#9, please tell Steven J. Vaughan-Nichols and others. For them, every outage for a Windows-based system is due to Windows, and every outage for a Linux-based system is due to everything but Linux.
I think the other issue here is comparing uptimes between dissimilar systems. A PBX should boast 5 9's. A server room switch (and I wish a consumer-grade switch!) should hit 5 9's. A tightly-controlled, mature, predictably-accessed back-end database for a banking system should crank out 5 9's. Yeah, it'd be great if BPOS/O365/Google Docs also hit 5 9's already, but this is relatively uncharted territory, particularly at this scale. Give them a few years to mature.
Don't get me wrong--I'm not crazy about these services so far, and uptime is one of the reasons. The amount of service downtime doesn't really bother me, it's just that you don't have any control over it. When you host your own, you can schedule necessary maintenance around the requirements of the business. The downtime figures are also not an aggregate of total customer-side downtime, just the backend system. If you lose your internet link (either locally or somewhere upstream), have a routing issue, have a firewall issue, or some other reason that prevents access to the service (DDOS?), those will drop the "actual" downtimes as perceived by end users. And at the end of the day, that's the only one that counts....:)
|