Disaster recovery as a service, or cloud-based disaster recovery (DR), makes testing a DR plan too easy to overlook -- the opposite of traditional self-managed BC/DR facilities.
One of the primary deterrents to most business continuity/disaster recovery (BC/DR) plans is that recurring testing must occur to ensure preparedness for when calamity strikes as well as to prove compliance for organizations with regulatory mandates. Testing may not only be arduous due to the complexity of bringing replacement systems online, but doing so without proper preparation may affect the original servers that are actively serving users.
When asked about the frequency of DR testing, Enterprise Strategy Group research respondents using cloud-based disaster recovery services are four times more likely than those self-hosting their BC/DR solution (20% vs. 5%) to perform weekly recovery tests to determine if and how quickly they could recover from a major outage. The contrast in testing frequency between those using cloud-based disaster recovery services as opposed to organizations relying on internally supported processes almost certainly stems from a couple of key factors:
- Most service providers not only offer “pay only for what you use,” but some even include discounts/waivers for DR tests. Either way, to pay only minimally for secondary storage resources and only incrementally more for the CPU resources consumed for a few hours or few days makes testing economically feasible -- instead of the more expensive processes related to traditional BC/DR service providers or self-managed BC/DR sites.
- Modern disaster recovery as a service (DRaaS) solutions often include the ability to “sandbox” or partition virtual machines so testing can be done without impacting the still-functional production servers. While most DRaaS offerings consider this capability to be “table stakes” or an essential feature requirement, sandboxing is often much more difficult in typical on-premises solutions using traditional virtualization management tools.
Combining these factors, testing failover preparedness in most DRaaS solutions is quite frankly too easy not to do, which is exactly the opposite of what limits the testing of many traditional self-managed BC/DR facilities. On the other side of the argument, considering the ease of DRaaS testing or even including the mandate that every BC/DR solution must be tested, it's admittedly disheartening that 15% of DRaaS and 9% of self-hosted BC/DR solutions aren't tested even annually, according to that same ESG research.
That replication feature is not a BC/DR solution
Regardless of whether you prefer DRaaS or self-hosted BC/DR, it's important to note that the replication technology within your storage systems or backup software isn't a BC/DR solution. Replication is a data movement technology that provides BC/DR solutions with the IT resources they need. Real BC/DR products and features manage development of the BC/DR plan and/or orchestrate the testing and failover of the resources at the secondary site. BC/DR is much more about understanding the business impact of key system failures, documenting recovery procedures, and then affecting corporate culture for preparedness and testing. That being said, for most organizations, if you don't have your data and key systems replicated, then the rest of your BC/DR plan is moot.
Without testing, you have a BC/DR hope, not a BC/DR plan
If you're not testing the recoverability of each key system at least quarterly, please do or you'll almost assuredly discover that those systems aren't as resilient as you thought (and when you need them the most). For those who struggle to maintain a self-hosted BC/DR solution or to test it regularly, DRaaS may be your answer.
Whether self-hosted BC/DR or DRaaS, the most important thing to remember is that the best BC/DR test is the one that fails, even partially, because then you know what to improve. If your BC/DR tests consistently result in all green checkmarks, you either have one of the top BC/DR solutions in the market today, which is well deployed and well managed, or (more likely) you're not testing thoroughly enough.
If you don't regularly test your BC/DR solution, then you don't even have a BC/DR plan, you have a BC/DR hope. With virtualization making servers portable and DRaaS solutions providing a second site that is cost-effective and has the manageability to orchestrate service recoveries, BC/DR is no longer just a hope or a plan, but can actually be something achievable for organizations of all sizes.
About the author:
Jason Buffington is a senior analyst at Enterprise Strategy Group. He focuses primarily on data protection, as well as Windows Server infrastructure, management and virtualization. He blogs at CentralizedBackup.com and tweets as @Jbuff.