I see many folks asking if SAAS (software-as-a-service) will be hurt by data center outages. So I figured I would take a shot at it.

Anything over the IP network is going to be slightly less reliable than the TDM network. The TDM network had 100 years to create the 5 nines reliability that customers have come to expect. IP networks have not had that long - 15 years max to design some kind of reliability.

One big difference is that IP networks have more points of failure. Routers and switches on each end. Smart jacks. Cabling.  Sure TDM has RJ-11 cabling, but when was the last time that failed? Until it disintegrates it can usually carry some kind of voice signal. It's data that requires a better connection, more wire integrity. The signal carrying the data is more susceptible to interference, like the way an AM radio signal can knock out a DSLAM. 

Luckily for carriers, the consumers have been taught by cellular usage that convenience comes with some trade-offs. So dropped calls, blacked out areas, no coverage, roaming charges, unclear calls, and the like have helped to lower customer expectations about services.

That's a good thing for SAAS - whether it is Hotmail, Googel Apps, Salesforce, blogs, or what-have-you. Outages happen. Total redundancy is way too expensive for most companies to actually implement. Too many moving parts from machines that get cheaper and cheaper. Let's remember that The Cloud and virtualized apps are still sitting somewhere on a server and hard drive. Many points of failure there, but people have outages on premise based servers too. And on PC's. So there is some understanding.