ROC Part 3: Margin of Safety in CS&E?
Today marketing claims of 5 9s of availability (99.999%) but customers achieving 2-3 9s (99% to 99.9%)
Like Civil Engineering, perhaps we will never make systems dependable until we add a margin of safety (“margin of ignorance”) for things we don’t (or can’t) know
- No more “that failure doesn’t count”
Perhaps we need to “over engineer” by a 1-2 9’s to deliver in practice what we claim in theory?