informative)

insanewombat (656212) | more than 4 years ago | (#29496931)

Most of downtown Toronto is cooled by lakewater - enwave energy [enwave.com] provides district cooling for most of the major buildings in the downtown core. This includes 151 Front St. [151frontstreet.com] , one of the major datacentres in the area. See here [151frontstreet.com]

Other cooling techniques (0)

Wowsers (1151731) | more than 4 years ago | (#29496987)

I might just be a bit far out with this (non sea water) solution, but why don't they just install a heat exchanger to generate their own electricity? It generates electricity whilst not costing the earth in installation / cooling costs.

Re:Other cooling techniques (1)

jabithew (1340853) | more than 4 years ago | (#29497377)

Second Law of Thermodynamics; the heat is probably of too low-grade to be used to generate useful electricity. The temperature gradient between ambient and the temperature the heat is being released at likely isn't big enough either. The ratio between Thot and Tcold limits the efficiency; the higher the ratio, the higher the efficiency.

See this Wikipedia [wikipedia.org] for a reasonable discussion of this concept.

Already old news (1)

thewils (463314) | more than 4 years ago | (#29497087)

Didn't they try this in New Orleans a couple of years back?

Wait, back that up, reverse it. (1)

blhack (921171) | more than 4 years ago | (#29497093)

Am I the only one that reads things like this as:

"Check out this great new way to heat our oceans using our datacenters!".

You guys realize that the energy doesn't just disappear, right?

Re:Wait, back that up, reverse it. (2, Insightful)

RobVB (1566105) | more than 4 years ago | (#29497321)

The heat certainly doesn't disappear, but you're just pumping heat into cold water with this system. The transfer of heat from a warm to a cold substance is a process which increases entropy, which means it's a spontaneous process (it doesn't take any energy to do it).

Air-conditioning, on the other hand, transfers heat from a cold to a warm substance (the cooled air inside becomes cooler, the warm air outside becomes warmer), which is not a spontaneous process, meaning you're using extra energy. This extra energy is ultimately wasted as extra heat in the warm substance.

Also, the extra energy which A/C uses generally comes from burning fossil fuels at relatively low efficiencies, emitting even more heat into the atmosphere.

All in all, you're putting a LOT less energy (heat) into the earth if you're using a spontaneous process to cool your stuff.

Re:Wait, back that up, reverse it. (1)

nsayer (86181) | more than 4 years ago | (#29497345)

"Check out this great new way to heat our oceans using our datacenters!"

You say that like it's a bad thing.

We're going to have datacenters (the fact that you're posting to /. makes you a hypocrite otherwise). So which is worse? Cooling them with electrically powered air-conditioning? Or using something else as a heat sink?

Re:Wait, back that up, reverse it. (1)

glitch23 (557124) | more than 4 years ago | (#29498307)

The heat goes into the ocean, which is already cool or cold depending on depth. The warmer water will warm up the cooler water so the temperature distribution averages out. The heat will dissipate from the water eventually from the surface of the body of water and enter the atmosphere just as if it was placed there originally by using A/C. At least, that's what my brain tells me but I've never had a class in fluid dynamics or thermodynamics.

Cooling from the sea! (2, Interesting)

spun (1352) | more than 4 years ago | (#29497095)

"Welcome, humans! I am ready for you! Fish, plankton, sea greens and cooling from the sea. Fresh as harvest day. Overwhelming, am I not? Are you, too, startled? Am I too removed from your kin?"

Re:Cooling from the sea! (1)

camperdave (969942) | more than 4 years ago | (#29497745)

Ah! Someone besides me watched Logan's Run last weekend.

Environmental Concerns (2, Insightful)

Laptopdude (1240858) | more than 4 years ago | (#29497101)

Raising the temperature of a body of water by even a few degrees can have disastrous consequences; from outright killing species, to producing algal blooms that deplete oxygen levels (and then kill species). I mean, think about it. Water resists temperature change much more than air, so a sudden increase is bad news to creatures that just aren't made to deal with it. Also, a recent study has found that increased carbon dioxide levels are making marine life more susceptible to fluctuating temperature and oxygen levels.

But, of course, just one place in the ocean using this method isn't going to have that much effect. It's if and when this cooling strategy starts to catch on that we have to worry about affecting our environment, and weigh the consequences of air conditioning (fossil fuel emissions) versus heat pollution.

Re:Environmental Concerns (5, Informative)

RobVB (1566105) | more than 4 years ago | (#29497453)

The total mass of the oceans [hypertextbook.com] is about 1.4*10^21 kg. The total mass of the atmosphere [hypertextbook.com] is about 5*10^18 kg. That means the oceans weigh about 300 times as much as the atmosphere.

The heat capacity of water [npl.co.uk] is about 4000 J * kg ^ -1 * K ^ -1. The heat capacity of air [engineeringtoolbox.com] is about 1 kJ * kg ^ -1 * K ^ -1, or about 1000 J * kg ^ -1 * K ^ -1.

So since there's 300 times as much water as there is air, and the heat capacity of water is 4 times larger, heating up the atmosphere by 1200 degree Celsius would take the same amount of energy as heating up the oceans by 1 degree Celsius. This may not prove or disprove your point, I just started thinking about numbers when you said "raising the temperature of a body of water by a few degrees".

Re:Environmental Concerns (2, Funny)

mujadaddy (1238164) | more than 4 years ago | (#29497753)

heating up the atmosphere by 1200 degree Celsius would take the same amount of energy as heating up the oceans by 1 degree Celsius.

Wow. Let's do that!

Re:Environmental Concerns (0)

Anonymous Coward | more than 4 years ago | (#29498393)

The total mass of the oceans [hypertextbook.com] is about 1.4*10^21 kg. The total mass of the atmosphere [hypertextbook.com] is about 5*10^18 kg. That means the oceans weigh about 300 times as much as the atmosphere.

The heat capacity of water [npl.co.uk] is about 4000 J * kg ^ -1 * K ^ -1. The heat capacity of air [engineeringtoolbox.com] is about 1 kJ * kg ^ -1 * K ^ -1, or about 1000 J * kg ^ -1 * K ^ -1.

So since there's 300 times as much water as there is air, and the heat capacity of water is 4 times larger, heating up the atmosphere by 1200 degree Celsius would take the same amount of energy as heating up the oceans by 1 degree Celsius. This may not prove or disprove your point, I just started thinking about numbers when you said "raising the temperature of a body of water by a few degrees".

Except we don't know exactly what raising the temperature of the oceans by one degree will do.

Re:Environmental Concerns (2, Insightful)

Khashishi (775369) | more than 4 years ago | (#29497671)

I knew the environment would come up. Of course, pretty much anything humans do is going to be deleterious to the environment. But put things in perspective. It's more thermodynamically efficient to transfer heat to the ocean directly, rather than burn fuel to create electricity to power a heat pump which is used to transfer heat into the air. The power plant also needs to be cooled, either by evaporating large amounts of water in cooling towers, or by transferring heat to an ocean or lake. Which do you think is better for the environment?

Re:Environmental Concerns (1)

RobVB (1566105) | more than 4 years ago | (#29498087)

Which do you think is better for the environment?

Secret option number three: join the Amish.

Gulf Coast rejected something like this (1)

madcat2c (1292296) | more than 4 years ago | (#29497197)

There was a story of a LNG (Liquid Natural Gas)facility that lost approval for using this method to do cooling for a facility (or something in the LNG process) with Gulf coast seawater. The problem is that all the baby fish and other sea life die in the system. They were worried about the local shrimp and fishing industry getting wiped out. They made them use a closed loop system to solve this problem.

Re:Gulf Coast rejected something like this (1)

aXis100 (690904) | more than 4 years ago | (#29497721)

Umm..... I get your point, but the refrigeration on LNG plant would be about 4 orders of magnitude more powerfull than a data center.

Optical Computing (1)

transami (202700) | more than 4 years ago | (#29497267)

They should be more involved in getting optical computing on the table. That more than any other tech will have a profound effect on energy requirements. And given some of the latest R&D, the tech is getting very close to reality.

Environmental concerns (1)

nsayer (86181) | more than 4 years ago | (#29497279)

I thought the greeniacs were all up ons about nuclear plants using seawater for cooling because the heated exhaust invariably caused altered conditions at the point of discharge. And far be it from us horrible ebil humans to actually change the environment. That's just wrong.

So what makes this different?

Always wondered about ARSC's bill (1)

Pvt_Waldo (459439) | more than 4 years ago | (#29497361)

During the winter, does the Arctic Regional Supercomputing Center [arsc.edu] spend money (energy) heating their offices while cooling off their computers at the same time?

glaciers (0)

Anonymous Coward | more than 4 years ago | (#29497397)

hurray! bringing the warmth to the poles with even less inbetween steps! this efficiently gives us more melting water, which in turn can cool our datacenters even more. So at one point in the future you're google query is being cooled by water which once was ancient glacier ice. (that's alot of 'c-'s , which are those things which we were using in the first place to cool everything)

Pollution (1)

cybereal (621599) | more than 4 years ago | (#29497607)

For decades we've recognized this exact same kind of exploitation of coastal waters as pollution. Why would this suddenly be acceptable for a data-center, and how will they avoid the associate ecological devastation?

Solving the wrong problem (1)

CityZen (464761) | more than 4 years ago | (#29498061)

Instead of figuring out better ways to cool hot datacenters, I think a better investment would be to figure out how to make the data centers run cooler in the first place.

Each new generation of semiconductor technology typically allows 2 ways to improve: faster computing for same power, or less power for same computing.
You'd think that at some point, the second choice makes more sense than the first.

Of course, there is the problem of who can make these improvements. This is limited to the chip companies, whereas just about anyone who knows how to use a screwdriver can work on the cooling problem.

I solved my own version of the problem by switching to notebook computers instead of desktops. I get disgusted every time I look at the power requirements of modern desktops: they burn over 100 watts when sitting idle. That's just stupid. Most notebooks don't use that much even running flat-out.

I wonder how many data center boxes are burning 100 watts just sitting idle?

Load More Comments
Slashdot Login

Need an Account?

Forgot your password?

Submission Text Formatting Tips

We support a small subset of HTML, namely these tags:

  • b
  • i
  • p
  • br
  • a
  • ol
  • ul
  • li
  • dl
  • dt
  • dd
  • em
  • strong
  • tt
  • blockquote
  • div
  • quote
  • ecode

"ecode" can be used for code snippets, for example:

<ecode>    while(1) { do_something(); } </ecode>