[techslave]

> a novelty

not a novelty. the true bigiron vendors (not sun) had been doing this for decades.

mainframe reliability puts the upstart unix systems to shame.

[devchix]

Up to that moment, hardware maintenance meant having to power cycle the server. True HPC systems like those that ran at the US National Labs didn't find its way to the general market, and still haven't as far as I know.

[techslave]

even the AS/400 didn’t require power down.

this is actually so ancient it’s hard to find docs. here’s something from 1976. (the report is 1990 but the hardwares dates to ‘76). https://www.hpl.hp.com/techreports/tandem/TR-90.5.pdf