Slashdot Mirror


Data Centers Breathe Easier With Less Oxygen

PC World is reporting that some companies are looking at a new method of fire protection in their server closets, oxygen-deprivation systems.""Wood stops burning when the oxygen content falls to 17 percent and plastic cables between 16 to 17 percent, said Frank Eickhorn, product manager for fire detection at Wagner Alarm and Security Systems GmbH in Hanover, Germany. Wagner makes electric compressors that use a special membrane to remove some of the oxygen from the outside air, a system the company calls OxyReduct. The excess oxygen is exhausted, and the remaining nitrogen-rich air is pumped inside the data center."

3 of 392 comments (clear)

  1. Maybe I am stupid but . . . by vecctor · · Score: 0, Redundant

    From TFA:

    At 15 percent oxygen, it's safe for humans to enter. The lower oxygen content of the air is similar to being at an altitude of about 6,000 feet, Eickhorn said. He demonstrated with a lighter inside a sealed atrium Wagner has on display at Cebit. It won't light.

    6,000 feet? People regularly burn wood above 6,000 feet don't they? Basically, someone in a mountain state with a wood stove, or camping. People in the Andes are probably burning wood even higher than that.

    Does someone know how this is supposed to work? Am I missing something?

    --
    Why, yes I have been touched by His noodly appendage. And I plan to sue.
  2. I call bullshit by Secrity · · Score: 0, Redundant

    Does this mean that things don't burn above 6,000 altitude? I guess that I just imagined having camp fires above 8,000 feet in the Rockies. I saw the remains of a wooden building at over 14,000 feet that had burned to the ground. Something doesn't smell right with this article.

  3. I disagree with article by fred+fleenblat · · Score: 0, Redundant

    I think it is wrong because:

    1. The most critical element to a fire isn't just the ratio of oxygen in the air but the *flow* of air to the flame. And in a server room EVERY BOX HAS A FAN FEEDING AIR TO EVERY COMPONENT. So long as the air is exhasuted from the cabinet as fast as the fire uses it up, it will keep burning until the fuel is used up.

    2. I don't buy the 6000 ft thing. There is an 8000 ft tall mountain range nearby that catches fire every 8-10 years and darned if it doesn't just keep burning right on uphill, not stopping in the least at 6000 ft.

    3. I personally have seen open campfires, from plain old wood and kindling, above 10000 ft.

    4. Server rooms don't catch fire that often, and having smoke detectors and a sprinker system are proven (though messy) solutions. Halon (or whatever replaced it recently for ozone layer preservation puposes) can often put out the fire while the rest of the room keeps humming along doing whatever the servers need to do.

    5. An oxygen deprived fire is slow and smoldering. It is exactly the kind of fire that will heat up nearby components to near ignition point then flash over when oxygen is supplied (e.g. geek opens server room door) and which is hard to spot while it's ramping up. A nice healthy blaze will make itself known (smoke, heat, visible flame) sooner and is actually safer.

    6. There are some extra costs the article doesn't mention: typical drywall and studs + drop ceiling is not exactly air-tight. So you'll have to do some construction work to keep all that oxygen out and to avoid dropping the O2 levels in nearby inhabited spaces.

    7. The machinery required to remove O2 is itself composed of mechanical and electrical components and requires power to operate, and further it has to vent the O2 someplace...this O2 rich exhaust has to be carefully managed or it will create and extreme fire hazard all by itself.