4 ms·

Examining and Learning from Complex Systems Failures

33 points by wallflower 10y ago

awinter-py 10y ago
very much want to read this but uptimeinstitute is currently, and very ironically, down (maybe because of a complex system failure? or just load)
- yagibear 10y ago
  Archived at https://web-beta.archive.org/web/20170417232443/https://journal.uptimeinstitute.com/examining-and-learning-from-complex-systems-failures/ https://web-beta.archive.org/web/20170417232443/https://jour... or use the Google cache version from the web link at the top of the thread
  - a3n 10y ago
    Thanks for the Google cache reminder. When I view that, it also spins forever, with no page display below Google's header. If I click on the Text link, then I see the text version. I wonder if there's a problem in the source itself, rather than the server.
a3n 10y ago
> 16. Safety is a characteristic of systems and not of their components. > 18. Failure-free operations require experience with failure (Cook 1998). These two things in particular caught my eye, especially as I'm not in system design. So with 16, you can't just slap together components with sufficient -ility rating, you have to study and design the safety of the system as a whole. Those components, in two different designs, are exposed to different stresses in different ways, and live within different monitoring and mitigation systems. With 18 ... I wonder what the value would be to a promising new engineer who opted to not take the big offer from Google or Apple, and instead worked down among the unwashed for awhile, gathering experience. There's probably a lot of value in having experienced failures in a system that has no relevant mitigation, and you have to figure out what to do next, as the first person to have seen it. Luxurious to work at Google, where if it hits the fan, there's lots of systems and experience to handle it. Character building if you alone had to figure out how to keep the ship from sinking, from an "iceberg" that no one had thought of and planned for.

33 points by wallflower 10y ago

very much want to read this but uptimeinstitute is currently, and very ironically, down (maybe because of a complex system failure? or just load)

yagibear 10y ago
Archived at https://web-beta.archive.org/web/20170417232443/https://journal.uptimeinstitute.com/examining-and-learning-from-complex-systems-failures/ https://web-beta.archive.org/web/20170417232443/https://jour... or use the Google cache version from the web link at the top of the thread
- a3n 10y ago
  Thanks for the Google cache reminder. When I view that, it also spins forever, with no page display below Google's header. If I click on the Text link, then I see the text version. I wonder if there's a problem in the source itself, rather than the server.

a3n 10y ago

> 16. Safety is a characteristic of systems and not of their components. > 18. Failure-free operations require experience with failure (Cook 1998). These two things in particular caught my eye, especially as I'm not in system design. So with 16, you can't just slap together components with sufficient -ility rating, you have to study and design the safety of the system as a whole. Those components, in two different designs, are exposed to different stresses in different ways, and live within different monitoring and mitigation systems. With 18 ... I wonder what the value would be to a promising new engineer who opted to not take the big offer from Google or Apple, and instead worked down among the unwashed for awhile, gathering experience. There's probably a lot of value in having experienced failures in a system that has no relevant mitigation, and you have to figure out what to do next, as the first person to have seen it. Luxurious to work at Google, where if it hits the fan, there's lots of systems and experience to handle it. Character building if you alone had to figure out how to keep the ship from sinking, from an "iceberg" that no one had thought of and planned for.