The decision to replicate a virtual machine either from on-premise to a service provider or between geographic infrastructure as a service locations, should be based on the Recovery Point Objective and Recovery Time Objective of the particular virtual machine. Simply put, how much data could the business sustain to lose and how much time can the business sustain without the application being available. Does your organisation understand the true costs of downtime? Do you have a thorough understanding of which applications or systems are most critical to your businesses performance? Is there a robust plan to restore systems in order of criticality? When was it last tested? If you're unsure of any of these questions, let's talk. If we can't help, we'll know who can.