[Beowulf] removing tiles around heavy racks?

[Robert G. Brown]

On Tue, 3 Jan 2006, Mark Hahn wrote:

>> I've not studied the physics of it, but I can't see where the
>> surrounding tiles would have an effect.  You obviously don't want to
>
> sorry!  I should have explicitly said I'm worried about stability,
> not pure downward load-bearing-ness.  I mean, the floor is level,
> but not microscopically, and if someone leans on a rack, I'd like
> not to have the whole machineroom suddenly shear ;)
>
>> remove the tiles the rack is sitting on.  And those tiles should be
>> supported by a metal grid.  As far as I know, the metal grid and the
>
> presumably the grid (stringers) do most of the lateral stabilization,
> but they're not all that hefty themselves.  obviously with all/most
> tiles in place, the mesh is MUCH stiffer.
>
> but I guess that's the real point: it's a question of how much lateral
> thrust might possibly be encountered, and how well, say, a tile-less
> mesh of stringers+feet would handle it.

Actually, I don't think that tiles ordinarily do anything at all for
stabilization.  If they did, they would bind up in their holes, but I
think on most floors any given tile is likely to be easy to pull up.

The structural engineering of a raised floor appears to be produced by
three components.  The "pedastal", the "crosspiece" (a.k.a "stringers")
and a variety of optional "diagonal" pieces.  The pedastal structures
bear the load of the racks (and floors may well be a symmetric array of
pedastals, so to speak). The crosspiece/stringerss link the pedastals
together to form a regular grid and increase the floor's capacity.  They
produce a variable amount of structural stability in two ways -- to
distribute and increase load bearing capacity directly (by being thick
vs thin) and against parallelogram-type deformations.  Thick, wide
crosspieces won't rack (deform) easily.  Finally, to maximize stability
and load bearing potential, a variety of diagonals can be added to the
squares and the pedastal supports to produce rigid structural triangles.

All of these components can be varied with respect to
thickness/strength, and a wide range of potential loads, electrical
properties, cooling access patterns, and so on.  When properly
assembled, the tiles are really there primarily to be walked on or
otherwise distribute loads onto the understructure, which is pretty much
completely structurally sound with or without tiles at all.  That is, I
don't believe that they themselves are responsible for much in the way
of capacity or stability, aside from needing ones that are strong enough
to not break under the vertical loads of the racks as you move them
around.

GIYF -- you can find all sorts of stuff with "raised floor tiles
stability data" and similar strings.

However, I like one of the original responses.  Where health and
personal safety are at issue, the right thing to do is to talk to the
techs of whatever company(s) you are considering to do the work as you
go.  They should be able to fully answer your questions in such a way
that you can fully understand -- and hence have reason to trust -- those
answers.  I'd guess that with a raised floor you get what you pay for --
probably not a good place to scrimp, especially if you want it to be
adequate for generations of hardware that might gradually increase in
weight.  Raised floors can be engineered for pretty much anything you
like -- I even found a site that sells earthquake-proof extra-strength
shock-absorber pedastals so you can install oversized vibrating loads
right in the middle of your server room floor safely.

I'd guess that for the MOST part, data-center raised floors are
overengineered as SOP.  A badly constructed one or inadequate one that
collapsed would leave the seller open to all sorts of very expensive
lawsuits, given that the cost of the hardware on the floor is likely
tens to hundreds of thousands of dollars per square meter.  Loaded racks
still weigh in at around (or a bit less than) a ton per square meter,
which is a lot but not THAT much on a solid subfloor.  Under ordinary
circumstances, racking torques should be relatively small even as you
move stuff around.

    rgb

>
> thanks, mark hahn.
>
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
>

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu