Re: DATACENTER: power requirements

[Martin Hannigan <hannigan@shore.net>]

> On Fri, 4 Jun 1999, Jason Emery wrote:
> 
> > I am dealing with some upper management who is under the impression that we
> > should measure the amps drawn by the various pieces of equipment we are
> > installing to determine the "actual" amperage required. Theory being that we
> > can save money on our installed power circuits by only installing what is
> > needed. Does anyone have any feedback on any "real" reason not to do this.
> > Is power really that expensive? Any rough figures on the cost of installing
> > and cabling power? Thanks a ton.
> > 
> > Jason Emery
> > Network Engineer
> > Avenue A Media
> 
> 
> 	If you set up the device, got it operating at 99%+ load, raised
> the room temp to max, and then made accurate measurements of peak and
> nominal current draw...  Well, it should come out about the same as what
> the manufac rates the device as.
> 
> 
> 	If you install only enough to supply the devices at nominal load
> you could have some fun, what happens if too many devices are powered on
> at once?
> 
> 
> 	In other words, you are dealing with a pointy hair, feed em to the
> paper shredder.

Ok, but shouldn't the basic rule of thumb be you're using what you 
are fused at? 

In any event, it probably doesn't matter. If you are going into a colocation
scenario and they sell you 100A, I'm sure they're not going to say if
you can prove you only use 90A, we'll discount you. I think colo pricing
is pretty static in that respect.

-M