AOL: Outdoor Server Huts Are the Future 146
1sockchuck writes "While Facebook and Apple are investing in huge data cathedrals, AOL has decided to go in a different direction: a distributed network of rack-sized server huts that live outdoors. AOL is taking the concept for its unmanned data center and shrinking it into a 'micro data center.' AOL envisions a distributed network of these units, allowing it to quickly roll out new IT capacity for hyperlocal news sites and create its own content distribution network."
CBG (Score:5, Funny)
envisions a distributed network of these units, allowing it to quickly roll out new IT capacity for hyperlocal news sites and create its own content distribution network
Ahh someone's been visiting the Corporate Bullshit Generator [sourceforge.net] LOL!
Re: (Score:3)
What content is it AOL has that people want to see?
Maybe they should find existing sites to put the servers.
How about Pizza Huts? Help heat the pizza ovens.
Re: (Score:1)
Re:CBG (Score:5, Interesting)
Yes, you can. You obviously wouldn't use the servers as the sole heat source in a pizza oven, but you could use their waste heat to keep the oven warm, so that the mean heating element doesn't have to use as much energy to reach the operating temperature. Granted the savings would probably be minimal, but it can be done. But it probably would take a while to recoup the investment in actually putting it all together.
There's a lot of energy-saving things involving heating and cooling that can be done that people don't bother with because the return on investment would probably be too long. For instance, everyone has a refrigerator in their house, which produces a fair amount of heat in cooling its interior (there's coils on the backside, for cooling the compressor). In the winter, this is no problem, since you want your house warm anyway, but in the summer this is counteracting your house's A/C, making it work harder to keep it cool inside. What if there were some way of diverting that heat to the outside of the house in the summer, but in cooler months keeping it inside? It could be done with some custom ducting, a fan, and temperature-controlled louvers. But you never see this because that's a lot of work, and therefore expense, when the gains are probably minimal.
Similarly, clothes dryers produce a LOT of waste heat, but this is usually just ducted outside. My house actually has a diverter box in the ductwork so that in the wintertime, I can move the flap and let all the hot air blow into the house (through an additional screen to keep the lint down), which reduces my heating bills. Unfortunately most houses don't bother with this, though it'd probably have a much bigger effect than the refrigerator idea above.
As energy costs rise, I imagine we'll be seeing a lot more energy-saving strategies like these.
Re:CBG (Score:4, Insightful)
You must like it humid in the house.
Re: (Score:2)
Humidity traps for these systems are readily available. They don't work amazingly well, but they do remove a notable amount of the moisture from the air in question.
Re: (Score:2, Offtopic)
Humidity traps for these systems are readily available. They don't work amazingly well, but they do remove a notable amount of the moisture from the air in question.
Well which is it? Either they don't work well, or they do.
Well which is it? Either you own a dictionary or... well, I see you don't.
Re: (Score:2)
"Brevity is the essence of wit."
Re:CBG (Score:4, Funny)
"Brevity is the essence of wit."
I see what you
Re: (Score:2)
Now you're starting
Re: (Score:2)
...
Re: (Score:3, Informative)
During the winter months (when the heating would be necessary), additional humidity is often beneficial. Air inside can get *dry* with the temperature difference's effect on relative humidity.
Re: (Score:2)
I keep hearing about that, but I have not experienced this, at least in California. The dry days, evidenced by bloody noses, happen in the summer with the AC running. We also had a humidity gauge awhile back. It varied by rainy days and such, but otherwise not much.
Re: (Score:3)
Take a short drive to Phoenix. It's dry in the winter (it's dry all year, except during the "monsoon season" in the summer), and while not exactly "cold", it can be chilly enough to want a little bit of heat, and the small amount of humidity that dryer exhaust would add to the air isn't a problem, instead it's probably welcome.
Re: (Score:2)
Re: (Score:2)
Hence I dislike going to Tahoe.
On the coldest nights and mornings at our house, we'll hit around 30F. We keep the inside at something like 60, but the days heat makes the night heating not too bad. When the nights are in the 30s, the days are usually in the 50s.
Re: (Score:2)
I live in Phoenix. A little extra humidity inside in the wintertime is a welcome thing.
But.. When it is very cold it is dry. (Score:2)
Re: (Score:2)
So you're saying that in Soviet Russia, dryers un-dry you?
Exactly! (Score:2)
Nobody is dry in Russia...
**Hic**
Re: (Score:1)
On the surface, your thinking looks good-- find heat transducers and reservoirs to store it. Distributing servers causes massive redundancies of the fixed gear needed to store and reuse heat. Convenience memes that create dysfunctional residential electrical power states are a problem, but people don't like sharing, don't like living in dorms, don't like high, up-front costs of shelter, and so you're fighting many problems.
As to AOL's distributed "NOC", the rationale for it is almost ludicrous. The local/re
Re: (Score:2)
Someone at AOL must know this entire idea is simply ludicrous but if it gets the financial backers of AOL (whoever the hell they are) to let the company live another day then its worth a long shot.
Re: (Score:2)
No, it doesn't. The backers get strung out and lose the investment. The mgmt at AOL/HuffPo/etc. don't get changed. Civilians that invested in AOL get screwed. Living for another day isn't necessarily a good thing. When the entire economy is down, then the tendency for everyone to take their lumps at once is a well-known way to keep financial pain from having an effect on stock price. But that isn't the case here. AOL is portending that this is somehow visionary, and solves big problems, where it doesn't do
Re: (Score:2)
How is living another day not a good thing? Yes, the investors and backers may be screwed, but who cares about them? The only important people are the executives, and as long as they get more bonuses, that's all they care about.
Re: (Score:2)
As to AOL's distributed "NOC", the rationale for it is almost ludicrous.
And it's not even clever. Some "datacenter on a container" providers already exist, that go way beyond a couple of racks, have neatier insulation and are completely mobile if you need it. And if you want small, there are half-container solutions.
Re: (Score:3)
Re: (Score:3)
If you still have soap in your clothes when you dry them, you're either using way too much soap (quite likely; lots of people make this mistake, thinking they need 4x as much detergent as they really do), or you need a new washer. The rinses (multiple) are supposed to get rid of all that. "Wash byproducts"? Like what? There shouldn't be anything going into the dryer except wet, clean clothing. The clothing material thing I might buy, but even there you should be using heat appropriate for the fabrics;
Re:CBG (Score:4, Informative)
Your refrigerator idea won't save energy in the summer. The efficiency of the fridge is related to the temperature difference between the coils and the air around them. If they are exposed to a hot outdoors, the fridge will simply work harder. If you expose them to the cool indoors, the fridge gets a break at the expense of the house's A/C which has to pump the extra heat away.
In other words, something has to overcome the temperature difference to push the heat outside. It's either the fridge or the fridge + A/C and in neither case are you going to see any savings.
Re: (Score:3)
This is nonsense. Your home A/C isn't 100% efficient, and the refrigerator's compressor motor generates its own waste heat (in addition to the heat it's pumping out) which can be vented to outside with no additional energy required.
Re: (Score:2)
Neither is 100% efficient. Running one harder or the other harder will make no significant difference. The compressor was partially my point. If you expose the fridge coils to the outdoors, the motor has to run harder, so it may even be less efficient. In addition to this, A/C units tend to use less power per unit of heat exchanged (i.e. are more energy efficient) so having them do the work is better, meaning that the GP's idea won't work (as I said).
Re: (Score:2)
Until you can be bothered to run the numbers on your scheme, you are just an ignorant loud-mouth that likes to argue.
Re: (Score:3)
No, unless you have a perfect reversible refrigeration cycle, you will save a little energy.
You might be able to save a considerable amount of energy in the winter, though, by rejecting heat directly to the outdoors, or using "free" economizer cooling when (and if) the temperature outside is below the refrigerator temperature. Of course, if you're heating with electricity, anyway, it wouldn't really make a difference.
An better energy saving idea, th
Re: (Score:2)
I've thought about a number of different options for saving energy related to heating and cooling.
In a house you could use a diverter to run water from the bathtub drain through a radiator in the basement in winter instead of just sending all that heat into the sewer.
In the summer time you could make use of a radiator on the water supply to dump heat from the house into the water that's been cooled by the ground, then send that heat into the sewer.
When I see an indoor rink with a pile of ice from the zambon
Re: (Score:2)
The dryer thing sounds like a fabulous idea to me. The fridge I kinda have my doubts. If you went the fans and duct work route the compromise to weather envelop of an otherwise well insulated house would likely result on the loss of more energy then you spend having the AC move the waste heat out. Keep in mind most energy loss is around doors and windows, solid wall with insulation and house wrap over it has a pretty high R value.
The better approach to the fridge might be some valves to select of a secon
Re: (Score:2)
What should really happen is that the fridge and the AC shouldn't be separate units. The AC should offer a refrigerant tap and communication feedback line for the fridge so they can be combined into a single system so that the fridge's compressor is completely bypassed during the summer.
Re: (Score:2)
If you are in climate you use the AC all year round or in one where you turn it around and use it to pump heat in in the winter I can see that. There are lots places where the AC unit is simply not used though during 6+ months of the year and it would be way to large to use to efficiently keep a fridge cool.
Re: (Score:2)
That's why I said, "bypass" and not "replace"
The problem could also be solved by having a pony compressor unit for use when the AC demand is limited to the refrigerator. It shouldn't matter that the refrigerant is delivered to other parts of the house as well if the refrigerant is actually a room temperature liquid that cools by endothermic phase transition.
But yeah, modern refrigerators use what.. 50W average? Unless there's a huge COP (and it can't be more than 15 for a 4C fridge in a 25C room), it's pr
Re: (Score:2)
But yeah, modern refrigerators use what.. 50W average? Unless there's a huge COP (and it can't be more than 15 for a 4C fridge in a 25C room), it's probably less than the lighting or a couple sleeping adults.
My point exactly. Even if we did come up with some elaborate way of improving the total system efficiency by piggy-backing onto the house's A/C or whatever, the actual energy savings would be quite minimal compared to the expense required to achieve that gain. At current energy prices, it just isn't w
Re: (Score:2)
It's a recipe for mold and moisture damage, unless you were to go to considerable trouble to get rid of excess humidity, especially in the space where the dryer is located. (I had an unvented dryer in my basement once for a year or so, I stuck the flex vent out the window in the summer, but tried to let it vent indoors in the winter - it wasn't a great idea) YMMV, though, as this depends partly on the climate.
Re: (Score:2)
All that moisture would just help mold and mildew grow.
Mold and mildew isn't a problem where I live; it's a desert here in Arizona, there's barely any humidity in the winter air at all. Humid air from the dryer is a good thing.
There are coils, not there is (there's) coils. But I digress.
Wrong. "There's [plural words]" is perfectly acceptable colloquial speech. I'm not writing a formal dissertation here, this is Slashdot.
http://itre.cis.upenn.edu/~myl/languagelog/archives/002447.html [upenn.edu]
Re: (Score:2)
Re: (Score:2)
Actually, yes it does, when you're talking about language, especially the English language. Language is defined by how people use it. The "proper" usage is defined by how it's used by the people who speak it. So if it is common, by definition, it IS acceptable.
If most people start saying "There is coils", then by definition, it is correct.
Re: (Score:2)
AOL owns the Huffington Post, amongst a lot of other sites, that get a lot of hits.
Re: (Score:2)
oops, AOL, not amazon, my bad.
Re: (Score:3)
Corporate Bullshit Generator [sourceforge.net]
Bookmarked. I'm going to try to use a line from there in my PP presentation to Sales on Monday. Ahh, here we go:
"Market-driven cost efficiencies influence our perspectives."
Makes sense (Score:4, Funny)
Their clients still live in huts, so it makes sense to move their servers there too..
If only (Score:2)
Re: (Score:1)
Have they thought this through? (Score:2)
Re: (Score:2)
also what about being hit by cars / trucks?
Hit by wind blown debris??
Stuff getting jammed / blocking air vents?
Flooding?
Re:Have they thought this through? (Score:4, Funny)
Re: (Score:2)
"The data inside"? This is AOL. They can't give their data away.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
It's a lot easier to seal the perimeter of one 500,000 square foot data center than of 500,000 one square foot centers.
Re: (Score:2)
It's a lot easier to seal the perimeter of one 500,000 square foot data center than of 500,000 one square foot centers.
True but the impact isn't the same. If you miss 100 of your 500k one square foot centers, only 0.02% of your servers are compromised.
If you miss 100 spots in your 500k sq ft data center, 100% of your servers are at risk.
Lol (Score:3)
Counting the minutes... (Score:3)
Re: (Score:2)
It definitely wouldn't work where I live, in Phoenix Arizona. Half the area's inhabitants are meth-heads who'll steal anything that isn't bolted down.
Re: (Score:2)
It definitely wouldn't work where I live, in Phoenix Arizona. Half the area's inhabitants are meth-heads who'll steal anything that is or isn't bolted down.
FTFY
Re: (Score:2)
Maybe, maybe not. Remember, in many places in the country (including meth-head infested areas like Phoenix), we have copper thieves getting themselves electrocuted so they can steal electrical cables, plus people stealing catalytic converters from underneath parked vehicles. I'm not sure how large this "server hut" is, but I sure hope it isn't on wheels like a trailer or else it'll be easily stolen.
Re: (Score:2)
Yep, and they also have portable sawzalls so they can steal catalytic converters!
Re: (Score:2)
Or the data?
Maybe they have some good encryption and intrusion detection, maybe not. Maybe you have to toast the first few boxes before you figure out how to suck the data out of one without it suspecting.
Re: (Score:2)
You must not have read the article!
“Security is always a concern, but in general these devices have more in common with equipment you would see in an equipment yard for most buildings than standard data center environments,” he said. “Wide-scale deployed units won’t have the fancy paint jobs and logos, so they will very much look at home in those equipment yards.
See? It's even better than security through obscurity, it's security through paint jobs! I hear covering it in dog shit a
Why? (Score:2)
I don't see the use case for these. Any sizeable office that wants server capacity is going to have a room indoors where they can put it - pouring a concrete slab outside and running power + network + water to it doesn't seem much better than dedicating a small server closet inside the building. I can see some niche market in remote sites (mining sites, research sites, etc) that need more servers than they can stack in the corner of their office trailer, but if they have that many servers, this single rack
Re: (Score:3)
"They are probably more interested in a datacenter in a shipping container."
Indeed. ISO containers are cheap and you can send one to your site with everything you need AND have storage or workspace INSIDE the container protected from the weather.
If your "AOL in a can" breaks in foul weather, then you need to COVER it to OPEN it.
You don't need to pour a slab for ISOs as they are supported by the corner fittings. A railroad tie under each end is usually just fine.
I use ISOs for personal shop buildings and sto
Re: (Score:2)
I don't sell ISOs, but do have morbid nasty love for them.
Where should somebody aspiring to this level of sickness start?
Re: (Score:2)
http://tech.slashdot.org/story/08/09/22/1448226/microsoft-innovates-tent-data-centers [slashdot.org]
^^
A solution in search of some problems (Score:2)
Right off the bat, you're going to combat environmental issues, but the biggest problem I see is how you secure these systems, their code and data in an adequate manner.
In short: stupid!
If There's Ever A Company To Trust (Score:5, Funny)
If there's ever a company to trust when it comes to predicting the future, it's AOL. Why just three years ago they predicted they'd have to move away from dial-up since broadband would be the wave of the future and look at how right they were.
Re: (Score:1)
Not to mention they can charge double since broadband was their idea. Along with Al's.
Interesting choice of color (Score:2)
Security? (Score:3)
Isn't one of the tenets of network security physically securing access to the servers? How would they prevent someone from tapping into the boxes and either sniffing traffic or directly stealing content?
Bonanza! (Score:4, Funny)
No Bad movie references yet? (Score:1)
Come on guys, this is Slashdot, no one has yet made the comparison of this idea to one of the worst movies of all time, that being Manos: The Hands of Fate [wikipedia.org] given the name of the CTO at AOL that is announcing this "breakthrough" in computing?
And here's an idea to help with the cooling on this outside enclosure - don't paint it BLACK!
Cool (Score:1)
This has seemed like an obvious idea to me for a while. I figured Google would be the first. But I'm glad somebody is finally doing it. I wonder how many of these AOL could possibly roll-out? Hundreds, at least. Thousands?
Bargaining chip (Score:2)
Sounds like a bargaining chip as they negotiate for colo space in switching centers. It might be useful in some special situations where you can terminate dark fiber into your own box and save on backbone costs, then have a short distance link into some other facility.
I would have expected this from Comcast or Verizon. If a local box held a few thousand hours of video cache, including recent TV and movies, most requests might be satisfied locally, unloading the upstream network.
Humidity? (Score:2)
How will they handle humidity? They've been designing servers to handle higher temperatures to save on air conditioning, but they won't deal well if the humidity gets too high. I expect these outdoor servers will get lost in the fog.
They'll also have issues with air pollution, insects, and animals.
There are plenty of data centers that use outdoor air for cooling, but they filter it and monitor the humidity.
Re: (Score:2)
What matters to electronics is relative humidity not absolute humidity (and the tolerable range is pretty wide 5% to 95% according to wikipedia) snd the absense of condensation. Condensation is generally caused by rapid changes in air temperature that leave solid objects colder than the surrounding air.
So as long as the following conditions hold I don't think humidity will be too much of a problem.
1:internal temperature is higher than outside temperature (it will be unless you are employing refridgeration t
'Live' outdoors (Score:2)
I for one welcome our new Living Server Overlords.
And for the broader view... (Score:2)
I could not think about this without wondering what Compuserve are up to ... so I looked at their site http://webcenters.netscape.compuserve.com/menu/about.jsp [compuserve.com] and cracked up when I saw the Netscape logo at the bottom... These people still think that they are important :-)
Re: (Score:2)
They have actually updated their "dialer" software for Windows 7, as if there is anyone who upgraded to Windows 7 but connects to the internets via dialup. There must be a group of computer programmers from the 90s that have been kept in a cave somewhere that still maintain the CompuServe network and are mostly unaware of what has changed in the last 15 years. Maybe 5 years from now they'll find out about that Facebook thing everyone has been talking about.
Similar to electricity distribution AC vs DC (Score:2)
Anyways the pendulum goes back and forth on these things, and if period doubling is occurring that means that fine grained rapid deployment is required to keep pace with that. Won't be long before data centres are riding electric trains, semi-trucks (a la Walmart's warehouse on wheel
Does this mean... (Score:2)
will it have? (Score:2)
...and (Score:2)
Jabba the server Hutt.
If, and ONLY if... (Score:2)
...they make each "hut" look like a Tardis.
Then, we can talk.
AOL: We hate our IT workers.... (Score:3)
Server huts... so you get to work outside when you work on these things. Rain, snow, 120 degree heat. No thanks.
I worked long enough in the Cellular industry to be abused by working on these god-for-saken "huts" they think they invented... The Cellphone industry has had them cince the late 90's.. it's not new, they are not innovative. And they SUCK to work on.
Re: (Score:2)
sounds like AOL to me (Score:2)
Hmmm. (Score:2)
New account, 100% MS AstroTurf... One wonders...
Re: (Score:3)
"I can clearly see that they don't care about the environment."
Citation fucking needed NOW.
The tiny servers-in-a-can model is cute, (the stupid black paint job has to go, try that in AZ in August and you couldn't even open the door without burning your hand), but centralization makes for much easier operation and maintenance for large server farms.
Here's a citation about google and the environment (Score:3)
Re: (Score:2)
Centralization puts similar objects where a small group of techs can have immediate hands-on access, store spare parts and spare machines, have backup power with economies of scale (large diesel or Capstone turbine gensets), and directly guard and monitor equipment.
That's simple physical, observable reality not requiring citation.
Isolated units which are small involve sacrificing advantages of immediate service. If no workee, dispatching a tech by car or truck with tools and parts is required. If it gets se
Re: (Score:2)
Re: (Score:2)
Citations please? The have been articles describing things Google has done that lowered the cost and power consumption of servers. They did things like using higher efficiency power supplies, and avoiding some extra DC-DC conversion.
Although large businesses can often be criticized for choosing higher profits over environmental responsibility, servers farms are clearly an area where better environmental impact generally means lowered costs. More efficient servers cost less to power, maybe be more reliabl
Re: (Score:2)
Avoiding DC-DC conversion? I thought they made more use of it to simplify the power supplies themselves, the connectors, and the motherboards.
Re: (Score:2)
IIRC the consolidation I'd read about was partly a matter of skipping going from DC to 120 VAC 60 Hz and back to DC again following individual UPS batteries. I believe the motherboards had an efficient high-frequency switcher for the bulk of the load, but not a full DC-DC converter. There was no isolation transformer and no rectifier beyond commutating diodes.
The DC distribution voltage wasn't really low, which helps avoid excessive conductor losses and those in the preceding rectifier. I might not be rec
Re: (Score:2)
Per rack is a different story and a lot less watts, but still, moving things to DC means moving everything to DC including test setups. There's far too much mucking about until you can dump a faulty server on a table and plug it into a DC socket.
Re: (Score:2)
True as anyone who's done HVAC maintenance knows.