Any good disaster recovery plan (DR) mitigates against a variety of different scenarios that might bring a service down or otherwise cause an incident.
Good DR plans also connect those scenarios to goals such as high availability (HA), plus metrics for measuring performance against those goals. Common examples include data recovered — how much was lost? And time to recovery — how fast was the service restored?
These principles have long been in place and remain largely true today. But much of the thinking about DR remains ingrained in on-premises infrastructure, which unnecessarily limits companies from bolstering their DR strategy and response to today’s incidents with new ways of doing things in the cloud.
It’s time to modernize the way we think about and approach DR. I see five interrelated ways that DR in the cloud can help organizations of all types improve their approaches to preventing and recovering from data loss and downtime while ensuring they’re meeting their SLAs. Here’s how DR in the cloud can help:
1) Run on world-class infrastructure
To start, the major cloud platforms give virtually any organization access to world-class infrastructure that, with few exceptions, beats what they run on-premises. From compute to network to storage and other IT resources, the major cloud vendors are able to offer top-notch infrastructure at accessible prices.
Few companies are able to replicate what the major cloud platforms can do in their datacenters. The cloud platforms also invest significantly in their own reliability and resiliency, meaning if a server or other resource goes down, they can likely be back online faster than most companies can accomplish on their own. This in and of itself can immediately improve your DR position, especially in organizations that are running on aging infrastructure
2) Implement more creative recovery designs
Cloud also enables more creative recoverability patterns than you can achieve on-premises. There are cloud recovery designs that are simply not feasible or would be prohibitively expensive to implement on-premises.
You also have more flexibility to create recovery designs that match your particular goals or requirements. You can design for performance and uptime with a multi-site DR pattern, for example. Or you might decide to design for cost and opt for the “pilot light” approach to DR, in which you replicate a core set of critical resources rather than the entire solution.
The pilot light method is a particularly good example: Hyperscalers allow you to pay-as-you-go, so you can built a pilot light DB Server that only switches on once per day to apply new logs before shutting down again reducing uptime spend by (23/24 =) 95% – this type of pattern is just not viable on-premise.
It’s worth adding here that hybrid architectures mean you don’t have to dump all of your existing on-premises investments and your workloads can still get much of the cloud DR experience
3) Benefit from automation and orchestration
Cloud is highly programmable, which is one of the major reasons that it can enhance existing DR strategies in ways that would be difficult or impossible to do entirely on-premises. This means cloud offers much greater automation and orchestration capabilities for DR solutions.
Much of the creativity in recoverability designs can be attributed to automation. While you can apply automation on-premises, the major cloud vendors make cutting-edge capabilities accessible for virtually any company.
Those technologies might otherwise be out of reach for organizations with limited in-house skills in automation and orchestration. Each of the major cloud vendors offer a managed distribution of Kubernetes for container orchestration, for example, which is an increasingly popular choice for HA systems.
In general, automation is embedded into many cloud and cloud-native technologies, and as such reduces the manual overhead typically required of some traditional on-premises recoverability strategies
4) Improve DR while keeping your costs under control
Again, one of the overarching benefits here is that cloud is making all of this — from elite infrastructure to modern automation technologies to a wider menu of recoverability patterns — more attainable from a budget perspective.
Much of what you can do in the cloud would be incredibly expensive to implement on-premises. That means cost makes it a non-starter and organizations stick with legacy DR planning that may no longer meet their needs. This is one of the broad promises of cloud — world-class infrastructure and services at hyperscale pricing — applied specifically to DR. DR resources should not be as big of a roadblock to recoverability as they may have been for some organizations in the past
5) Work with a vibrant ecosystem
The other good news is that there is a robust, growing ecosystem of partners with deep expertise in cloud. Some of these companies are specifically developing recovery software and solutions for cloud that can be implemented far more rapidly than with a DIY approach.
This is a bonus: You get all of the benefits above while letting someone else do much of the work for you. Essentially, cloud has made it so that there are no excuses for outdated, ineffective recovery patterns.
Also important is the rising frequency and cost of ransomware being used to attack systems in a wide range of industries, from healthcare to financial services to government and more. DR in the cloud can help in these cases because the platforms offer various data replication options and point-in-time database restores as a ransomware recovery strategy.
Countless businesses have already made the leap to the cloud. It’s time for their DR plans to do the same.
ABOUT THE AUTHOR:
Eamonn O’Neill is Co-Founder, Director & Chief Customer Officer at Lemongrass