There was a time when the enterprise was on its own if it wanted to build infrastructure for disaster recovery (DR). When it came to housing critical data and applications, the local data center was the only viable option. But maintaining near-perfect reliability with this approach came at a high price. In most cases, it required building an entirely new data center from scratch, establishing the architecture to continuously move required data and then ensuring a smooth operational transition should a calamity befall the primary site.
In today’s world of tight budgets and razor-thin margins, however, this approach is no longer an option. Fortunately, increased cloud and colocation adoption is stepping in to fill the void through all-cloud or hybrid DR infrastructure. Still, enterprise executives should be aware that moving from a secondary data center to a cloud/hybrid model is no walk in the park. The process can be highly complex, with many finite details to attend to—and errors can be substantial. For this reason, proper planning and testing is the main element in making sure the transition is as smooth as possible.
The first step, before any actual changes to the DR architecture, is to take stock of the entire IT environment that requires preservation in an emergency. The best way to do so is to identify which services and functions would be most harmed by extended downtime and then identify their basic resource needs and long-term strategic goals. Next, the enterprise must address several important considerations before, during and after the cloud transition.
Gaining From the Cloud
A good place to start is to determine exactly what you’ll achieve by shedding the secondary data center in favor of cloud-based DR. If the legacy data center has been in service for several years, the most likely benefit will be a more streamlined infrastructure and a more direct access plan. Also, the DR process should be much quicker and trouble free, given that most data in the cloud will reside in virtual machines that are far more flexible than fixed hardware.
But the best benefit of the cloud/hybrid model is that it may be half the cost, or less, compared with a traditional solution. Many providers specialize in disaster recovery as a service (DRaaS) and have fine-tuned their practices to better compete in an increasingly crowded market.
To achieve the highest cost/benefit ratio in the cloud, however, the enterprise must focus on a number of factors.
Software Licensing and Support
Most agreements are lengthy and complicated, and the fine print often has a dramatic effect on the move to the cloud and the new model’s ongoing performance. This is why you should go over licensing agreements carefully, ensuring you’re familiar with details for everything from replication, backup and DR software to wide-area connectivity and other technologies. A good rule of thumb: if you can get by without it, don’t buy it.
Skills and Training
Although DRaaS is easier to manage, it still requires some retraining. A managed-service provider (MSP) that specializes in DR can guide you in creating and securing the environment while your IT staff gets up and running with the new backup and recovery operations.
Some costs defy classification because they’re unique to each enterprise. A long-term lease for the backup facility may need to be broken, or various aspects of the business model may be affected by the transition. Determining your obligations ahead of time goes a long way toward avoiding surprises in the cost/benefit analysis.
Many Clouds, Many Benefits
We often refer to the cloud when most enterprises are gravitating toward a multicloud model. This trend offers numerous benefits, such as allowing the enterprise to choose the right provider for the right service, and it reduces dependency on a single provider to support the core business model. From a DR perspective, a multicloud approach vastly reduces the chance that a single event will knock out an organization’s entire IT infrastructure, and it also helps to maintain top-notch service to customers even during peak activity.
This dynamic also applies to the selection of telecom connectivity to and from the cloud. As usual, competition is the best way to achieve high quality at low cost, so it never hurts to keep your carrier options open and to make sure your provider knows that your business isn’t unconditional.
Too often, organizations embark on grand initiatives without defining what success will look like. DR is no different, but fortunately it can be gauged according to two metrics:
- The recovery-time objective (RTO): How long can you go without critical resources before the business impact is too great to bear?
- The recovery-point objective (RPO): How much data can you lose in a given event?
Few organizations can withstand more than a few minutes of downtime when it comes to critical applications and services. Data tends to suffer from the same time crunch, since most modern applications provide effective service with data that’s more than a few hours old.
This is probably the most important reason why it’s imperative to shed aging backup infrastructure in favor of a hybrid approach. Not only does the cloud excel in speed and reliability, but emerging options such as edge computing can improve in these metrics even further by keeping data close to those who need it.
DR has long been the most extreme example of a cost center. The costs involved in building and maintaining backup infrastructure are usually just as high as or higher than building primary infrastructure, not to mention the time and effort to move data, test recovery tools and processes, and attend to other operational needs. And all of this effort supports an infrastructure that will (hopefully) never be used.
By any measure, the cloud offers a better approach to DR, just as it has for countless other aspects of IT. Not only do service-based hybrid solutions provide major time, cost and resource-consumption advantages, but they do so while improving DR performance. These facts give the enterprise a clear path: ditch the backup data center and invest in hybrid infrastructure instead.
About the Author
Laz Vekiarides is cofounder and CTO at ClearSky Data. For over 20 years, he has served in technical and leadership roles to bring new technologies to market. Most recently, he was executive director of software engineering for Dell’s EqualLogic Storage Engineering group.