The Cloud and Disaster Recovery Bookmark has been added
The Cloud and Disaster Recovery
Deloitte on Cloud blog
While most people realize that disaster recovery (DR) capabilities are built into public clouds, in many cases, the reality is that it’s your responsibility. Moreover, what you think is there may not be.
July 26, 2018
A blog post by David Linthicum, managing director, chief cloud strategy officer, Deloitte Consulting LLP
Most enterprises have some DR capabilities, even if it’s just on their most critical systems. When data was on-premises, organizations had to stand up systems that were geographically different. The choices for DR systems are hot stand-by (active/active), which are ready to take over at any given moment and are always running, or, stand-by systems, also geographically different, that can be stood up in short order (active/passive).
Not placing the primary and redundant systems close to each other means that if the need arises to deal with a natural disaster that takes down a data center in one part of the country, or one part of the world, there is another data center available that is located nowhere near the primary. This helps insure that the backup data center will be able to take over.
For public clouds this is a feature that is largely built-in. Since cloud providers run many data centers all over the world, and do so by regions, you can configure an acceptable DR architecture. These typically deal with primary and secondary regions, which can either be active/active, which are dual redundant systems that run at the same time, or active/passive, which could mean there are backups of the data that will need to get up-and-running as part of the recovery.
There seem to be two schools of thought when it comes to DR and cloud computing. One school believes that DR is built into public clouds, and thus you don’t have to deal with it. The other school believes that clouds won’t be able to recover from a DR scenario quickly, and thus you need to set up redundant traditional on-premises systems, as well as an active or passive redundant backup. I argue that both schools of thought are wrong.
The right answer is likely somewhere in the middle.
Keep in mind that you’re responsible for your own DR, no matter what’s built into the public cloud. The good news is that public clouds understand what you need and can provide basic DR services.
While public clouds have their own resiliency built into the clouds that will help you get through most types of major outages, even outages that you may not know occurred, you should still account for major disasters that might affect operations. For example, a hurricane takes out a public clouds region which you have selected as the region for your primary systems. Thus, you need to select a secondary region that is nowhere close to the primary region, and make sure that the secondary region has current data and current copies of the applications. In other words, you maintain a complete copy that’s ready to take over if the primary region goes down.
These cloud DR mechanisms should be set up by you, and not the cloud provider—that’s the "bad news." The good news is that the public clouds do much of the work for you. It’s just a matter of configuration.
Interested in exploring more on cloud?