On January 6, 2009, salesforce.com suffered a 40 minute outage that was determined to be caused by a hardware failure. The automatic failover apparently didn't work, but the salesforce.com technical staff performed a manual failover and things were back up and running quickly.
It is unfortunate that outage like this happen. But they are inevitable. SaaS providers like inContact and salesforce.com work diligently to create systems that are very reliable. Reputable SaaS providers build strong redundant systems with automatic failover systems and manual processes to deal with events such as the salesforce.com outage on Monday.
Nobody is immune from failures like this. This happens to both SaaS providers as well as big and small companies alike who choose to maintain their own applications. Most companies don't deploy their systems in redundancy and keep a spare hardware replacements for the big expensive pieces in the architecture. The typical scenario of a company who is maintaining their own applications instead of utilizing SaaS providers is that they are usually waiting for a hardware vendor to provide new equipment. This could take several hours or even overnight shipments in order to get back up and running…not to mention the people involved who are always on call.
Good SaaS providers are very concerned about reliable service and they do deploy geographically redundant systems in order to avert the potential of outages. Good SaaS providers also employ a staff to monitor 24/7 the operations of the application and react quickly when badness happens. SaaS providers can be trusted. I believe salesforce.com is as trustworthy as they come.
I am impressed with the scope of the deployment of the salesforce.com application platform and their ability to bring the whole system back up quickly after a significant hardware failure. We at inContact are pleased to be part of the SaaS family and plan to continue the tradition of high availability in SaaS despite the inevitable combination of events that lead to unfortunate, but always temporary failures.