The outage of Amazon cloud services asked the following question – Do you blame your service provider or do you blame the architecture you used to integrate with your service provider?
While other entities went offline due to the Amazon’s outage, Netflix remained online because they did not rely solely on Amazon. Redundancies were implemented with other Service providers in the case of an outage. One must remember that the cloud is another platform. Hence architecture principles applied within your organization should be applied to the cloud.
Availability, Security and other non-functional attributes need to be applied in the cloud. A single cloud provider will affect the availability of your application of that cloud provider fails. Failover is solved by the addition of another service provider. You may argue that this would be costly; however will not be costly to lose customers?