3 ms·
Seems fair (this is a good argument). I don't think there's a huge amount of precedent. There are systems which are tightly coupled and thus all-or-nothing (m
by rdl 14y ago
Seems fair (this is a good argument).
I don't think there's a huge amount of precedent. There are systems which are tightly coupled and thus all-or-nothing (most hosting pre-cloud, power generation facilities, etc.), and systems which are highly distributed (networks). Generally for the distributed systems, no one reports a global uptime percentage, but a per-link uptime target and percentage.