Bikecommutr
80 pts. | Sep 4 2007 3:02PM GMT
The best way to do this is empirically. Buy a power monitor and sample some racks. See what the actual draw is over a week. Assuming a good sample, you can extrapolate to the rest of the data center fairly accurately.
At worst, look at each power supply in the devices and add up peak/average load. Remember the formula watts = amps * voltage and it all becomes easier.
Brycenesbitt
20 pts. | Sep 17 2009 6:24PM GMT
Looking at the vendor specifications is essentially useless. Go out and get a Kill-A-Watt P4400
, and plug it in. The only downside is unless you have redundant supplies, you have to unplug first.
For examples of how far off “specified” and “measured” power usage is, see my
switch power consumption page at
Brycenesbitt
20 pts. | Sep 17 2009 6:25PM GMT
At:
<a href="http://www.obviously.com/tech_tips/Switch_Power_Consumption" title="http://www.obviously.com/tech_tips/Switch_Power_Consumption" target="_blank">http://www.obviously.com/tech_tips/Switc…</a>
<a href="http://www.obviously.com" title="http://www.obviously.
" target="_blank">www.obviously.com</a>






