I am glad that SOTT is back up and running.
Your IT guys might want to look into redundant fail-over systems,
clustering, or web server failover farms with smart routers with
failover redirection since you already do have several web-host
providers hosting cassiopaea[.com/.org], sott, and so on, so
it is possible to "integrate" all your sites together at each of the
web-host providers but with fail-over and/or "crossover" support?
You should think about your databases too as they should also
be redundant. There are software (and hardware) solutions that
can replicate databases on the fly, for example.
In short, redundancy should be seen at a global level, from a
geographical, hardware, and software point of view. The design
of the system should be well thought out.
As for "mirrored" drive systems, if one mirror breaks, the other
mirror should have kept on going depending on the mirroring
hardware or software used. The system *should* have kept
running and all it should have taken was a hot-swapped drive with
automatic mirroring recovery that would have resulted with no loss
of downtime. It always depends on the software/hardware used but
hardware solutions are always best depending on the vendor. With
expensive harware raid cards, I have used RAID 0,1 and 5
configurations and had it running for a very long time except in cases
where the motherboard, power supply, or power failure at the plug.
There are of course redundant motherboards, power supplies, and
so on but it gets expensive, and then there is clustering (more than
one computer and in geographically dispersed locations.)
But, with the proliferation of cheap web-hosting sites, you should be
able to design a system where if the primary web server (farm) fails,
other web servers (farms) should pick up where the connection was
dropped.
Of course, the hard part about redundancy is getting the knowledge,
understanding it, and applying it, and it takes resources, time, and
money.