"Pulling the wrong plug"? Sorry, Sam. Not buying it. At. All.
The only item in a DC that I can see causing that type of "outage" would be "pulling the plug (power, network, what have you)" on a SAN/NAS/Filer. It would then take time to rescan and rebuild block indexes across the disk-span. Assuming, of course, that they totally ignored the routine "rebuilding of indexes -- or "defrag" as windows used to call it -- which is also routine practice.
The biggest problem I have with that theory is, if you are an airline, and you know that this would be a single point of failure, do you not have redundant PSU's, fed from separate UPS's on separate power feeds into the DC? If it was network-related, did they not bond NIC's on separate switch fabrics?
In my previous lives (well, earlier in my career) I architect'ed and deployed such in far-less mission-critical infrastructure. The driver in those cases were typically Engineering "departments" at software dev houses cost tens or hundreds of thousands of dollars per hour of lost work-time. Not even personal safety...
/nerd