Put it this way. 24-48 hours of down time for us if something as catastrophic as the data centre melting would lose a few customers as they would realise it's something out of our control.
One indication that it is about a lot more than simply restoring backups. What else outside the servers was damagedin this incident? In any another incident here or elsewhere, what non-server, essential equipment and services could possibly be damaged or disrupted? How many of us would really, genuinely, be prepared to offer worthwhile SLA's when customers are are so very price sensitive? (read the small print)
Which just goes to show that the level of service required depends on the type of website. We couldn't afford to be down for more than an hour, let alone three days. Our DRP strategy would allow us to recover from a complete blow-out at one geographic site, although that worst case scenario could mean an outage of 24 hours.
