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!

Why Waste Servers' Heat?

timothy posted more than 3 years ago | from the because-heat-makes-us-murderous dept.

Data Storage 204

mikejuk writes "A new paper from Microsoft Research (PDF) suggests a radical but slightly mad scheme for dealing with some of the more basic problems of the data center. Rather than build server farms that produce a lot of waste heat, why not have distributed Data Furnaces, that heat home and offices at the same time as providing cloud computing? This is a serious suggestion and they provide facts and figures to make it all seem viable. So when it gets cold all you have to do is turn up the number crunching ..."

cancel ×

204 comments

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

Not new. (4, Interesting)

forgotten_my_nick (802929) | more than 3 years ago | (#36856162)

This isn't a new idea. Some buildings like this already and IIRC IBM also marked this as one of their next 5 in 5.

Re:Not new. (1)

Daniel_Staal (609844) | more than 3 years ago | (#36856214)

And the problem with it in most cases is that servers shrink over time. Every new generation is both smaller and more power-effiecent, or they just plain get moved.

Which means that in 15~20 years you are barely supplying enough heat to overcome heat losses in the system. And the homes and offices have no heating.

Re:Not new. (1)

AngryDeuce (2205124) | more than 3 years ago | (#36856290)

And the homes and offices have no heating.

In my office, that would be an improvement.

In my personal experience, one of the biggest downsides to working in an office of predominately women (like I am now) is that the vast majority are freezing once the temperature drops below 75F. There are people running space heaters under their desks currently because we recently got a male office manager who, in his first managerial decision, turned the air down so that all the men weren't sweating down their backs and through their shirts all day long.

Of course, in the winter, the office temp. has always been pegged at about 87F. Plants wilt, you can see heat haze when looking from one end of the office to the other...

Re:Not new. (3, Funny)

Midnight Thunder (17205) | more than 3 years ago | (#36856430)

Some of them are cold inspite of the actual temperature ;)

Re:Not new. (0)

roman_mir (125474) | more than 3 years ago | (#36856656)

office of predominately women (like I am now)

- whaaa? Are you saying you are a predominantly woman?

the vast majority are freezing once the temperature drops below 75F

- it's clear what to do now - women should be used as cooling exchangers.

Re:Not new. (0)

Anonymous Coward | more than 3 years ago | (#36856844)

Trans girl here; estrogen affects metabolism.
I never used to feel the cold but now I do. It's an objective thing, not just women being wimps.

To reach a compromise, I still say wear more layers rather than wasting all that energy.

Re:Not new. (1)

AngryDeuce (2205124) | more than 3 years ago | (#36857150)

I'm all for layering, but I have a feeling that I would get sent home if I removed enough layers to be comfortable in our office in the summer time.

Funny though, suggest that they put on a sweater in the winter to keep warm, as opposed to turning the heat up so high paper starts to char, and you'd think you'd asked them for sexual favors or something. You know it's bad when fighting over the thermostat settings spills over into email and ends up going like 4 steps up the ladder before a random VP tells everyone to stop wasting their fucking time...

Re:Not new. (1)

The Archon V2.0 (782634) | more than 3 years ago | (#36857498)

To reach a compromise, I still say wear more layers rather than wasting all that energy.

Precisely. I'm a guy who tolerates heat well but cold poorly. I've got my jacket on right now because of the air conditioning where I work. I don't bother other people because of it, I just put the damn jacket on. In winter I add a sweater and cardigan under the jacket, still nice and toasty.

Oh, that's excellent timing. My cubicle neighbor just started complaining about the temperature. I think I'll add earmuffs.

Re:Not new. (0)

Anonymous Coward | more than 3 years ago | (#36856662)

Mainly older women I would say. In my office we have a group of women in their 50s/60s who bitch constantly if anyone dares turn the a/c on but all the other women are fine. Segregation of the workforce would help (ie. Stick the whining bastards in an office on their own and they can melt themselves for all I care)

Re:Not new. (1)

Truekaiser (724672) | more than 3 years ago | (#36856758)

What also works, and studies back it up is just putting in a dummy thermostat. people who complain it's too(hot/cold) suddenly stop afterwards.

Ah, yes. (1)

Oxford_Comma_Lover (1679530) | more than 3 years ago | (#36857600)

What also works, and studies back it up is just putting in a dummy thermostat. people who complain it's too(hot/cold) suddenly stop afterwards.

Efficiency through lying. How normatively cute. :)

(The other common placebo example is the close door button in elevators.)

Re:Not new. (1)

perlchild (582235) | more than 3 years ago | (#36856292)

They get more efficient... for the same amount of computing power. However, we ain't stopped needing more/faster computing power recently, quite the opposite.
I would have expected Microsoft to make a proposal about power plants with computing power, not mere "smarter" buildings though.

Re:Not new. (1)

Anonymous Coward | more than 3 years ago | (#36856500)

This is exactly what happened at my last job. The building was put up with no heating system at all - they counted on the data centers producing enough heat to heat the 4-story building. We all started chuckling when the directors in their corner offices had to bring in space heaters because they ended up freezing with all their windows and not enough heat coming from the data center.

This was 15 years ago so although Microsoft may think it's radical that tells you how far behind the technology they really are. And my old job was perhaps 100 miles from Redmond as the crow flies.

Re:Not new. (1, Funny)

ozmanjusri (601766) | more than 3 years ago | (#36856726)

servers shrink over time. Every new generation is both smaller and more power-effiecent,

This is Microsoft research remember.

Their bloated OSs have kept chip designers busy building faster, more complex CPUs for decades.

Re:Not new. (2)

fatphil (181876) | more than 3 years ago | (#36856884)

But 20 years ago, we just alt-tabbed between windows, and they just drew themselves as quickly as possible. Nowadays, we (not me, it's a complete abomination, IMHO) want high resolution alphablended 3D wibbly-wobbly animations in order to switch between programs. Pulling a figure out of my arse, that must be about 100x as much work. (The folk interpretation of Moore's Law supports a 57x increase in that period.)

Likewise, some browsers are now doing web searches in the background with every character you type in the search box - that's way more than 100x work than just rendering 1 more character and waiting for you to click 'search'.

You have it backwards (1)

goombah99 (560566) | more than 3 years ago | (#36857080)

And the problem with it in most cases is that servers shrink over time. Every new generation is both smaller and more power-effiecent, or they just plain get moved.

Which means that in 15~20 years you are barely supplying enough heat to overcome heat losses in the system. And the homes and offices have no heating.

Sigh... Let's think about this shall we. The reason we currently discard our serveres every 3 years is that the operating costs per compute cycle savings exceed the capital costs of new servers. But these servers have negative operating costs. They will never go obsolete in terms of operating costs per computer cycle. They will only go obsolete when the wall clock time for a calculation becomes undesirably long. For certain kinds of servers (such as ones that are bandwidth starved) the machine will fail from old age first. This very long life cycle means it will be worth the extra cost of putting in very high reliability components.

The reason the paper is important is that it works through the less obvious but more important details. For example, there will be pluses and minuses to distribution. Residential electricity costs more than COmmerical electricity. Bandwidths will need to be upgraded. One of the big costs is the data center itself and that includes air circulation, facility maintanence, floorspace for egress not just cooling power. Don't forget about bathrooms, lights, floor cleaning and 403Bs for all the people who maintain the facility.

Mod up parent (0)

Anonymous Coward | more than 3 years ago | (#36857134)

parent is only thread praising this.

Re:You have it backwards (1)

FooAtWFU (699187) | more than 3 years ago | (#36857340)

Two points. First, it's not certain that the machines will have negative operating costs - reduced operating costs, sure, but using electricity for heating has always been a silly-expensive way to do it (exclusive of certain areas with ridiculously good hydro power). Secondly - 403(b)s? I'm pretty sure those are for educational institutions, non-profits, and the like... not businesses and utilities.

Re:Not new. (1)

watanabe (27967) | more than 3 years ago | (#36857156)

Nope.

Energy use per cm^3 has risen dramatically over the last 20 years. By your stated measures it should be dropping. Datacenters no longer have space budgets, they have power budgets, waste heat is one THE big problems with computers and datacenters right now.

Re:Not new. (1)

Anonymous Coward | more than 3 years ago | (#36856274)

Indeed. My college uses this in their "Green" building. They use the server heat to warm the side of the building that isn't in the sun for balance, and both sides have metal structures outside the windows to act as shade from the sun to try to keep it cooler as we do get a lot of sun around here.

That is only what I know, I wouldn't be surprised if they also used the heat generated to heat up water in the building as well or some other use.

Re:Not new. (2)

Midnight Thunder (17205) | more than 3 years ago | (#36856422)

It is not new, but maybe getting a company like Microsoft talking about it will mean people will actually take notice?

This is also probably why in colder climates the server farms should be downtown, where the excess heat can be taken advantage with the least loss, due to distance.

ETH in Zürich Switzerland (1)

toby (759) | more than 3 years ago | (#36856750)

...has done this for decades.

Re:ETH in Zürich Switzerland (1)

pluther (647209) | more than 3 years ago | (#36857202)

So has Tektronix in Beaverton, Oregon.

What a novel idea (4, Interesting)

suso (153703) | more than 3 years ago | (#36856164)

Nobody's ever thought of that before. I thought this "paper" was going to have some kind of design for a way to do it or something. Actually, recently I've been thinking about the way some barns are constructed. Where they have have windows at the apex of the roof. I guess that channels the heat up and lets it out right? Is it possible to put turbines up there that are driven by heat?

Re:What a novel idea (0)

Anonymous Coward | more than 3 years ago | (#36856184)

Yes because your two sentence is clearly just as substantial and practical to implement, not to mention as beneficiali to mankind as a complicated research paper with all facts and figures by a respected institution. Drongo, did you even read the PDF? Put a fork in suso and turn him over, he's done.

Re:What a novel idea (0)

Anonymous Coward | more than 3 years ago | (#36856244)

Put a fork in suso and turn him over

Careful, to him that's foreplay.

Re:What a novel idea (4, Informative)

suso (153703) | more than 3 years ago | (#36856258)

Oh really, then why when I search for datacenter in a barn, this comes up first: Microsoft puts data centre in a barn (Jan 2011) [itnews.com.au]

Yeah, I'm done alright.

Re:What a novel idea (1)

errhuman (2226852) | more than 3 years ago | (#36856704)

I had to check that one wasn't an April Fools first.

City boy knows nothing about barns. (1)

Anonymous Coward | more than 3 years ago | (#36856452)

Actually, recently I've been thinking about the way some barns are constructed. Where they have have windows at the apex of the roof. I guess that channels the heat up and lets it out right?

No, that's not their purpose at all. Those aren't "windows". They are hatches. They aren't used for ventilation purposes, and they aren't used to let light in.

Most barns have at least two levels. Some that are built into the side of a hill can have three or more levels. Above those levels there is what's called the "loft". The hatches you're referring to lead to the loft.

The loft is usually separated from the lower floors by a wooden (or metal or concrete, in newer barns) floor, with one or two stairways leading down to the lower levels. There are very few pathways for light or air to move from the loft to the lower levels.

The hatch you're referring to is used to more easily move items, goods and material up to the loft, which can otherwise be difficult to access. Outside of the hatch there is usually a rope-and-pulley apparatus that can be used to lift or lower items. The items can be placed into or removed from the loft using the hatch.

Re:City boy knows nothing about barns. (1)

Anonymous Coward | more than 3 years ago | (#36856556)

No, that's not their purpose at all. Those aren't "windows". They are hatches. They aren't used for ventilation purposes, and they aren't used to let light in.

I think he's referring to a cupola which is a small louvered structure on top of a barn often used to vent excess heat.

Temperature? (2)

Kraftwerk (629978) | more than 3 years ago | (#36856174)

640K(elvin) ought to be enough for anybody.

This is novel? (2)

AngryDeuce (2205124) | more than 3 years ago | (#36856194)

My PC has been doing double duty as a space heater for years.

Re:This is novel? (0)

Anonymous Coward | more than 3 years ago | (#36856902)

It's the only thing keeping me on Windows.

Re:This is novel? (1)

hedwards (940851) | more than 3 years ago | (#36857046)

During the winter I don't bother heating most of the house on average days because my computer does a good job. It gets toasty during the summer, but not until the afternoon and it just gives me an excuse to go outside and risk the evil day star's menacing photons.

But, I think the real problem with doing this on a scale substantial enough to make a difference is really that you have them on all the time and you don't want to have to go running around to a million different server rooms monitoring that they're still running properly. Making sure that nobody has gained unauthorized access and getting to the proper room when one does break.

Most large buildings don't use central heating for a reason, they've pretty much all got heat exchangers, hydronic pumps and all that jazz because it's a lot easier to manage the heating and cooling that way than it is to deal with this sort of BS heating system.

Issues (1)

JamesP (688957) | more than 3 years ago | (#36856198)

The main issues are efficiency and temperature.

Sure, when you have something "for free" efficiency is moot. But you would still have to have pumps to transport the heat.

Hence the 2nd point: temperature. I'm thinking you can have the water around 50C/120F tops by that method. So if you get the water at that temp, pump it out to the offices, how much of your heating needs can be fulfilled there? How much heat will be lost in transport?

Re:Issues (0)

Anonymous Coward | more than 3 years ago | (#36856228)

50C running water for the sinks in the toilet should be OK. Or is it 60C that they like to limit it to?
That can still be "topped up" if needed.

Re:Issues (1)

cynyr (703126) | more than 3 years ago | (#36856394)

Simply using them as "pre-heat" will usually greatly limit the amount of heat needed to be added by other "non-free" means and in some cases can greatly increase the efficiency of the other heater. Can i hook my desktop and server up to a pre-heat loop in my house? at least then the heat would go somewhere useful, granted i think all the extra pumps would outweigh any energy savings. In a data center environment, it should be rather easy to use one large pump. Most data centers are a fairly constant load once they are up and running.

I would love to have my heatpump be able to swap outside coil for water heater pre-heat coil when in cooling mode. Granted that would mean an extra water cooled condensing coil, and a receiver to handle the varied refrigerant volume, but both of those issues are overcome-able. This would mean that my heatpump all of a sudden gets very very efficient in the summer. Granted in the fall/spring when in heating mode it doesn't help out the water heater, by a good high turndown condensing water heater should take care of that or maybe even a 2 stage tankless or tankless and point of use small tank one say 3 gallons(would have to be sized based on the incoming line volume from the whole home tankless.

Re:Issues (1)

FishTankX (1539069) | more than 3 years ago | (#36857062)

It might be possible to use a heatpump sourcing heat from the 50c air. That would make it fantastically efficient.

Been done for years. (0)

Anonymous Coward | more than 3 years ago | (#36856206)

This has been done, at least to some extent, for many years.
There is a local school where I live which is heated by waste from an Ericsson server park.
One massive datacenter I was at initially used the servers to heat the swimming pool.
Finally as the server park grew the water reached above 40 degrees c and people were starting to complain.
Then they concluded they could heat surrounding areas with this heat as well.

Stating the obvious (1)

gilesjuk (604902) | more than 3 years ago | (#36856236)

Hardly radical. Power stations have done it for years, some other food processing factories have used the heat to warm up greenhouses to grow tomatoes.

A radical idea would be putting data centers in a cooler climate so they can be cooled more with ambient temperatures.

Re:Stating the obvious (2)

bz386 (1424109) | more than 3 years ago | (#36856264)

Hardly radical. Power stations have done it for years, some other food processing factories have used the heat to warm up greenhouses to grow tomatoes.

A radical idea would be putting data centers in a cooler climate so they can be cooled more with ambient temperatures.

For example like Google's Hamina data center? http://www.google.com/datacenter/hamina/ [google.com]

Re:Stating the obvious (1)

cynyr (703126) | more than 3 years ago | (#36856432)

most new data centers are trying to do everything they can to use "free" cooling methods. indirect evaporative cooling, air side economizer, pre-warming the incomming city water for the heaters. litterly everything, they then also simulate the cooling systems based on outdoor temps for every hour of every day in an ASHRAE [ashrae.org] standard year for the location of the data center.

Re:Stating the obvious (1)

Lennie (16154) | more than 3 years ago | (#36856554)

In the Netherlands there is already a project with datacenters and greenhouses.

Re:Stating the obvious (1)

Lennie (16154) | more than 3 years ago | (#36856580)

Building datacenters in different climates doesn't really help.

The reason companies like Google have as many datacenters isn't just about redudancy. But the biggest reason is the latency between the user and the server. Datacenters need to be close to the user to get the data to the user quickly.

That is why many companies use CDN's and one of the reasons why Google started the SPDY project (because of TCP-slowstart, SPDY tries to make HTTP faster by re-using TCP-connections).

It's already being done. (1)

jgreco (1542031) | more than 3 years ago | (#36856238)

We used to have a minimal heating bill in the winter back when we kept a few racks of servers on-site. Our gas bill has gone up substantially as we've moved to virtualization.

Combined heat and power (3, Informative)

pfafrich (647460) | more than 3 years ago | (#36856242)

Combined heat and power [wikipedia.org] (CHP) schemes are a increasingly common using the waste heat from some process to provide district heating. Temperatures from a server farms might be a bit on the low side but it changes the situation when you look at the heat as a resource to be used rather than a waste item.

Re:Combined heat and power (1)

Midnight Thunder (17205) | more than 3 years ago | (#36856472)

As good as all this is, I would like to see smarter server farms. For a load balanced systems there would be hot, warm and hibernate. All this would be linked to a master controller that would rev up and down the machines as needed. Also, each server would be of the form of mini-ITX, to pack more in and designed to be reach a certain maximum threshold before bringing on supporting systems.

I am also wondering whether in a hosted environment whether there would be a way to give each customer a virtual machine that could be moved around hardware as load requirements dictates?

Re:Combined heat and power (1)

h4rr4r (612664) | more than 3 years ago | (#36857424)

Vmware does this already with power management, which is the same as heat in the end. It is something citrix xen also does and should not be very hard to get kvm to do.

Patent Pending (1)

aixylinux (1287566) | more than 3 years ago | (#36856250)

I smell a patent application coming on....

Multi-functional Space Heaters (1)

Brewmeister_Z (1246424) | more than 3 years ago | (#36856262)

That is what I call my computers sometimes. My office is always the most comfortable room during the cold seasons since that waste heat keep it warmer (and the fact that it is in the upper level of the home). However, in the hot season (now), it is unbearable to work on much in my office. I had 5 computer running yesterday (4 were client computer getting serviced) and it was 85 F in my office with fans going and the AC nearly runs 24/7 lately due to the heat wave. This is a 10 F difference to the middle level where the thermostat is located. This house will built in 1956 and is a partial split level. The office and bedrooms are above a garage and utility rooms on a ground level slab and the mid level is over a crawl space. Poor insulation (by today's standards) in the walls is another factor as well.

Re:Multi-functional Space Heaters (1)

vintagepc (1388833) | more than 3 years ago | (#36856442)

Heh... A family member of mine had that too -> Had a workstation and a 4-PC cluster for modelling computations in his office. Often had to open the window during the winter, and joked that the office should reimburse him for the heating costs they saved.

Not NOW. (1)

retroworks (652802) | more than 3 years ago | (#36856268)

Damn it, NONE of us in the USA are in any mood to talk about heating our damn houses or buildings. It's 45 degrees centigrade. Can't we just save this discussion for a couple of months. It's not like a new idea or anything.

Re:Not NOW. (1)

cynyr (703126) | more than 3 years ago | (#36856444)

HERE HERE! and that is in the great white north that is Minnesota!

Re:Not NOW. (2)

ColdWetDog (752185) | more than 3 years ago | (#36856684)

You guys have to move to Alaska. It's a nice, comfortable 55 degrees F. And my rendering cluster (a pair of old dual xeons) is making the basement nice and comfy. The Lab is currently sleeping under the rack that the computers are on because the heat is deflected downward.

Re:Not NOW. (1)

couchslug (175151) | more than 3 years ago | (#36856892)

"It's 45 degrees centigrade."

Visiting foreigner?

Re:Not NOW. (0)

Anonymous Coward | more than 3 years ago | (#36856998)

The western desert lives and breathes
In forty-five degrees
How can we dance while the earth is turning?
How can we compute when our servers are burning?

Re:Not NOW. (1)

hedwards (940851) | more than 3 years ago | (#36857064)

Liar, no American would ever refer to it as 45 degrees centigrade you impostor. For one thing it's 45 degrees Celsius and for another thing nobody in the US knows that that number means.

Problematic assumption (1)

xkuehn (2202854) | more than 3 years ago | (#36856270)

The least expensive DFs will use the existing home broadband connection

We have caps around here. And so will the rest of you, when the telecoms companies get their way.

Re:Problematic assumption (1)

tgd (2822) | more than 3 years ago | (#36856332)

How much data do you really think something like that will move around?

Netflix and bittorrent, and an American subculture of people who seem to need 10 hours of video entertainment a day are why caps are a problem.

Business servers and people who go outside tend to not run into caps.

Been there, done that... (3, Informative)

julf (323835) | more than 3 years ago | (#36856356)

http://www.guardian.co.uk/environment/2010/jul/20/helsinki-data-centre-heat-homes

Is this design available for home dryers? (0)

Anonymous Coward | more than 3 years ago | (#36856364)

It would be nice to direct the heat (minus the lint) from my dryer to heat my home during the winter.

Re:Is this design available for home dryers? (1)

vintagepc (1388833) | more than 3 years ago | (#36856462)

The problem you tend to get is that hot air is _very_ humid. If you've ever watched a home reno show where the original builders f*cked up the dryer vent install, you'll know how bad the mold can get from that...You'd need some way to dehumidify the air first... and that cost would probably put you at par with running a space heater or so.

Re:Is this design available for home dryers? (1)

russotto (537200) | more than 3 years ago | (#36856772)

You can't use dryer air directly both because of lint and humidity, and exhaust products in the case of a gas dryer. You'd have to use a heat exchanger. This would probably be OK for the inside air, but you'd have to do something about the condensate from the heat exchanger (air from a dryer is typically saturated, so cooling it will cause condensation), and also somehow keep the lint from clogging the exchanger.

bitcoin mining (0)

Anonymous Coward | more than 3 years ago | (#36856370)

imagine a home where all heating is done by tiny computational devices that mine bitcoins http://www.bitcoin.org -> getting a reward for heating and cooking is something i like to have ;-)

Why limit the conversation? (5, Interesting)

bradgoodman (964302) | more than 3 years ago | (#36856386)

Why limit the conversation to just servers, when this occurs everywhere in common life?

Why does my refrigerator take heat out of the inside, and dump it into my house - requiring my A/C to then take it and again put it outside?

Why does my A/C in a house take all the heat and discharge it outside into the atmosphere, which meanwhile a pool heater is running 5 feet away using energy to generate more heat for the pool?

Why do people call incandescent light bulbs "energy wasters", when then can (in the cooler months) defray the work needed to be done by a household heating unit?

Why does the Pizza place down the street run their heater in the winter yet has these giant metal exhaust ducts running from their pizza ovens, venting heat to the outside world? (Why no fins/blowers on these ducts to disperse heat into the pizza-joint?)

The point is - people think of heating and cooling on a "unit" basis - and not on a systemic basis of an overall building - or even area. HVAC systems in buildings get this - sort of - they are not single machines - but a system of different, interconnected machines which are each interconnected, performing different tasks - sort of like organs in a human body. This approach needs to be thought of everywhere where cooling is required, and/or heat is generated.

Re:Why limit the conversation? (2)

kanweg (771128) | more than 3 years ago | (#36856596)

And why does the A/C try to release heat into the atmosphere at a moment when it is already hot (instead of dumping it into a cold buffer it prepared during the night), i.e. at a time when it is the hardest to get rid of that heat?

And why does the A/C try to release heat at a time when electricity demand is already at peak level (instead of during the night)? It would save money building power plants (and lower the electricity bills) if they didn't.

For a fraction of the defense budget, Americans could have saved more energy than their wars in the Middle East secured. And it wouldn't have pissed that many people off also. However, American culture halts progress, I guess. Isn't it amazing. When it comes to going to war, individual Americans don't pick up their gun and defend some sandy patch themselves for a while. There the DO realize that concerted government (military) action is more effective for the greater good (which in this case isn't very good. The US would be quite surprised if Arabs sent drones to Texas to secure their oil interests and kill some civilians as collateral damage. If something is the right thing, it shouldn't matter who does it.). But when it comes to energy, somehow unconcerted action based on individual sellf-interests would be the way to go and government should be left completely out of it.

Bert
With the US defense budget of 10 years, a significant portion of the US energy production could have been sustainable, even if it were spent on the most expensive one of them all: solar PV. Would have made the US probably the world's most efficient producer of solar panels. Now the export product is high velocity lead and copper.

Re:Why limit the conversation? (1)

russotto (537200) | more than 3 years ago | (#36856708)

And why does the A/C try to release heat into the atmosphere at a moment when it is already hot (instead of dumping it into a cold buffer it prepared during the night), i.e. at a time when it is the hardest to get rid of that heat?

A "cold buffer"? You know of a practical way to produce this "cold buffer"? Freeze a ton of water, then efficiently release the heat from same? Easier said than done.

And why does the A/C try to release heat at a time when electricity demand is already at peak level (instead of during the night)?

The A/C causes the peak demand.

Re:Why limit the conversation? (0)

Anonymous Coward | more than 3 years ago | (#36856984)

You know of a practical way to produce this "cold buffer"?

Heat exchanger in the swimming pool.

Re:Why limit the conversation? (3, Interesting)

fph il quozientatore (971015) | more than 3 years ago | (#36856606)

Still, it has always stricken me as peculiar that in the wintertime people spend energy to heat the kitchen up to 20-25 C, and inside it there is a little fridge working as hard as it can to bring the temperature back to exactly the same value as outside.
Not to mention that this refrigerator is typically located just next to the electric cooker...

Re:Why limit the conversation? (1)

bradgoodman (964302) | more than 3 years ago | (#36856732)

Yes, exactly! I'd thought of this - forgot to mention it in my post!

Re:Why limit the conversation? (4, Insightful)

ColdWetDog (752185) | more than 3 years ago | (#36856742)

Still, it has always stricken me as peculiar that in the wintertime people spend energy to heat the kitchen up to 20-25 C, and inside it there is a little fridge working as hard as it can to bring the temperature back to exactly the same value as outside. Not to mention that this refrigerator is typically located just next to the electric cooker...

Convenience and cheap energy. For residential buildings, the money saved generally doesn't amount to enough to support the infrastructure required to transfer and control heat. However, in larger buildings, this sort of thing is rather normal. In theory, you could make smaller units for the house that would take hot air from the refrigerator and dump it into the living room in the winter or preheat the water for the hot water heater, but the ducting involved would either be rather ugly or have to be built in to the house. Wait until heating / cooling gets really expensive, then the savings might justify the hassle.

The other big problem is that we're not talking about a lot of heat. Put your hands on the back of a modern refrigerator - it's warm, not hot. To move energy with low heat values gets harder (read bigger ducts / fans) and less worthwhile. Put your hands on the exhaust of a city sized natural gas fired thermal power plant and you've got some significant BTUs pumping out - it then becomes worth your while to do something with it.

Re:Why limit the conversation? (0)

Anonymous Coward | more than 3 years ago | (#36856994)

Modern refrigerators (produced in the last 10 years) typically use on the order of 60-70 watts or so. Not exactly a lot of potential savings for a lot of effort to route cold air from the outside for a few months of the year.

Re:Why limit the conversation? (1)

russotto (537200) | more than 3 years ago | (#36856610)

Why limit the conversation to just servers, when this occurs everywhere in common life?
Why does my refrigerator take heat out of the inside, and dump it into my house - requiring my A/C to then take it and again put it outside?

Why does my A/C in a house take all the heat and discharge it outside into the atmosphere, which meanwhile a pool heater is running 5 feet away using energy to generate more heat for the pool?

Mostly because trying to use small temperature differences is difficult; solutions tend to be bulky, inefficient, costly, inflexible, and/or complicated.

Figure out a practical and flexible way to do it and you could get rich. So you've got to take the heat from the refrigerator and dump it outside in the winter (either with ducting or plumbing), but in the summer keep it in. Your HVAC outside unit can heat the pool in parts of the spring, summer, and fall (but what if the pool is already too hot? No A/C for you?), but you probably don't want to use it as a heat pump to freeze the pool in winter.

The reason pizza ovens don't have fins and fans for wintertime heating is likely cleanliness.

Re:Why limit the conversation? (1)

Anonymous Coward | more than 3 years ago | (#36856644)

> Why limit the conversation to just servers, when this occurs everywhere in common life?

> Why does my refrigerator take heat out of the inside, and dump it into my house - requiring my A/C to then take it and again put it outside?

This makes refrigerators easier to install and ship, otherwise you would have to plumb the lines to an external radiator, and get special equipment to reprime it each time it is relocated. Additionally you would want the heat inside during the cooler months, requiring two radiators and a valve.

Of course, you could easily make this modification yourself if you believe it will save you money, just need some refrigerant line, insulation tape, a priming pump and some refrigerant.

> Why does my A/C in a house take all the heat and discharge it outside into the atmosphere, which meanwhile a pool heater is running 5 feet away using energy to generate more heat for the pool?

Because you didn't specify your pool or A/C to be configured this way when you had them installed. Likely this wouldn't be that practical, because the months you want your pool heated, are going to be the months you are moving heat indoors, which means the A/C ODU will be getting cold. You generally never want to cool your pool (except maybe in Arizona?), and if the weather is warm enough to use A/C, then it's probably unnecessary to be heating your pool.

> Why do people call incandescent light bulbs "energy wasters", when then can (in the cooler months) defray the work needed to be done by a household heating unit?

Because in the summer or any other time electric heating isn't in use, they are using more energy to produce undesired heat. In cool weather, resistive heating is still less efficient than heat pump systems.

> Why does the Pizza place down the street run their heater in the winter yet has these giant metal exhaust ducts running from their pizza ovens, venting heat to the outside world? (Why no fins/blowers on these ducts to disperse heat into the pizza-joint?)

This would at first seem a mystery, as ideally you would want to keep the heat in the oven. I'm guessing it is a gas or wood burning oven, so removal of the exhaust gas is a neccesity for health reasons. A radiator could be added to the exhaust ventilation, but it would need to be designed with controllable vents so that it can be excluded from the exhaust system when heating is not required.

Likely though, they simply haven't considered it. Or someone has decided that it would be more expensive than operating heaters, rightly or wrongly.

> The point is - people think of heating and cooling on a "unit" basis - and not on a systemic basis of an overall building - or even area. HVAC systems in buildings get this - sort of - they are not single machines - but a system of different, interconnected machines which are each interconnected, performing different tasks - sort of like organs in a human body. This approach needs to be thought of everywhere where cooling is required, and/or heat is generated.

Yes, but in a residential home, there are not really that many sources of heat, or the small scale of them makes it non-viable to install transfer systems between them.

Re:Why limit the conversation? (0)

Anonymous Coward | more than 3 years ago | (#36856918)


Why does my refrigerator take heat out of the inside, and dump it into my house - requiring my A/C to then take it and again put it outside?

Because it's a very small amount of energy we're talking about. Routing ductwork to the outside would cost far more than the little bit of energy you saving in cooling costs.

Why does my A/C in a house take all the heat and discharge it outside into the atmosphere, which meanwhile a pool heater is running 5 feet away using energy to generate more heat for the pool?

Because the people who make the pool heater are different than the people who make A/C. Also, limited market.

Why do people call incandescent light bulbs "energy wasters", when then can (in the cooler months) defray the work needed to be done by a household heating unit?

Because electric heat is far more expensive than gas heat per unit energy.

Why does the Pizza place down the street run their heater in the winter yet has these giant metal exhaust ducts running from their pizza ovens, venting heat to the outside world? (Why no fins/blowers on these ducts to disperse heat into the pizza-joint?)

Good question. Ignorance? Complexity of temperature control? Limited market?

because (0)

Anonymous Coward | more than 3 years ago | (#36857212)

our political/economical/monetary system dictates that "waste" does not exist unless you can profit by eliminating it.

Re:Why limit the conversation? (1)

lapagecp (914156) | more than 3 years ago | (#36857326)

The problem with a lot of these questions is analogous to "Why don't you stop your car to pick up a penny when you drive by one." Its a tremendous amount of work for very little pay out. It doesn't make any sense to try an get the heat produced by your air conditioner to your pool heater when a very cheap solar heater will give you a much greater payout for much less.

Re:Why limit the conversation? (1)

bradgoodman (964302) | more than 3 years ago | (#36857390)

My home A/C runs about $250 a month in hot summer months. A pool heater can run a couple hundred, if heavily used in the same timeframe. Shouldn't these two negate each other, rather than add up? These are not minor expenses!

Re:Why limit the conversation? (0)

Anonymous Coward | more than 3 years ago | (#36857456)

Rethinking building designs and the standards of living which is now ingrained, is nothing new.

Personally, I've been doing it for years. Asking what-ifs for every possible scenario I've observed. Have I implemented anything? Nope. Not a builder.

Sadly, the cynic in me knows getting such improvements and alternate designs into the mainstream building system, or even into prototype is almost a futile effort.

It seems 'Change', has become the enemy of infrastructure. Amusing really, since it's a proven constant.

Energy wasting technology (1)

sjbe (173966) | more than 3 years ago | (#36857526)

Why do people call incandescent light bulbs "energy wasters", when then can (in the cooler months) defray the work needed to be done by a household heating unit?

Because they are energy wasters. The furnace in your house is FAR more efficient at producing heat than the incandescent bulbs. It's not even a close comparison. Your basic point about how we should be using (and re-using) waste heat as much as possible is a good one but that isn't a reason to use energy wasting technologies for their by-products.

I do love the idea of using waste heat in useful ways but let's not generate waste heat on purpose.

Links: Some Good Working Examples (2)

1sockchuck (826398) | more than 3 years ago | (#36856402)

As others have noted, there are many good examples of data center reusing waste heat. Here's a list of examples [datacenterknowledge.com] of server heat being recaptured to warm homes, offices, greenhouses and even swimming pools. This is common enough that The Green Grid recently released guidelines on the best way to integrate heat recapture in key efficiency metrics like PUE (Power Usage Effectiveness).

Remember the Cray 1? (2)

airfoobar (1853132) | more than 3 years ago | (#36856450)

It was a sofa. [google.co.uk]

Xeon-roasted chicken (1)

drolli (522659) | more than 3 years ago | (#36856486)

nothing more to say about this.

Bitcoin miners have known about this for a while (2)

mathimus1863 (1120437) | more than 3 years ago | (#36856494)

I realized this at the end of winter when I had 8 high-power GPUs running in my condo mining Bitcoins, and my central heating was not running anymore. You put your hand behind one of the quad-GPU computers on full load, and it feels like a blowdryer, running 24 hours per day. Seemed to have no problem heating 1200 sqft. This seems to apply to GPUs more than anything, though. I don't know how many CPU servers can produce 1.5 KW of heat...

Re:Bitcoin miners have known about this for a whil (0)

Anonymous Coward | more than 3 years ago | (#36856668)

A core i5 2500k is rated at 95W full throttle, overclock and you get more.

Re:Bitcoin miners have known about this for a whil (0)

Anonymous Coward | more than 3 years ago | (#36857078)

McDonald's and KFC are now accepting bitcoins.

I used them earlier today to purchase happy meals for my kids, and a huge bucket of gravy from KFC for myself. I know it isn't real gravy, but it still tastes good, and it was worth it just to be able to use my bitcoins.

At one place I worked (0)

Anonymous Coward | more than 3 years ago | (#36856628)

The server room was in a small space behind the bathroom. We had to vent the server room into the bathroom and I joked that we could use the 4 inch pipe as a hand dryer. The next day an elbow was on the pipe and a pipe went down to hand level. It was one of the best hand dryers I have ever used.

Energy out = energy in - work done? (1)

KreAture (105311) | more than 3 years ago | (#36856814)

By increasing the number-crunching we increase the power requirements of the CPU and the energy in, and heat out.
This much is self-evident.

However, is number-crunching work? If so how much work?
Feeding n watts into a system and crunching math at m mips. Does that leave us with n-k*m power out where k is some magic constant not necessarily unrelated to m already and varying with the machine?
Do you then want a low k-factor or will they simply increase the power in so that you don't freeze?

Re:Energy out = energy in - work done? (0)

Anonymous Coward | more than 3 years ago | (#36857026)

k is 0 due to conservation of energy. Some people claim a non-zero k (but extremely small) based on the information entropy of what you're computing.

That could work. (0)

Anonymous Coward | more than 3 years ago | (#36856922)

Heat collected in the data center could power furnaces on the lower levels, which would send warm steam to higher floor, so that the environment for directors would be cozy and moisty -- lower levels being cramped and hotter (because of the furnaces) seems a fair price to pay.

Directors would be walking amid the warm fog -- in a kind of Heaven while workers would sit all day long next to furnaces. They wouldn't last much, but hey, who wants to live in such hell forever?

Heater? (0)

Anonymous Coward | more than 3 years ago | (#36856952)

I guess this doesn't help in places where you don't need a heater. Nobody I know ever uses a heater in South Florida. If only the surrounding heat could be put to good use! It sure takes quite a bit of current for central A/C to move the heat outside.

The Article Doesn't Really Make the Point (1)

stevelinton (4044) | more than 3 years ago | (#36857000)

The idea, I think is that these servers are your cloud infrastructure, rather than being used for any local purpose.
Imagine a future where computer technology is a bit more stable than it is now, so a server has a 10 year or so useful life before becoming obsolete. Also you have fibre to every apartment building or office block.

Now, you want to convert some electricity into heat for whatever reason. So you buy/rent a "brick" of servers of suitable size, probably an all-solid state affair with no moving parts at all, plug it into the power and the internet and arrange to move heat out of it for whatever purpose you have. As far as you're concerned that's it, and this is cheaper for you than just buying or renting a conventional electric heating element (ie you get paid, or subsidised power for doing it).

As far as the user of the computation is concerned, they buy computation and related services from Amazon or someone, just as they do now.

The middle-man is running a complex management layer that migrates VM instances and data around the millions of "bricks" that they manage, and allocates each as much work to do as the demand for its heat output requires. Balancing the compute demand against the heat demand requires partly scale, partly non-urgent background jobs, partly blanacing load between time zones and hemispheres and partly a few conventional data centres that can fill in any gap.

Prime 95 (0)

Anonymous Coward | more than 3 years ago | (#36857100)

I remember a cold winter when my heater was on the fritz and the only way I got through the night was a couple of Pentiums running Prime 95.

Why not use (0)

Anonymous Coward | more than 3 years ago | (#36857138)

Peltier devices and harvest the heat? OOPS, that would be possible and ALMOST useful, almost.
Take the heat and use it for a useful purpose... too green possibly?

My PC sits right underneath the thermostat (1)

sandytaru (1158959) | more than 3 years ago | (#36857152)

Just the way the house got laid out, unfortunately. I had to switch to a more thermal efficient system just so the heater would actually run in the winter time.

This is news?!? And they get a paper out of it?!? (0)

Anonymous Coward | more than 3 years ago | (#36857210)

University of Sherbrooke (in Quebec) has had this --- heats up some of its buildings using heat from their (rather large) cluster. I am sure they are not unique or at the cutting edge (Canadian universities are nowhere near as well funded as American ones).

At least there are other people at MS Research that do truly original and insightful work.

BOINC (0)

Anonymous Coward | more than 3 years ago | (#36857406)

I used to heat my living room with just my PC running a few interesting BOINC projects in the colder months of the year:

http://boinc.berkeley.edu/

Not new.. FOOD (1)

Chuby007 (1961870) | more than 3 years ago | (#36857484)

This is not new, I still fry my eggs on an old AMD processor, after that is done, I just place my coffee mug on top, it stays warm all day...
Load More Comments
Slashdot Login

Need an Account?

Forgot your password?