There are a lot of non-fire reasons that I like the presence of an EPO switch and I would design one in regardless. However, its requirement in the fire-code makes the expense easier to justify. I design framerooms with many smaller UPS's distributed throughout the racks, as opposed to a central pile of batteries. As a result, I have gel cells everywhere. I usually use Liebert GT's, every two racks, with power pod and interlock, at the bottom of the rack. Some of the non-fire reasons; Ground-faults placing 120VAC on the rack frames. Flooding event. Data intrusion prevention (Emergency SCRAM).

From: Sean Donelan [mailto:sean@donelan.com] Sent: Saturday, November 04, 2000 5:56 PM

Several people have asked where can they obtain a copy of the US Air Force draft.

Engineering Technical Letter (ETL) 00-3: Fire Protection Engineering Criteria - Electronic Equipment Installations

http://www.afcesa.af.mil/Publications/Drafts/ETL00-3Draft.pdf

I find it interesting how the designs created by people who have to operate an installation are different from the designs from consultants and vendors, or even the National Fire Protection Association. Or commercials on television like those being run by Janus.

But back to my original question. What is the real fire risk for ISPs and collocation operators. Is it burned buttered popcorn in the microwave setting off the FM-200 system? Is it home-made computers? Is it the Emergency Power Off switch?

State Farm Insurance has started a project to change the National Electrical Code for computer rooms and the requirement for an EPO switch. Is this something other ISPs and collocation providers would be interested in seeing changed? If so, we need to collect data and evidence to support the change.