Slashdot Mirror

Have you guys ever seen a datacenter? "Nondescript" is one word for them. Who would want that in their town?

Yes I have. Here is Apple's datacenter in North Carolina. It looks like a one-story building in the middle of nowhere that you can't really see from the road.

>"But here's the most interesting part. GeekWire reports: Amazon is moving ahead with a unique plan to use heat generated from data centers in the nearby Westin Building to warm some of its new buildings downtown. The system transfers the heat from the data centers via water piped underground to the Amazon buildings"

Factories and businesses that generate waste heat have been doing that for at least two centuries now, all over the world. Where I work, some 80 years ago they ran waste heat steam lines from the laundry building to other places on the campus, including 1/4 mile away for some residences. Data centers have also been doing it in many places for many years both on and off capus. http://www.datacenterknowledge...

It is great to hear, but really nothing new.

Thus why Google started encrypting their DC-to-DC traffic four years ago. Microsoft is also doing the same. I would assume Amazon does the same, although I can't quickly find any article which they claim this.

I get the Netflix analogy, but, ya know, Netflix is all-in on Amazon Web Services (not Azure).

"If you have not successfully tested a restore and you do not have a completely offline copy, you do not have a backup."
OK, now that I have repeated it, let me add.

As a CEO, and a CTO, you MUST test backups and resiliency by artificially creating downtime and real-life events. You switch off the main server. Or instruct the hosting company to reboot the main server, unplug the main hard drive, and plug it back in. Then you sit up, and watch with great interest what happens.

THEN you will see, for real, how your company reacts to real disasters.

The difference is that if anything _really_ wrong happens, you can turn the hard drive back on and fire a few people.

Smart companies do it. For a reason. Yes it creates downtime. But yes it can save your company.

http://www.datacenterknowledge...

Merc.

There are very few CEOs who actually give a shit about that "IT stuff", and if anything really wrong happens, the company is going to lose more than fucking "downtime". Data loss usually holds a dollar value and impact. Doubly so for a public company, which the one getting fired for pulling dumbass stunts might be the CEO by the Board.

And the example company was maintaining half a dozen data centers at the time of the test, with a literal army of support staff running pre-tests prior to the main DR test. That hardly qualifies as profiling the average company infrastructure who can hardly afford to run a second backup data center or personnel to properly run it. The smart CEO knows how cheap they are, which is why we almost never hear of them being executed by the executive staff.

"If you have not successfully tested a restore and you do not have a completely offline copy, you do not have a backup."
OK, now that I have repeated it, let me add.

As a CEO, and a CTO, you MUST test backups and resiliency by artificially creating downtime and real-life events. You switch off the main server. Or instruct the hosting company to reboot the main server, unplug the main hard drive, and plug it back in. Then you sit up, and watch with great interest what happens.

THEN you will see, for real, how your company reacts to real disasters.

The difference is that if anything _really_ wrong happens, you can turn the hard drive back on and fire a few people.

Smart companies do it. For a reason. Yes it creates downtime. But yes it can save your company.

http://www.datacenterknowledge...

Merc.

> all economy runs down to energy.

^ THIS.

> society can get all of its energy needs without energy being built using big data infrastructure.

I'm not sure how you missed the fact that (Big) Data leads to Knowledge which leads to Power and Energy and ultimate Money.

That's why Apple, Amazon, Google, Microsoft, etc. are all building huge data centers. They want a piece of the pie of influencing & controlling because ultimately it will bring profits.

There is plenty of evidence that most bitcoin mines are moving to use hydroelectric power and geothermal. They aren't doing so out of a sense of environmental altruism , but simply because "greener" power is less expensive. http://gizmodo.com/why-bitcoin... http://www.datacenterknowledge... https://www.cryptocoinsnews.co... http://dealbook.nytimes.com/20... http://www.coindesk.com/my-lif...

http://cumulusnetworks.com/blo...
http://www.datacenterknowledge...
http://opennetlinux.org/
http://www.opencompute.org/
http://www.wired.com/2013/03/b...

Get with the times, the Big Iron Networking gear (like usead at Google and Facebook) are switches running Linux.

they obviously know how to build massive datacenters

I wouldn't be so sure about that.

I'm not so sure about that. One wonders if this was the cause of Amazon rebooting every EC2 instance a while back. They run on top of Xen.

, but it was NEVER needed in practice.

Good for you that your center never has had a fire. Of course, rip out the Halon system (since clearly it isn't needed) and enjoy the day you do have a fire. It does happen, we had one about a year ago in Stockholm, http://postandparcel.info/5951... and in 2011 there was the NYSE, http://www.datacenterknowledge...

These were those I know of off the top of my head, I doubt they're the only ones.

There is a realization that with cheapest, lousiest cooling a substantial number of servers may kick the bucket, but replacement costs are still lower that energy costs long term. See this paper for example. Liquid cooling doesn't fit the bill for general computing, although it may for very specialized cases like quantum computers that need to be cooled by liquid helium.

Why on earth would they call for 240V 60Hz? (almost) Nobody uses that.

Whoah, google spent $7,300,000,000 on data centers last year and doesn't even know what voltage they run on? Time to sell my stock!

Google built its business by developing its own infrastructure, starting with custom servers in pizza boxes and lego. What crazy company runs out and develops its own filesystem for its own internal use? Google did.

What? Who would setup there racks that way! As if you wouldnt have them all coming into the same network and being aggregates somewhere and having routing determine which links are used. I seriously doubt they segregate parts of their server farm for different interconnects. That just sounds inefficient.

Anyone with a large customer base that is pushing out a LOT of data (like Netflix). For the most part, they pay CDNs to deliver their data to the last-mile networks. With the Comcast and Verizon deal, they are co-locating their servers in Comcast and Verizon's data centers. The content they deliver is fairly static, so it makes sense to do it that way. User account data and other volatile data probably takes a different route, but when you start a movie streaming, you'll connect to the closest delivery point on your network. For Comcast and Verizon customers, that's now a shorter route.

http://www.datacenterknowledge... It's been done before.

(sorry for the duplicated posting; the previous one was cut because of problems with the html marks)

In order to obtain a 90% reduction in the energy bill, cooling must account for 90% of the power of the DC. This implies a PUE >= 10. As a reference, 5 years ago virtually any DC had a PUE lower than 3. Nowadays, PUE lower than 1.15 can be obtained easily. As a referecence, Facebook publishes the instantaneous PUE of one of its DC in Prineville, which at the moment is 1.05. This implies that any savings in cooling would reduce the bill, at much, in a factor of 1.05 (1/1.05 = 0.9523).

On the other hand, I believe that this is not the first commertial offer for a liquid-cooled server, Intel was already considering two years ago, and the idea has been discussed in other forums for several years. I can't remember right now which company that was actually selling these solutions, but I believe it was already in the market.

In order to obtain a 90% reduction in the energy bill, cooling must account for 90% of the power of the DC. This implies a PUE >= 10. As a reference, 5 years ago virtually any DC had a PUE instantaneous PUE of one of its DC in Prineville, which at the moment is 1.05. This implies that any savings in cooling would reduce the bill, at much, in a factor of 1.05 (1/1.05 = 0.9523).

On the other hand, I believe that this is not the first commertial offer for a liquid-cooled server, Intel was already considering it two years ago, and the idea has been discussed in other forums for several years. I can't remember right now which company that was actually selling these solutions, but I believe it was already in the market.

Netflix is having all these problems because they use Cogent, the cut-rate morons of the transit world...

This has happened hundreds of times, long before they carried Netflix streaming video:

http://www.pcworld.com/article...

https://secure.dslreports.com/...

https://secure.dslreports.com/...

https://secure.dslreports.com/...

http://www.complaints.com/2008...

http://publicpolicy.verizon.co...

http://www.prnewswire.com/news...

http://www.fiercetelecom.com/s...

https://www.datacenterknowledg...

etc., etc.

I've seen what happens when a voltage regulator failure on a blade takes out the entire 12V rail on the blade chassis (as well as taking out the blade next to it). No one that cares about reliability is going to run a single chassis.

Yeah? What about nuclear war? EMP? Zombies? Ebola? Solar flares? Mice chewing through wires? I mean, yes, you're right, but it comes down to how many contingencies you can possibly plan for, and how much budget you have to plan for them all.

That's what geographical redundancy is for - it protects you from everything that could affect a single datacenter or, with enough distance, it can protect you from a disaster that affects an entire state, or even an entire coast.

I can't plan for a one in a million event -- but neither can cloud giants, as we've seen even the mighty Amazon go down. Shit just happens sometimes. We want to minimize it of course, but be realistic.

As far as I know, Amazon has never had a multi-region outage. In 2 years on AWS, we haven't experienced any multi-availability zone outages that affected us, we've split our servers across AZ's, and though we have experienced single-AZ outages, we haven't been hit by any multi-AZ outages. Once a quarter we do a live failover to the our backup site in an Amazon region on the other side of the country to test it out. Our monthly spend at Amazon is about half what we were paying in colocation costs (which doesn't take into account the cost savings on equipment and maintenance contracts).

If I were paranoid enough and given unlimited funding and development budget, I'd host the disaster recovery site using a different provider, but we have a lot of management scripts that rely on AWS API's to monitor, manage, and scale the site as needed (spot instances help keep the costs down for some of the big data analysis jobs we run) and we rely on some AWS services that would make it inconvenient to use a different provider. That's a known risk that ties us to Amazon, but so far it's been a fair tradeoff.

The disaster recover site costs us very little - we keep a bare minimum number of servers running there that can host our service with some more resource intensive features turned off, and after failover, we can spin up more servers as needed, but even if the backup region is out of spare capacity due to a large number of customers migrating, we can provide our core services using just the reserved instances we keep running all the time. We're working on generalizing the failover so we can spin up our site in any of AWS's regions.

Does anyone really run their border firewall on the same blade chassis that run their servers?

I didn't say it was a good idea or that I liked it; just that it can be done... and sadly, I have done it.

Isn't that your entire point? That a blade chassis with 2TB SSD's is cheaper and better than using a cloud provider?

Why suggest it if it's not a good idea? There are an unending number of ways to host an unreliable service.

That's what businesses say when they haven't had a week-long outage because a transformer blew a hole in the side of their colocation center.

Sounds like another one of those crazy random events (if that actually happened). Although... a week is a long outage for something like that. That sounds like someone isn't prepared, regardless of whether you've got iron or cloud.

It's usually not as spectacular as an actual explosion, but datacenter center outages do happen. You think a week is a long time to recover from a datacenter explosion? I'm impressed that in this case they managed to get temporary power to customers in only a week after the explosion blew out the electrical room walls several fee

Reminds me of this story of a data center currently in an old Boscov's that was interested in taking over the entire mall: http://www.datacenterknowledge.com/archives/2013/06/13/cybernap-looking-to-expand-into-entire-failing-mall/

Are you looking at the same pictures? At least half of the pictures is a condenser.

http://www.datacenterknowledge.com/wp-content/uploads/2013/11/applied-control-novec.jpg

??? How old ARE you? (OMG: I'm only 55 -- maybe I really am older and more paranoid than I thought.)

Let me get this straight: you gave away control of your unencrypted files to someone who wasn't a known personal friend and then am surprised that something happened to them??

I treat on-line services slightly differently: I keep local copies of EVERYTHING that goes out, and I'm surprised when it's still accessible online 5 minutes later, never mind 5 years later. And controlling exactly who has access to it? That's just a fantasy -- really. It's actually binary: either it's out there and they MIGHT have it, or it's not and they DON'T.

I do run Dropbox and use KeePass as a password manager. The credential store is encrypted, but even then the stored password there just isn't "quite right". Phone camera pics get uploaded to Dropbox. At times I'll AES encrypt and email or use Dropbox and expose. For stupid pics I'll just dump 'em out there straight. But I know what's exposed and encrypted-exposed. The latter die soon after they're used.

You store important and critical (tax receipts, lawyer-enforced) notices that might cause breach of contract? And you put control of that in someone else's hands, paid for or not? What kind of an IDIOT are you? Then again, you must not think much of the breaching penalties. That's great, I'm glad you're so confident at everyone always doing the right thing everywhere and nothing bad ever happening.

Me, if I'm going to have a some contract or data leakage it'll be because *I* did it myself and have no one else to blame. Then again, it's obvious digital computer files and paid services will stay around forever: Just ask MegaUpload, GeoCities, and LavaBit. Oh, and the data center located in the Twin Towers? Onsite backups sure came in handy there. Some got thru better than others: One, Two

Then again, there's this brand new data center that will hold all of your data for years -- all for free! I'm sure you can retrieve all of your data from that.

Really, I'm glad things are going so well for you, with the exception of a few bumps. And local storage doesn't solve everything either -- drives can be stolen, warrants can be served, computers can be hacked and data downloaded. But damn it, for 99.9% of my data, I'm 100% directly responsible for it. Offloading everything to the cloud is just offloading responsibility, never mind anything at all to do with the NSA.

Oh, one last thing. Even if all of the employees in the ISP, supporting companies, 3rd party vendors and everyone involved are all above reproach. are you sure? And even say all of the software is 100% vetted and accurate (ignoring accidental software bugs): oops.

Paranoid? Probably, but then again most things don't deserve multiple layers of defense. Only a few do, and of those only a select few get vetted, encrypted, backed up, and rotated offsite. But as for "What would you need if everything was suddenly gone (house fire) and you could only keep a couple of things?" Well there's your answer.

Good luck with it all; hope you produce a updated

VPN via Sweden, are you freakin kidding me - you might as well cc all your data to GCHQ directly!? Sweden's NSA Spy Links “Deeply Troubling”, or check out the professors blog for ongoing abuses on all fronts by the Swedish authorities. Whatever cred Sweden may have established during the cold war years, they have more than used up and are still digging down. The country (well its political leaders) can't be trusted - not a good place to do business anymore.

If any country near the UK has some semblance of credibility, perhaps try Iceland as the first hop for your VPN. They are even trying to promote themselves as a naturally cooled server hub, which is nice...

How easily people forget that AWS is Amazon's excess server capacity.

Is this common knowledge? I've never heard this before. Do you have a source?

Vogels also noted that Amazon eats its own dog food. As of Nov. 10, 2010, all of the web traffic for Amazon.com is being served by Amazon Web Services, he said.[1]

[1] Miller, Rich. Amazon Cloud Now Stores 339 Billion Objects. Data Center Knowledge, June 22nd, 2011.

You might also like: to try using google before asking questions thus easily answered on slashdot.

has their office and their servers in that same building can't afford a DR Plan.

Every business can afford an IT DR plan. If you can have a disaster, you need a recovery plan. That plan might be to take the off site tapes and run down to Best Buy and get a new computer. But it's still a DR plan. Every serious business should have contingency planning. What happens if you lose your phone service? A power outage? A tornado? Flooding? A disgruntled employee? A drunk driver through your front door. How does your business survive when your customers can't contact you or you can't provide whatever puts money in your accounts. The cost of the contingency plan will be directly related to how long you can afford to be without those resources.

As far as moving your business to the cloud, what happens when the 'cloud' eventually has a failure. And it will at some point. Some will be large, some will be small. Will your business be important enough to encourage a 3rd party to respond quickly to restore your service. Because now the response to any outage is directly related to how much income you provide for that 3rd party, not how serious that outage affects your business.

Pushing more traffic into Verizon's network than you pull, means that Verizon's users are requesting data from you.

Umm, no it doesn't. We're not talking about last-mile links here, we're talking about backbone. If I'm Cogent, and I need to get traffic from San Francisco to New York, I can dump that on Verizon's network (or anyone else I'm peering with) and their network will dutifully forward the traffic all the way to NY. The end-point could be AT&T, Comcast, or even another Cogent customer, but dumping it on Verizon's network saves Cogent money, not having to utilize their own backbone.

And this is exactly what Cogent has been repeatedly accused of doing in the past, by pretty much EVERY TIER-1 ISP. Here's just a few examples:

https://secure.dslreports.com/shownews/92749

https://www.datacenterknowledge.com/archives/2008/10/31/peering-dispute-between-cogent-sprint/

http://www.prnewswire.com/news-releases/level-3-issues-statement-concerning-internet-peering-and-cogent-communications-55014572.html

http://www.datacenterknowledge.com/archives/2009/10/22/peering-disputes-migrate-to-ipv6/

http://www.cybertelecom.org/industry/cogent.htm

Pushing more traffic into Verizon's network than you pull, means that Verizon's users are requesting data from you.

Umm, no it doesn't. We're not talking about last-mile links here, we're talking about backbone. If I'm Cogent, and I need to get traffic from San Francisco to New York, I can dump that on Verizon's network (or anyone else I'm peering with) and their network will dutifully forward the traffic all the way to NY. The end-point could be AT&T, Comcast, or even another Cogent customer, but dumping it on Verizon's network saves Cogent money, not having to utilize their own backbone.

And this is exactly what Cogent has been repeatedly accused of doing in the past, by pretty much EVERY TIER-1 ISP. Here's just a few examples:

https://secure.dslreports.com/shownews/92749

https://www.datacenterknowledge.com/archives/2008/10/31/peering-dispute-between-cogent-sprint/

http://www.prnewswire.com/news-releases/level-3-issues-statement-concerning-internet-peering-and-cogent-communications-55014572.html

http://www.datacenterknowledge.com/archives/2009/10/22/peering-disputes-migrate-to-ipv6/

http://www.cybertelecom.org/industry/cogent.htm

What, the Microsoft Cloud Services going offline!? That'll never happen!

http://www.datacenterknowledge.com/archives/2013/02/25/windows-azure-cloud-crashed-by-expired-ssl-certificate/

WTF 80 deg F (approx 27 deg C) is quite warm in a Data-centre especially in a "cold aisle" and 95% humidity is criminal.

You're used to classic datacentres, where the goal was "shove as much cold air into them as possible", i.e. "the lower the temperature the better". It all depends on how the datacentre was built, how its cooling system is/was engineered, and an almost indefinite number of variables. References for you to read (not skim) -- the study in the PDF will probably interest you the most:

http://www.datacenterknowledge.com/archives/2011/03/10/energy-efficiency-guide-data-center-temperature/
http://www.geek.com/chips/googles-most-efficient-data-center-runs-at-95-degrees-1478473/
http://blog.schneider-electric.com/datacenter/2013/05/06/getting-comfortable-with-elevated-data-center-temperatures/
http://www.cs.toronto.edu/~nosayba/temperature_cam.pdf (PDF)
http://www.dummies.com/how-to/content/data-center-temperature-and-humidity-range-recomme.html

TL;DR -- 80F is not "quite warm" for a datacentre designed/built within the past 10-11 years.

Ugh, they aren't even in the same ballpark. Microsoft use their cloud server infrastructure for all sorts of things.

THIS is the difference between a shitty cheapskate gaming company that mistreats their employees and a legitimate enterprise vendor.

EA provisioned literally *dozens* of servers for Simcity 5. MSFT is provisioning 300,000 servers out of countless servers operating the Azure network.

While I agree with you on many points.

It is possible serverrooms are going to look very different in the coming years:

http://www.datacenterknowledge.com/archives/2013/01/22/silicon-photonics-the-data-center-at-light-speed/
http://www.wired.com/gadgetlab/2010/07/silicon-photonics-50-gbps/
http://www.opencompute.org/ocp-summit-iv-videos/

And thats pretty worthy, taking into account that 9.999999% of its uptime, hey, is like 7 9s!

Maybe they cut a new back door while the guard was watching the front one.

I don't care if it's onerous. And I don't really care why TWC decided not to peer with Netflix. None of that should affect how Netflix interacts with me as a customer. All I should have to care about is "is my pipe wide enough to receive the stream". If it is, Netflix should send me the stream. If not, Netflix should not send me the stream. The rest of it is none of my concern.

Netflix is the one changing the game, here. They are really the first B2C entity telling customers that they will treat them differently based on the topological structure of the internet between the customer and the business. That's really unprecedented.

What if Apple said "we're not going to deliver large apps to customers on networks that don't peer with us" or "we're going to charge an extra _% to deliver large apps to customers on networks that don't peer with us". People would go BALLISTIC. So why is it OK for Netflix to do that?

Would you rather have them turn it on and charge everyone more, probably loosing customers in the process? Maybe they should just charge TWC customers? As much as you don't like their decision, it is not some arbitrary position. Netflix is barely holding on. Their net income went from 226 million in 2011 to 17 million in 2012 in the face of skyrocketing content licensing costs. Compare that to TWC with a net income of 1.7 billion for 2012. Netflix could turn on Super HD for everyone, but they would have to raise prices and would end up loosing customers. Or, they could just charge extra to people on ISPs who won't or can't peer with them. Which would be better? Or would you rather have Netflix just go out of business? No business, not even an Internet based business, has a moral obligation to provide a service at a loss.

http://www.google.com/finance?q=NASDAQ%3ANFLX&fstype=ii&ei=9KAGUcDzGPCy0QGc7QE
http://www.google.com/finance?q=NYSE%3ATWC&fstype=ii&ei=caAGUejWObGh0AGoRw

Treating customers differently based on Internet geography is not unprecedented. I'd bet that it happens a lot but just isn't publicized. Comcast already started doing this. First, Comcast had their spat with Level 3 where they wanted L3 to pay Comcast for delivering data that Comcast subscribers were downloading. Then Comcast followed up by exempting their in-house video streaming service from the data caps. According to surveys, 64% of US broadband customers are under a data cap. Time Warner Cable is not currently under a cap, but that is because the customer base went into open revolt when they tried it in Texas. Netflix is not changing the game, it is responding to conditions on the field.

http://www.datacenterknowledge.com/archives/2010/11/29/level-3-vs-comcast-more-than-a-peering-spat/
http://techland.time.com/2012/04/16/netflix-ceo-takes-swing-at-comcast-xfinity-over-net-neutrality/
http://gigaom.com/2012/10/01/data-caps-chart/

Oh, and the reason Apple could not get away with a similar scheme is because Apple is the 2nd most valuable company in the world. Size makes a difference, and Netflix really is not a very big company.

Oh noes! you need nearly a whole rack of storage for a $237,000,000 film that raked in over 2 billion in box office sales.

That's a drop in the water compared to the rendering farm at Weta Digital that processes all that raw data, which was 34 racks in 2009.

The final render of Avatar was 2.8TB

http://www.datacenterknowledge.com/archives/2009/12/22/the-data-crunching-powerhouse-behind-avatar/

Weta Digital has over 3PB of storage
http://techday.com/netguide/news/the-tech-behind-weta-digitals-success/22643/

If the owners of the building could run cooler I would think they would.

Have a look here for more background.

Basically, they're describing four types of data centers. Have you seen the Google data centers with their heat curtains and all that? I surely don't work in any of those types of data centers. Some of the fancier ones around here have hot/cold aisles, but the majority are just machines in racks, sometimes with sides, stuck in a room with A/C. Fortunately it's more split systems than window units these days!

The conventional wisdom was that AC is cheaper than downtime/hardware so they told the building owner what to run the temperature at and they paid for it. Some of those assumptions are now being challenged.

I do dig energy effecient IT - I focus on this whenever I spec gear - but many people just 'go big', 'go cheap', or 'go IBM' (for various values of 'IBM'). Focusing on operating heat is an after-the-fact approach if you have opportunity to cut down on heat (freebie: do you put SSD's in front of your big drives to keep them cooler?)

With that said, there's one very good reason to run a cold room: power failures. I typically see places with decent to nice UPS units, but the A/C units are almost never on battery backup, and generators are too rare (even when they're there, they're rarely sized for or connected to the A/C). A data room can get hot in a hurry without A/C and if you're running at 65, you get to 95 much less slowly than you do when you're running at 82. Yeah, if you're a government contractor you just buy a CAT diesel and go about your day, but for many businesses the monthly cost of A/C is weighed against the purchase of the generator to make it able to sustain those kinds of conditions.

It looks like you may be right. At least that is what GoDaddy is saying now. http://www.datacenterknowledge.com/archives/2012/09/11/go-daddy-network-issues-not-hackers-or-ddos-caused-outage/

Well, you say that, but they may have an opportunity now: http://www.datacenterknowledge.com/archives/2012/09/04/intel-explores-mineral-oil-cooling/

As of 10/11, Phoenix had 830K, with plans to add 160K and maybe even a million more feet later. source.

However, just one building in North Carolina has 500K of space (source), and Apple is by no means alone out there, with at least Google and others. Central Oregon and Washington are also big; Facebook has 300K in Prineville and another 300K on the way. Apple is also building in central Oregon, Google has a large facility, and yahoo and microsoft have large facilities in Central Washington.

Phoenix is a player, but by no means has "as much DC capacity as the rest of the US combined." They may have more colo space, and more individual 100k+ size units than elsewhere, to but consider all of the domestic DC capacity you are including self-builds in that statement.

Indeed. I'm sure that once they factor this into their ever-so-accurate estimates, they'll have numbers indicating this data center consumes more energy than the state of Texas.

If you actually care about the environment more than you hate Apple, you'd realize that Apple is more green in terms of how it makes (AND recycles) its products than any other major electronics manufacturer. Environmental groups just like to eviscerate Apple for PR, even though it's one of the most transparent and aggressive on protecting the environment and green tech.

The funniest thing? In a few years we'll see every other vendor following Apple's lead, as they always do.

Apple helped create the EPEAT standards alongside the other stakeholders who helped define it.

Apple even has a contract to recycle products from ANY manufacturer, for free, with free shipping fees and boxes provided. What other vendor does this? Who puts their money where their mouth is on the environment?

Apple's products, in real, practical terms, are MORE recyclable, in terms of recyclable content contained therein, and the ability to actually recycle them — albeit by using Apple's programs for things like iPhone, iPad, and now the Retina MacBook Pro — and that many other EPEAT-certified products may be (and are) markedly worse than Apple's products in this sense, but can still be certified because they are able to be disassembled with conventional tools. How does that make them "more green"?

EPEAT alone isn't the end-all, be-all of green certifications. Organizations use EPEAT because it is a metric; a box that is easy to check; an easy way to define the "greenness" of a product. Apple helped develop the EPEAT standard, and has been one of the most committed and transparent manufacturers to green tech, environment, and recycling. No other major vendor has this level of transparency.

And Apple is STILL targeted by folks like Greenpeace, even as Apple is pursuing green more aggressively than its competitors, with Data Center Knowledge noting:

Greenpeace’s continuing use of this methodology, in light of Apple’s disclosure and permit data, raises several possibilities:

- Greenpeace is having difficulty developing estimates that accurately incorporate data center operations and power usage.
- Greenpeace is predisposed to cling to estimates that make Apple look less “green” because it generates more headlines for its awareness campaigns.

EPEAT didn't cave on anything — but the next generation of EPEAT would do well to consider the real, end-to-end recyclability and carbon footprint of electronic products.

As I also work (as a contractor) for a agency (that I've mentioned before, if you really care) ...

I see it as being two things:

1. Lobbying. Just like the 'one card' fiasco (what, we issue cards for 5 years even if the person's got 2 years left on their contract? Because it's such a pain & expense to issue a card? Oh, and it can get them into other installations that they have no business going to?), it's all a matter of lobbyists selling 'solutions' to minor problems that end up having major repercussions that the high up decision makers don't seem to care about.
2. There's been a requirement for the past few years for agencies to reduce the number of 'data centers'. Unfortunately, they keep changing the definition of 'data center' until I think at this point a telco closet qualifies. (and unfortunately, I'm not exaggerating). So, the easiest reduction -- get rid of all of your servers, implications be damned.

And of course, we're getting all sorts of clouds now -- public cloud, public companies with specialized government clouds, agency clouds, departmental clouds and clusters, etc. And for the departmental one they're trying to get us to move over to, they can't tell us what the cost model is going to be. (other than we know it'll be cost / year, while our budget is built around infrequent hardware replacement costs + lower reoccurring sysadmin time)

This is all part of Facebook's new strategy.

Facebook will be building a huge new data center in northern Sweden to support the rapid global growth of its users. The new data center in Lulea, Sweden will be Facebook’s first facility outside the United States.

“It’s the next step in our ongoing strategy of building our own infrastructure and moving away from leased facilities,” said Facebook spokesman Michael Kirkland. “We are expecting this data center to continue to help us reduce latency for our users in Europe and beyond.”

http://www.datacenterknowledge.com/archives/2011/10/27/facebook-goes-global-with-data-center-in-sweden/

Has anyone ever tested if we actually need air conditioning for a server room?

You do need "air conditioning", since you do want to make sure the air is not too dirty or humid or dry or hot.
But yes you can do without conventional data center air conditioning:
http://www.facebook.com/note.php?note_id=10150148003778920&_fb_noscript=1
http://www.datacenterknowledge.com/the-facebook-data-center-faq-newest-page/

They're also trying in a warmer more humid area:
http://www.wired.com/wiredenterprise/2012/04/facebook-data-center-2/
Wonder how well that will work.

"cares about it's users"? Are you kidding me? Do you know how much it costs to run facebook? For bandwidth, servers, electricity, etc for 900 million users?

$600 million for equipment in 2011, and another $500 million this year. (source)

That's just for equipment. Plus you have to pay for developers, server admins, office space, etc, etc, etc.

Anyone who thinks they are going to start a service to replace facebook without making money their #1 priority is an idiot who will fail the moment they have to open a hundred million $ data center.

The effort to gain acceptance for DC distribution in data centers is being helped by a series of investments by ABB, and the growth of the EMerge Alliance, which is trying to unify DC proponents around a 380V standard. The challenge for DC is that customers don't ask for it, meaning multi-tenant facilities aren't likely to offer it. Also, Schneider says it is "not aware of any data centers moving off of their established, traditional power distribution to DC." In fact, NTT has at least five DC data centers in Japan, and ABB is backing a DC distribution project at a Swiss hosting company. In the US, there are numerous sites testing DC power, which is widely used in telecom infrastructure.

The effort to gain acceptance for DC distribution in data centers is being helped by a series of investments by ABB, and the growth of the EMerge Alliance, which is trying to unify DC proponents around a 380V standard. The challenge for DC is that customers don't ask for it, meaning multi-tenant facilities aren't likely to offer it. Also, Schneider says it is "not aware of any data centers moving off of their established, traditional power distribution to DC." In fact, NTT has at least five DC data centers in Japan, and ABB is backing a DC distribution project at a Swiss hosting company. In the US, there are numerous sites testing DC power, which is widely used in telecom infrastructure.

The effort to gain acceptance for DC distribution in data centers is being helped by a series of investments by ABB, and the growth of the EMerge Alliance, which is trying to unify DC proponents around a 380V standard. The challenge for DC is that customers don't ask for it, meaning multi-tenant facilities aren't likely to offer it. Also, Schneider says it is "not aware of any data centers moving off of their established, traditional power distribution to DC." In fact, NTT has at least five DC data centers in Japan, and ABB is backing a DC distribution project at a Swiss hosting company. In the US, there are numerous sites testing DC power, which is widely used in telecom infrastructure.