If you manage a cloud environment, you're probably familiar with conventional cloud backup and recovery techniques — like "warm standby" and "pilot light."
These are great strategies if you use a single cloud environment. They may even work well enough in most multicloud setups.
But when you migrate to a hybrid cloud environment, cloud backup and recovery becomes more complicated. Here's why, and what you can do to make sure you're prepared for a quick hybrid cloud disaster recovery.
Traditional Cloud Backup and Recovery
If you work with a single-cloud environment, the main question you'll have to answer when planning for backup and recovery is which balance to strike between cost and performance.
The various backup techniques out there — which AWS summarizes nicely — offer different tradeoffs between what it costs to support them and how quickly you can restore functionality based on them. Techniques such as pilot light allow you to minimize the resources you devote to disaster recovery preparation, but they lead to slower recovery times. On the other hand, warm standby or multi-site approaches enable fast recovery, but at a higher cost.
Why Hybrid Cloud Disaster Recovery Is Different
The challenge for businesses that have adopted a hybrid cloud architecture — one that integrates private infrastructure into a public cloud environment — is that the cost and performance calculus for disaster recovery plans in a hybrid cloud are considerably different from those in a single-cloud environment.
There are a number of reasons why:
Limited infrastructure scalability
For starters, hybrid clouds lack unlimited infrastructure resources because they depend in part on private infrastructure. You can't simply spin up new on-premises servers to recover from a hybrid cloud disaster in the way that you can launch replacement VM instances in a public cloud to create a recovery environment.
For this reason, techniques like pilot light and warm standby, which require rapidly scaling up infrastructure to replace failed cloud servers, don't work as well in a hybrid environment.
Data bandwidth limitations
Another key challenge in hybrid cloud recovery is slow data movement between the public and private parts of a hybrid environment.
When your environment runs solely within a public cloud, and your backups are hosted in that cloud, you can restore data from backups at very high speeds because you don't need to move data over the internet. But in a hybrid architecture, it could take days — possibly even weeks — to move backup data that is stored on your private cloud servers into the public part of your cloud, or vice versa.
Data privacy challenges
One popular use case for hybrid clouds is simplifying data privacy and compliance challenges by making it possible to keep some data on-premises while still taking advantage of public cloud resources.
From a disaster recovery perspective, however, this can complicate recovery plans. You need to manage backup and recovery operations in a way that ensures that data that needs to remain on-prem stays on-prem.
This also makes it difficult to do things such as create a standby environment in the public cloud to replace a failed hybrid cloud because an environment that runs only in the public cloud may not be able to meet the data security requirements of your original hybrid architecture.
Limited backup tool support
While there are plenty of data backup and recovery tools that work well with public clouds, few vendors cater to the hybrid cloud market. Their tools aren't designed to recognize the difference between the public and private parts of a hybrid architecture, or to enable different data storage, retention, and security requirements for each component.
This means that backing up a hybrid cloud, and designing disaster recovery plans for it, requires more effort than simplifying deploying a backup and recovery platform and calling it a day.
Backup and Recovery Optimization for Hybrid Cloud
The above issues don't mean that conventional cloud disaster recovery strategies simply don't work in a hybrid environment. They can be adapted to support hybrid cloud, but it will take more work — and possibly require more financial investment.
You may, for example, need to set aside spare on-premises server capacity to host a recovery environment in case your production private servers fail. Or you might find that you have to establish separate backup and recovery operations for the private and public parts of your hybrid environment in order to support the different security requirements and infrastructure characteristics of each component.
You can back up and recover a hybrid cloud environment. But expect it to require more planning and more investment than backing up a standard cloud environment. Techniques such as pilot light and warm standby can still work in a hybrid architecture, but only if they're modified to support the unique requirements of hybrid cloud.
This story originally appeared on IT Pro Today, a Data Center Knowledge sister publication.