In a letter to nearly 1,000 patients, University Hospitals in Ohio says that a tank's remote alarm system, meant to alert an employee to temperature swings, was disabled for an unknown length of time. That led to the destruction of around 4,000 frozen eggs and embryos, double the original estimate:
Hospital officials say they doesn't know who turned the remote alarm off or how long it was disabled. They also said they were aware the tank in question needed preventative maintenance. Some of the eggs and embryos had been stored there since the 1980s. The hospital's investigation is ongoing.
"Right now we do not know whether it's mechanical or human or [a] combination," said James Liu, chairman of the department of obstetrics and gynecology at University Hospitals.
He says he doesn't think anyone intentionally disengaged the alarm. "Because it is a computer, we think it's unlikely that there was any kind of external force that was working to hack the computer or anything like that. We think it's unlikely," Liu said.
Previously:
Freezer Malfunction May Have Damaged Up to 2,000 Frozen Eggs and Embryos
Two Fertility Clinic Freezer Failures Occurred in a Single Day
(Score: 0) by Anonymous Coward on Wednesday March 28 2018, @11:14AM (2 children)
I wonder what the numbers are now.
I suppose we're mostly getting errors in non-essential data, such as video streams.
(Score: 3, Informative) by c0lo on Wednesday March 28 2018, @11:52AM
With the transistors getting smaller, I think is reasonable a cosmic ray incident will cause more extensive damage by flipping more transistors. The frequency may stay the same, just the damage is more intensive.
Now, I don't know how much of that study looked into the isotopic composition of the materials around. E.g the chips are covered in plastic and 14C (in 1 ppm abundance) is radioactive. 40K - potassium - will be present in the concrete that makes the building - half chances to decay in energetic gammas that penetrate the chips.
https://www.youtube.com/watch?v=aoFiw2jMy-0 https://soylentnews.org/~MichaelDavidCrawford
(Score: 3, Informative) by c0lo on Wednesday March 28 2018, @01:11PM
A more recent (2003) article (PDF) [iaea.org]
256 MB is 2048 Mbit - thus apple-to-apple comparison would say 0.0172032 errors/hour at the level of 2003 memory. Or 1 soft error every 58h = 2.6 days in 2003
That was 2003. The today transistors are much easy to upset (lower charge thus incident energy required) thus I expect that the number upset events to be increased.
Supplementary, the amount of memory cells in the same silicon volume increased zillion-fold - Moore's law hold quite strong. I wouldn't be surprised to hear of multiple soft errors per hour today.
Other stories
The Invisible Neutron Threat [lanl.gov] - SEU - single event upset
Google: DRAM error rates vastly higher than previously thought [computerworld.com] 2009
https://www.youtube.com/watch?v=aoFiw2jMy-0 https://soylentnews.org/~MichaelDavidCrawford