Large-Scale, Simultaneous Failure of a Product or Technology

Yesterday, 30GB Zunes (including mine) all over the country crashed.

Have there been other instances where several thousand (or several million) units of the same product or technology all failed simultaneously? I’m sure it’s happened before, but I can’t think of any specific instances right now.

Talk about planed obsolescence. Wow Just Wow.

i heard on the news that they were going to start working again at 12:00 midnight newyears. so are they working now? is everything good again. and what caused this to begin with?

as for the OP i remember hearind a brand of microwave ovens went on the fritz with Y2K. other than that, maybe one of the big auto recalls might count.

Well, since it was a boneheaded bug, I think it’s more unplanned

Well, but with things like car/appliance/consumer product recalls it’s more of a case of a production point failure that does not necessarily result in simultaneous loss of functionality of every instance of the application – most motorists can drive on w/o going up in flames until they get around to taking the car to the shop for the recall repair. This sort of simultaneity almost requires embedded control technology so it would be something that begins becoming more frequent in the future.

It’s maybe not quite what you’re looking for but it’s arguable that the potato blight which caused the famine in Ireland and other countries in the 1840s was an example of this. There was a similar thing with European vines in the 19th century when they all had to be grafted onto American rootstocks. Hopefully we’re not about to see another example with bees.

I expect that bananas may fail before bees, sadly…

Leap years have been a big source of bugs for commercial software products, as have clock counter overflows. Usually, the problems get patched or the products are replaced/updated before disaster strikes.

What exactly happened with the Zunes? Is the cause of the failure known?

Bug in firmware Microsoft wrote for the Zune: unit incapable of operating on final day of leap year.

Yes – the units were first sold in 2006, and obviously whoever wrote the firmware never thought they’d last until the next leap year.

Another such example is the 1970 US corn crop disaster, in which the introduction of a single gene, Texas male sterile, rendered corn plants susceptible to infection by * Helminthosporium maydis*, Southern corn leaf blight.

How about the thalidomide disaster?

The Distributed Proofreaders website failed for Opera (and some FireFox) browser users on Dec 31st, and started working again at Jan 1st West Coast time (which is what the server runs on).

It’s working now, but efforts to trace it down seem to indicate a bug in a commonly used subroutine wget setCookie, which uses a slightly non-standard format for the expires date. And apparently this format gets processed OK by other code, except on the last day of a leap year. So the programmer just called standard subroutines, one of which happens to have a deeply buried bug that only occurs for a few hours out of every 35,064 hours.

I wouldn’t be surprised to see that the Zune problem is related to this, too – the description seems similar, and the fact that it fixes itself when everybody has rolled over to the new year.