Amazon and 2nd Watch have published numerous white papers and blog articles on various ways to use Amazon Web Services™ (AWS) for a disaster recovery strategy. And there is no doubt at all that AWS is an excellent place to run a disaster recovery environment for on premise data centers and save companies enormous amounts of capital while preserving their business with the security of a full DR plan. For more on this, I’ll refer you to our DR is Dead article as an example of how this works.
What happens though, when you truly cannot have any downtime for your systems or a subset of your systems? Considering recent events like Hurricanes Sandy and Katrina, how do critical systems use AWS for DR when Internet connectivity cannot be guaranteed? How can cities prone to earthquakes justify putting DR systems in the Cloud when true disasters they have to consider involve large craters and severed fiber cables? Suddenly having multiple Internet providers doesn’t seem quite enough redundancy when all systems are Cloud based. Now to be fair, in such major catastrophes most users have greater concerns than ‘can I get my email?’ or ‘where’s that TPS report?’ but what if your systems are supporting first responders? DR has an even greater level of importance.
Typically, this situation is what keeps systems that have links to first responders, medical providers, and government from adopting a Cloud strategy or Cloud DR strategy. This is where a Reverse DR strategy has merit: moving production systems into AWS but keeping a pilot light environment on premise. I won’t reiterate the benefits of moving to AWS, there are more articles on this than I can possibly reference (but please, contact the 2nd Watch sales team and they’ll be happy to expound upon the benefits!) or rehash Ryan’s article on DR is Dead. What I will say is this: if you can move to AWS without risking those critical disaster response systems, why aren’t you?
By following the pilot light model in reverse, customers can leave enough on premise to keep the lights on in the event of disaster. With regularly scheduled s to make sure those on premise systems are running and sufficient for emergencies, customers can take advantage of the Cloud for a significant portion of their environments. From my experiences, once an assessment is completed to validate which systems are required on premise to support enough staff in the event of a disaster, most customers find themselves able to move 90%+ of their environment to the Cloud, save a considerable amount of money, and suffer no loss of functionality.
So put everything you’ve been told about DR in the Cloud in reverse, move your production environments to AWS and leave just enough to handle those pesky hurricanes on premise, and you’ve got yourself a reverse DR strategy using AWS.
-Keith Homewood, Cloud Architect