Single points of failure pop up in the strangest of places. I’ll never forget the day when a colleague noticed a lonely single orange (fiber) cable coming out of a giant disk array. We were working in a data processing facility where the world’s largest computer manufacturers dropped off their latest-greatest technology for us to evaluate and bang on. The manufacturer was so impressed that he noticed an oversight on their part that they flew him out several times to meet with their engineers and review their designs in person. Sometimes seeing the obvious stuff makes you the expert.

In a similar vein, I was just reading a post in alt.folklore.urban that claims a US Navy Vessel was almost completely disabled when Inertial Navigation Systems (INS) went offline:

For close to a week and a half an Aegis class destroyer ended up piloted through shallow extreme-Northern Arabian Gulf waters by a combination of extremely cautious steering, celestial navigation, dead reckoning, and the occasional check by landmarks if we got too close to an oil platform.

[…]

More nerve-wracking for the rest of us was the fact that all our weapons systems with the exception of the Phalanx Close-In Weapons System are dependent on the INS for levelling information, the failure of which turned them into so much useless scrap. Those of us dealing with our one offensive weapons system, Tomahawk, regarded it as a grand vacation during which we didn’t have to notify the entire chain of command up to CentCom that we would be down for routine maintenance.

I’d be surprised if they don’t regularly practice running the ships on reduced navigability or impaired systems, but with today’s rapid-development and release industries it seems more prudent then ever to double-check for redundancy, in case of failure.