Forgot your password?
typodupeerror
The Internet Earth

Nuke Site Converted Into Green Data Center 125

Posted by samzenpus
from the glowing-data dept.
1sockchuck writes "If you had 100,000 servers, would you put them on top of a former nuclear fuel facility? One of the world's largest web hosts, 1&1 Internet, is building a new data center on a site in Hanau, Germany previously used by Siemens to produce mixed oxide rods made from enriched uranium and plutonium. The site has been cleaned up, and 1&1 is converting it into a 'green' data center powered by renewable energy and using free cooling to save on air conditioning costs."
This discussion has been archived. No new comments can be posted.

Nuke Site Converted Into Green Data Center

Comments Filter:
  • sssss (Score:4, Funny)

    by Anonymous Coward on Thursday November 13, 2008 @09:00AM (#25745471)
    I cant hear the name Siemens without giggling
  • 1&1 (Score:5, Insightful)

    by h4rm0ny (722443) on Thursday November 13, 2008 @09:03AM (#25745491) Journal

    1&1? They should worry more about where they site their customer service! I was with them for a while and when they screwed up my billing it took a long, long time to untangle the mess. Mainly because the different departments were all sited in different places and none had the authority to do what needed to be done to sort it out. 1&1 - hateful, money-grubbing company. Thank you, rant over. I will now pay the karma hit with pleasure. :)
    • Re: (Score:2, Insightful)

      by andy19 (1250844)

      Mainly because the different departments were all sited in different places and none had the authority to do what needed to be done to sort it out.

      I always thought this was standard among all customer service departments.

    • 1&1? They should worry more about where they site their customer service!

      They outsource their customer service to the Philippines if I remember correctly. If you want to speak to someone who knows English, call their sales line. I interviewed for them and I can say I wouldn't want anything to do with them.

    • I have had absolutely no trouble with 1&1, they've been my host for about 4 years. If you're dealing with a customer service rep in any industry that doesn't have the authority to do what needs to be done, then you have to ask to speak to someone who does. It's a shame you're missing out on a great service provider because you didn't make a painfully obvious request.
      • Yeah, well... and what if they have no authority to connect you to someone who has the authority?
        That's why I hate the Postbank in Germany. They do not even have phone numbers for their branches.

        Oh, and the fact that they forged my signature (No joke! I still have proof.) does not make it any better...

      • I have had absolutely no trouble with 1&1, they've been my host for about 4 years. If you're dealing with a customer service rep in any industry that doesn't have the authority to do what needs to be done, then you have to ask to speak to someone who does. It's a shame you're missing out on a great service provider because you didn't make a painfully obvious request.

        Been there, done that. There is nobody to talk to. Really. I wound up finally leaving a message on the cell phone of some guy in Pennsyl

        • by h4rm0ny (722443)

          My problem was different - it was a less serious cock up on the domain side of things - but the consequences were similar: Unwanted billing of a credit card and, when I cancelled that, repeated attempts, referral to debt collection agencies who threatened me with various consequences if I didn't pay 1&1's incorrect charges. I bounced between two departments, neither of which had the power to change things by themselves and no apparent desire to. It dragged on for a long time and wasted a lot of my time
  • Green power (Score:5, Informative)

    by askanis42 (1138835) on Thursday November 13, 2008 @09:04AM (#25745501)
    1&1 is also using "green" power generated from wind, water and solar energy for their datacenters and office buildings. see: (German only) http://www.1und1.info/xml/order/popupGruenerStrom [1und1.info]
  • Cleaned up? (Score:5, Interesting)

    by jDeepbeep (913892) on Thursday November 13, 2008 @09:05AM (#25745505)

    The site has been cleaned up

    Oddly enough, TFA says nothing about the site being cleaned up.

  • ... and convert a Green Data Center into a Nuke Site.

    THAT would be news.

  • by dkf (304284) <donal.k.fellows@manchester.ac.uk> on Thursday November 13, 2008 @09:09AM (#25745533) Homepage

    When they said "Green Power", did they prefix it with "Glowing"?

  • Does anyone else get the feeling that the summary wants us to react in a certain way?
    Would you put your servers on the NUKELEURZ? WOULD YOU!?

    I'm not feeling the fear here.

    • Better an old nuke site than an old Pet Semetary [imdb.com]
    • by jopsen (885607)

      Does anyone else get the feeling that the summary wants us to react in a certain way? Would you put your servers on the NUKELEURZ? WOULD YOU!?

      Yeah...
      I wonder if there's more radiation than normal?

      Does anybody know if could be bad for the servers, chips etc...
      I'm sure they wont die immediately, but I wouldn't be surprised if radiation is bad for todays frail processors...

      • I wonder if there's more radiation than normal?

        Does anybody know if could be bad for the servers, chips etc...

        I'm sure they wont die immediately, but I wouldn't be surprised if radiation is bad for todays frail processors...

        They made nuclear fuel rods here. Uranium and Plutonium. Both alpha emitters. Alpha particles can be stopped by a sheet of notebook paper. Or the paint on the walls of your datacenter. Or the tiles on the floor of your datacenter. Or even the cases that your servers are in.

        So,

  • by Anonymous Coward
    It's a data center.
    It's a former nuke producing facility.
    It's green.

    Is there anything to see here?

  • by Rogerborg (306625) on Thursday November 13, 2008 @09:17AM (#25745589) Homepage
    And what's their plan to deal with the Deep Crows [penny-arcade.com]?
  • by splutty (43475) on Thursday November 13, 2008 @09:18AM (#25745603)

    I've read quite a bit about this whole idea of free cooling, and as far as I've been able to conclude, the basic premise is that the replacement cost for failures very much outweighs the costs for cooling it properly.

    If you realize that the last decade or so, most components can easilly be overclocked with proper cooling, and will function quite well in a wide range of temperatures, it's not hard to imagine that operating temperatures of anywhere between -10C and +40C are generally fine for most equipment.

    The only thing that would be affected, in the sense of less cleaning of air, would be movable parts components, like harddisks, fans, etc.

    With the prices on HDDs and the ease of use and availability of any sort of RAID configuration you can think of, the actual costs for replacing these parts when they fail, could very well be a fraction of the costs that would be required to make them function 'properly'.

    All in all it seems an economically very viable option, with the added advantage of using a lot less energy overall.

    • Re: (Score:2, Informative)

      by tomhudson (43916)

      With the prices on HDDs and the ease of use and availability of any sort of RAID configuration you can think of, the actual costs for replacing these parts when they fail, could very well be a fraction of the costs that would be required to make them function 'properly'.

      Even if the hard disks were FREE, the cost of replacing them, both in downtime, and in labor, and in higher risk of cascading failures (second drive fails when restoring a raid5, requiring a full restore from backups), are more than the co

      • Re: (Score:3, Insightful)

        by partenon (749418)

        Depends on your needs. If you have a big company, with tons of servers working in a distributed cluster, then one server can completely fail without having any hit on the performance of the services. And as failures are exceptional cases, those big companies prefer to have failures in some specific components/machines than to have to pay a far higher energy bill.

        But if you have the "traditional setup", with tons of machines, each of them responsible for a specific system/application, then of course: if anyt

      • Re: (Score:3, Insightful)

        by QuantumRiff (120817)

        Have you ever actually looked at the FULL cost of proper cooling? Not just the AC units, but the power draw, the labor to keep the parts running, and filters clean, the HUGE generators to keep these large AC systems running when the power goes out? More than half of most datacenters generators and UPS load is for cooling. You can buy a ton of hard drives for the cost of a 1MW diesel generator.

        • by tomhudson (43916)
          The cost of the hard drives is irrelevant. It's the data that's on them that counts. The bigger the drive in a raid, the more likely that the raid will fail during restoration, meaning you don't want to make it even less likely to be restorable because you're running the drives way too hot. But since you're running the drives way too hot, you need to build out extra redundancy, which also costs, both in capital costs, and in energy. TINSTAAFL.
          • TINSTAAFL.

            Whaaaaaaaaaaaaaat?

            • Re: (Score:3, Informative)

              by tomhudson (43916)

              TINSTAAFL.

              Whaaaaaaaaaaaaaat?

              This is slashdot - if you use TANSTAAFL, the grammar nazis will come after you about the double negative. If you don't recognize the Heinlein reference, and can't even google, it's time to turn in your geek card.

      • by Chyeld (713439)

        Even if the hard disks were FREE, the cost of replacing them, both in downtime, and in labor, and in higher risk of cascading failures (second drive fails when restoring a raid5, requiring a full restore from backups), are more than the cost of proper cooling.

        This is assuming that "proper cooling" actually extends the life of the drive significantly past that of "improper cooling". And truthfully, I can't say that observational experience or my limited reading on the subject backs such an assumption.

        Proper

        • by tomhudson (43916)
          ... or you might be stuck with those Chinese "Maxtorgates" that have a worse than 50% DOA, which solves the problem ...
      • by splutty (43475)

        You're not thinking on a corporate scale there.

        All reasonably big datacenters I've ever worked on have a combination of raids, with hotswappable spares, in most cases a RAID 1/0 solution.

        Also in terms of reliability, there are tons of ways to avoid double failures in drives. One of the easiest being to replace disks in a staggered fashion. Something that most major datacenters will do anyway.

        MTBF suddenly becomes a whole lot less important.

        I think you're overestimating the actual time/costs involved with re

        • by tomhudson (43916)

          As drives get bigger, the probability of a second failure gets larger as well. And we all know the rule - it's not a question of "if", but "when" a drive will fail.

          Then there's the environmental cost of disposing of those hard drives, which isn't ever mentioned in those "green" calculations - and the environmental cost of producing more drives to take into account the higher failure rates.

          Drives fail a lot more often than they used to, and its only when you try to recover files that you haven't touched

    • These guys think they are so smart, but if they hadn't cleaned up the site, they could have had free heating too.

  • by Critical Facilities (850111) * on Thursday November 13, 2008 @09:26AM (#25745655) Homepage
    This is marginally interesting, but light on specifics. I mean, the article claims that the new Data Center is going to use "renewable energy" to power it, however it doesn't explain what kind of renewable energy or how it's going to do so.

    Furthermore, while the air side economizer is a great idea (and more data centers should be using it), there is no description of what supplemental, mechanical cooling there will be in this facility. I can't honestly believe that there will never be a need for any cooling other than what mother nature is providing. Sure, geographically, it's bound to be cooler than say the southwest U.S. but there are still apt to be days in the summer where temperatures make it implausible to be on "economizer only".
  • Interference (Score:3, Interesting)

    by Eudial (590661) on Thursday November 13, 2008 @09:30AM (#25745691)

    1) Stray residual gamma rays knocks more electrons out of circuit A than circuit B.
    2) Resulting potential difference induces current.
    3) Resutling current flips a bit.
    4) Bit is saved on hard drive.
    5) Data is corrupted.
    7) ???
    8) (Absence of) Profit!

    • by ben0207 (845105)

      Gamma Rays? I'm more worried about the Hulk, myself

      "Server is down? HULK SMASH PUNY SERVER!"

    • HDDs dont need radiation to degrade, room temperature will suffice:

      1) Thermal instability of magnetic media leads to degrading of magnetic field. Results in loss of bit.
      2) This happens often enough that the ECC is overwhelmed.

      As capacity goes up, the energy per bit goes down. The MAJOR design challenge with HDD media is making it writable by a tiny head, yet not writable by random thermal induced fields at room temperature.
  • that is unexplained, i usually say something like "probably a stray cosmic ray"

    for the technically inclined, this usually elicits a laugh

    for the technically uninclined this usually elicits a stony face of seriousness

    try this comment sometime, its win win. its a good litmus test for the level of technical acumen you are dealing with in someone

    however, these guys can actually say this sort of thing with a straight face: "probably a stray gamma ray"

    • however, these guys can actually say this sort of thing with a straight face: "probably a stray gamma ray"

      Does the old routine of 'hmm' walk over to the right side of the building, look out the window, squinting hard for 15 seconds, and simply saying 'damn sunspots' not work anymore?

    • by Saffaya (702234)

      I can remember reading Dave Small (famous hardware hacker and entrepreneur, he made the MacIntosh emulator cartridge on the ATARI ST back in the days, also some 68030 accelerator cards) describing how he saw a character on his screen change in front of his eyes with no intervention, and attributed to a cosmic ray and his higher than normal altitude.

      So this begs the question, although modern servers do have ECC memory to correct such occurences, couldn't there be a weaker link in the server chain somewhere t

      • I used to get the cosmic ray answer from Motorola when one of our former systems would suffer a double-bit parity error and go TU. The first time the support weenie said it, it was mildly amusing, but by the third time they replaced all the memory, including swapping it with a system that nevercrashed, I figured out that it was their code phrase for "I don't know, I don't care, and I am going home." I have not had a Motorola system for eight years, and my cosmic ray problem left when they did. To me, that p
    • Cosmic rays? I always figure it's sun spots.

    • Re: (Score:3, Interesting)

      by aphyr (1130531)

      Actually, cosmic rays can and do cause errors. Muon flux where I live tends to be roughly one through your hand per second, and they're going a pretty hefty fraction of C. With memory size and transistors scaling further and further down, cosmic ray interference becomes a really big issue, which is why ECC is so important.

      http://ieeexplore.ieee.org/Xplore/login.jsp?url=/iel1/16/6912/00278509.pdf?temp=x [ieee.org]

      We're dealing with more delicate technology these days; It's only gotten worse since then.

  • by Ummite (195748) on Thursday November 13, 2008 @09:34AM (#25745721) Homepage
    There is a ton of places, like any northern places in Canada, where electricity is cheap and is really cool nearly all year long. I could think about Quebec province, in Canada. Electricity is approximatly 5 cents (canadian) per kw.h (like 4 cents US$) and it would cost nothing to cool down as much server as you want. Maybe some company already have such datacenters, but I could think about some google / microsoft datacenter going to canada, to save on electricity bills and cooling.
    • One Word: (Score:4, Informative)

      by oncehour (744756) on Thursday November 13, 2008 @09:45AM (#25745821)
      Taxes. Canadian Business Taxes are really pretty bad. Don't think that Microsoft and Google haven't already crunched the numbers. In all likelihood the cooling and electricity savings are outweighed by increased regulation and taxation.
    • Of course you can always go to the North/South pole and you're right, the cooling costs would dramatically decrease. However further you are away from civilization, the bigger are your difficulties to have enough electricity AND data connections (ok, Canada would probably be fine, however the problem remains in principle). This is imho the main drawback of Google's Off-shore Data Center [slashdot.org] and similar proposals.

      • by rHBa (976986)
        The inaccessibility would also add costs in terms of:
        1. Constructing the facility
        2. Supplying the facility (hardware, people, food etc)
        3. Paying employees enough to make them want to work in the coldest and most boring place in the world with little/no daylight for 12 months a year.
        • If you're smart about it, the facility would be unmanned, similar to LEGO's manufacturing facility.

          • by rHBa (976986)
            You'd still have to re-supply the facility with replacement/upgrade parts (HDDs, power supplies etc) and someone would have to fit them. I suppose you could have an automated 'production line' to service the machines but that's just another expense!
      • However further you are away from civilization, the bigger are your difficulties to have enough electricity ... ok, Canada would probably be fine

        More than fine. Quebec Hydro [wikipedia.org]. Data connections would be more of a problem, but laying fiber isn't that expensive. Taxes, on the other hand...

        • You still have to fight the speed of light. Google did a study showing that every second an end user waits for results, the chance of them going elsewhere rises quickly.

  • They can save energy by not having to turn the lights on.

    Everything has it's own "natural" glow.

    Hot water for the staff won't be a problem either.

  • by slashmojo (818930) on Thursday November 13, 2008 @10:00AM (#25745947)

    So no need for Ready Brek [wikipedia.org] to make the sysadmins "Get up and Glow"

  • I used to visit Hanau on business. I don't know whether it's changed, but it used to be full of nuclear engineers, metallurgists, and scientists working on some interesting technologies. In the (spotlessly clean) town centre (rebuilt completely after WW2) is a memorial to the Brothers Grimm, the philologists who collected the fairy tales. Hansel und Gretel are famous for stuffing the witch into her own oven, and one company in Hanau used to make extremely high temperature furnaces, but that's about the only
  • I find it funny (Score:3, Interesting)

    by Gates82 (706573) on Thursday November 13, 2008 @10:58AM (#25746663)
    As a graduating civil engineering student I find it funny the way people think about places or items which were formerly contaminated and now clean. The summary is a classic example of this mentality. Their building a data center on former nuclear facility site that has been cleaned. So what is the news?

    I would get this same reaction in my environmental engineering class concerning waste water treatment (gray to white not sewage to gray). Even though the engineering of the treatment plant was explained most of the students would not be willing to drink the water that came out of the facility even though it used RO or other methods that are used to purify water from natural sources. This makes absolutely no sense. Engineers who understand that all water is recycled anyway, and that there is no difference if it is done mechanically vs. naturally.

    If as educated individuals we cannot sell ourselves on the safety of the procedures how do we ever expect the uneducated masses to accept them?

    --
    So who is hotter? Ali or Ali's Sister?

    • If as educated individuals we cannot sell ourselves on the safety of the procedures how do we ever expect the uneducated masses to accept them?

      Kind of like that time we almost destroyed the ozone layer, right?

      I'll err on the side of caution until I'm absolutely sure, thank you very much.

      (But seriously, the mechanics of the CFC-ozone reaction are downright scary. Given a few more years, we could have done some serious irreversible harm)

      • Re: (Score:3, Interesting)

        by Gates82 (706573)

        Given a few more years, we could have done some serious irreversible harm

        Seeing how ozone is produced by a reaction with UV, the more UV that is allowed to pass the ozone layer causes an increase in natural ozone production. So in essence this becomes a self balancing system (as most of the earth's systems). So the hole would come to a natural equilibrium.

        I'm more concerned with the irrational fear associated with such things as the hole in the ozone. You think that it is gone? We have not heard much about it in the last 10 years cause it was of little concern; also, ozo

        • We have not heard much about it in the last 10 years cause it was of little concern

          It was in the news recently - the ozone hole this past year was the fifth largest ever. Of the 30 years we've been sampling the silly thing.

        • by vadim_t (324782)

          I find it great human arrogance to assume that we can to any large degree upset the natural balance of the earth to a level that it will not recover itself.

          Where does this "arrogance" nonsense come from? It would have been true in say, the middle ages, but today we're more than capable of something like this. Currently we can go against pretty much any natural process.

          Golf course in the middle of a desert? Been done. Building a city under the water table? Yep, New Orleans, and the only problem was incompete

        • by Draknor (745036)

          But to run tests on water and find that it is cleaner then "normal" municipal sources and not trust it is absurd.

          That's the catch though, isn't it - How good are your tests? I could tell you that your converted water looks clear & doesn't smell funny, so it must be clean, right? That might not pass muster today, but it probably did 200 years ago. And 200 years from now, people will be aghast at the daily things we do/consume that we think aren't so dangerous.

          I do agree that the only real risk we face

  • No matter how well they clean it up I'm guessing that there are more alpha and beta particles flying around there than on some random previously empty piece of land. And with chip geometries smaller than ever this might be an issue.
    • Radioactivity 101 (Score:4, Interesting)

      by Kupfernigk (1190345) on Thursday November 13, 2008 @12:59PM (#25748407)
      Erm...alpha particles (helium nuclei) are stopped by paper or air. Beta particles are stopped by quite thin metal foil. I think you mean gammas, and I suspect that these will be much lower than the background radiation (read, cosmic rays.)

      I recall that back in the old days when expensive ICs were packaged in ceramic and cheap ones in plastic, cheap memory was less prone to bit errors because some of the ceramics contained, as it turned out, significant amounts of radioactivity. Potassium, for instance, is noticeably radioactive in its natural state (one of its isotopes is unstable).

      Given that the concrete won't be made from raw materials collected on site, nor will the aluminum and steel in the server racks, and that the only really common beta emitter (tritium) produces electrons with less energy than those in an old style CRT, your fears are groundless.

    • Re: (Score:2, Informative)

      by Soruk (225361)

      Alpha particles [wikipedia.org] are pretty large entities, being helium nuclei (two protons and two neutrons), as a result can only travel a few centimetres through air so any machine's case will stop them completely.

      Beta particles [wikipedia.org] are electrons or positrons) and can reach about 9 metres [harvard.edu] through air but less than 5mm through aluminium.

  • by greg_barton (5551) <greg_barton@NOSpaM.yahoo.com> on Thursday November 13, 2008 @12:55PM (#25748321) Homepage Journal

    ...then it was already green. [blogspot.com]

  • > If you had 100,000 servers, would you put them on top of a former nuclear fuel facility?

    If you had 100,000 servers, would you put them on top of a former toaster factory?

"How do I love thee? My accumulator overflows."

Working...