Office 365, which launched today, is interesting, but in order for it to be a success, it's going to have to live up to its name; its predecessor, Microsoft's BPOS service, has rung up enough downtime in the last two months to give any potential customer reason to think twice about relying on the cloud.
Now, to be fair, all of the BPOS outages in the last year don't appear to add up to much more than a couple of working days of downtime, and that's only in the unlikely event that one was affected by every incident for the entire duration. But the unplanned interruptions show us just how difficult it is for anyone - even an organization as free from resource constraints as Microsoft is - to make this stuff work.Microsoft promises three nines of availability for Office 365; if my math is right, that's just under a minute and a half of downtime a day. Not bad, since even across the average working year of 250 days, that only amounts to six and a half hours; for a 24x7 view, that's a little more than nine hours. BPOS, for all of its faults, has proven that this much is doable.
But to push Office 365 to four nines, where it would truly live up to the "365" in its name, is the next big milestone. That would amount to less than nine seconds per day in downtime, or about 36 minutes over the working year; most organizations spend more time than that on fire drills, I hope. When Microsoft can deliver that kind of availability, then it will have conquered the cloud. Until then, "Office 365" is an aspiration, not a guarantee.