Slashdot Mirror


Google's Chiller-Less Data Center

1sockchuck writes "Google has begun operating a data center in Belgium that has no chillers to support its cooling systems, which will improve energy efficiency but make weather forecasting a larger factor in its network management. With power use climbing, many data centers are using free cooling to reduce their reliance on power-hungry chillers. By foregoing chillers entirely, Google will need to reroute workloads if the weather in Belgium gets too warm. The facility also has its own water treatment plant so it doesn't need to use potable water from a local utility."

15 of 132 comments (clear)

  1. What's next? by ickleberry · · Score: 5, Funny

    If it wasn't for the required internet connectivity google could go off the grid completely. But they already own so much fibre and the public internet seems to need google more than they need it.

    Soon they will generate all their own power from wind and solar, convert all their employees shit to power so they don't need the sewerage system either, send all their traffic through the network of low earth orbit satellites they are about to launch which also conveniently beam solar power back down to them.

    So basically at the end of the day they will be able to buy or swindle a plot of land from some country with low tax, bring in all their own employees, contribute absolutely nothing to the local economy and leave when the sun goes down. It's great really, saves them on lawyers that would otherwise help them pussyfoot through the swaths of modern over-regulation and the satellites will help them get past any censorship / connectivity problems.

    And if China start shooting down their satellites, Google will make satellites that shoot back

    1. Re:What's next? by Anonymous Coward · · Score: 5, Funny

      I am totally buying Google stock if they do this.

  2. Unreliable... by Darkness404 · · Score: 4, Interesting

    So basically everything gets rerouted on a hot day. Ok, that sounds fine until you realize that most of the outages of Google's products were due to, rerouting. And also, it seems odd that the cost of building a (hopefully redundant) datacenter that is this unreliable would be less than consolidating it with another one and using electrical cooling.

    --
    Taxation is legalized theft, no more, no less.
    1. Re:Unreliable... by martas · · Score: 5, Interesting

      well, it might be unreliable, but i think you're overestimating the reliability of normal data centers. even if failure is twice as likely at this data center than others, i think it still improves overall performance and reliability enough that it's worth building. or at least google seems to think so.

    2. Re:Unreliable... by j79zlr · · Score: 5, Informative

      If you have chilled water, you have a chiller, which means you have compressors. Process water or ground source water usually is not cold enough to be an effective cooling medium. You want a high delta T between the entering air temp and the entering water temp to induce heat transfer. Closed loop ground source water is extremely (prohibitively) expensive and open loop is quite a maintenance hassle due to water treatment. High efficiency chillers paired with evaporative cooled water towers with economizer capability is very efficient and reliable. Usually you can get down to around 0.5kW per ton with high efficiency chillers at full load and with multiple staged compressors you can do even better with part load conditions. The cooling towers are usually pretty low with around 0.05 to 0.15kW per ton. Use VFD's on the secondary pumps and cooling tower fans, and you can get cooling in at 0.75kW per ton for the whole plant at peak and even lower and part load conditions (95% of the time).

      I just designed a data center for a large Big Ten univeristy and there were no large air handlers involved at all. The system had two 400-ton chillers with the chilled water piped directly to rack mount APC fan coils. Without "green" being the basis of design, the chiller system still operates right at about 1kW/ton.

      --
      I'm not not licking toads.
    3. Re:Unreliable... by jhw539 · · Score: 5, Interesting

      You do not need a chiller to operate a datacenter in many environments at all. Based on the 2nd edition of ASHRAE's Thermal Guidelines for Data Processing Environments (which was developed with direct input from the major server providers), you can run a datacenter at up to 90F. Seriously, 90F into the rack. When it comes out the back of the rack, you collect the heat exhaust at 100-110F. "Chilled" water at 81F is more than enough to knock that 110F down to 90F- ready to go back into the front of the rack.

      The 81F water can be produced directly from open cooling towers (direct evaporation) whenever the wetbulb is lower than 76F (4 degree approach plus a 1F on your flat plate that isolates the datacenter loop from the open tower loop).

      You designed an efficient datacenter, but you're five years behind cutting edge (not actually a bad thing for most critical environment clients). The next wave of datacenters will have PUEs of 1.2 or less and redefine the space from a noisy but cool space to hang out to a hot machine room with industrial heat exhaust design.

      I actually just finished a chiller less 8MW schematic design and analysis for a bid. It was my second this month (the first was a cake walk - an extreme Twb of 67F, the second was west coast light conditions).

      PS: Secondary pumps? Seriously? Unless you have to boost up to 25 psi to feed a Cray or some other HPC I thought everyone who cared had moved onto variable primary-only pumping. (Sorry, feeling a bit snarky after hitting a 40 hour week on Weds...)

    4. Re:Unreliable... by j79zlr · · Score: 4, Informative

      The units were mounted on the roof, but were packaged AAON 2 x LL210 chillers (and a full 400 ton backup) with no exposed exterior piping. Glycol reduces the specific heat of the fluid and increases the specific gravity, so it can move less heat and takes more power to move. I only add glycol to the system if freezing is an issue.

      --
      I'm not not licking toads.
  3. Worth the tradeoff? by Clockowl · · Score: 4, Interesting

    Is it really worth to be dependent on the weather in exchange for a lower energy bill?

    1. Re:Worth the tradeoff? by symbolset · · Score: 4, Insightful

      If you're Google? Apparently the answer is "yes."

      More people can and should do this. 27C is plenty cool enough for servers. It annoys me to go into a nipple crinkling datacenter knowing they're burning more juice cooling the darned thing than they are crunching the numbers. A simple exhaust fan and some air filters would be fine almost all of the time, and would be less prone to failure.

      --
      Help stamp out iliturcy.
  4. Reroute the weather instead by basementman · · Score: 4, Funny

    Why not just reroute the weather? Once google gets into cloud seeding and all that they really will be SkyNet.

  5. Re:Unreliable...Probably not by Banzai042 · · Score: 5, Insightful

    Remember that even on hot days not all of the traffic through the datacenter needs to be rerouted, and I'd imagine that a location selected for a datacenter like this was chosen for the infrequency of days that will require rerouting. Do you know how much it costs to cool a datacenter, and how much this will save? I don't, but Google probably does, and they probably wouldn't make a decision to do something like this without comparing the savings with the potential cost from decreased lifespan of computers running hot and losses due to downtime. I would also imagine that Google will be working to greatly increase stability during rerouting, given the comments from the end of TFA about other power saving uses, such as routing traffic to datacenters where it's night, meaning "free cooling" can be used since it's colder outside, and off-peak electricity rates are in effect.

    I think the concept is interesting, and it makes me wonder if we'll see more datacenters built in areas of the world more conducive to projects like this in the future.

  6. Yakhchal by physicsphairy · · Score: 4, Informative

    The ancient Persians had a passively cooled refrigerator called the yakhchal which "often contained a system of windcatchers that could easily bring temperatures inside the space down to frigid levels in summer days."

    Perhaps the Google datacenter could employ some variation of their technique.

  7. Re:This might be a dumb question by david.given · · Score: 4, Interesting

    The short answer is yes --- water takes a staggering amount of energy to change temperature (it's one of the many properties the stuff's got that's really weird). A big lake makes an ideal dumping ground for waste heat. What's more, the environmental impact is going to be minimal: even the biggest data centre isn't going to produce enough waste energy to have much effect.

    (A big data center consumes about 5MW of power. The specific heat capacity of water is about 4kJ/kg.K, which means that it takes 4kJ to raise the temperature on one kilogram of water by one kelvin. Assuming all that gets dumped into the lake as heat, that means you're raising the temperature of about 1000 litres per second by one kelvin. A small lake, say 1km x 1km x 10m, contains 10000000000 litres! So you're going to need to run your data centre for ten million seconds, or about 110 days, to raise the temperature by one measly degree. And that's ignoring the cooling off the surface, which would vastly overpower any amount of heat you could put into it.)

    (The same applies in reverse. You can extract practically unlimited amounts of heat from water. Got running water in your property? Go look into heat pumps.)

    In fact, if you were dumping waste heat into a lake, it would make sense to try and concentrate the heat to produce hotspots. You would then use this for things like fish farming. Warm water's always useful.

  8. Re:Unreliable...Probably not by j79zlr · · Score: 4, Interesting

    That is where the ice storage systems become interesting and cost effective. In the states, usually half of a commercial energy bill is peak demand. If you can transfer that energy usage to night time to build up your ice storage and transfer your main power draw to off peak the savings can be very significant and create payback times in months not years.

    --
    I'm not not licking toads.