In his recent column “DR-as-a-Service may be the Cloud’s Killer App” my colleague Eric Slack predicted that disaster recovery as a service (DRaaS) will become the predominant cloud offering. I agree, DRaaS is almost too compelling for many organizations to pass up. But like any new technology it requires that organization ask new questions and potentially alter their processes to take full advantage of the offering. This is especially true in the case of DRaaS, because not knowing the answers to the right questions and altering your DR strategy accordingly could lead to complete business collapse.
Why DRaaS Exists
DRaaS is being offered primarily by online or cloud backup providers. These companies typically have a long history of providing backup and recovery solutions over a network connection. Most expectations are that they provide a means to get the initial backup done, a remote location (cloud) to store that backup and software that allows that backup to be incrementally updated as the original changes in your data center. And, of course, you expect that you’ll be able to recover files, applications and even complete servers from their cloud.
Therein lies the rub. Most of the advanced software technologies, like deduplication, that enable this incremental update over a constrained internet connection, won’t do any good in a full recovery. While there are other techniques available – like leveraging a local appliance, physically shipping storage devices from the cloud and using changed block or deduplicated recovery – all have their downsides. And more importantly, they all count on a surviving data center and an intact data set.
DRaaS, for cloud providers, is the ultimate workaround. They “simply” spin up your application(s) in their cloud and you are “instantly” back on-line. As we discuss in our data protection workshops, this is an excellent workaround for the cloud recovery challenges, but it does require that you ask new questions and potentially change the return part of your DR strategy.
DRaaS Changes The Cloud Relationship
The key component to understand about cloud recovery is that the relationship between you and the provider has now changed. Before DRaaS they were simply a repository, with only a responsibility to store and protect your data. Now, however, they are going to run your application and instantly change from a backup vendor to a compute vendor. Do they have service level guarantees around that application’s performance? In the event of a regional disaster like Hurricane Sandy, do they have a plan to provide constant performance to every one of their customers when they all declare a disaster?
Companies that provide compute as their primary focus, like Infrastructure- and Software-as-a-Service vendors, struggle to meet the service levels their customers rely on. You can barely go a day without reading about some service or another that was down. Just last week my bank (BBVA Compass) was down for most of the day, and eBay and Facebook had reported outages. In fact as I write this, the great Google is having intermittent issues. Can backup providers that have now become compute providers be counted on? The answer is “probably” but IT planners need to ask the hard questions and get the right answers to make sure they can.
As we continue our coverage on DRaaS we will provide you with the answers that cloud backup providers could give you and how to determine which ones make the most sense.