You can never know when a disaster will occur to your cloud infrastructure. You can, however, prepare a cloud disaster recovery plan that enables you to restore operations quickly whenever disaster does strike your cloud. This article explains how.
What Is Disaster Recovery?
A disaster is any type of unexpected event that disrupts your business’s IT workloads.
A disaster could be caused by the failure of a cloud provider’s hardware, such as the data center fire that disrupted service to some Azure customers in 2017. It could be the result of a cyberattack that makes applications unavailable, as another example. It could be caused by the accidental deletion of data by one of your admins.
You can’t know the exact nature of a cloud disaster until after it has occurred. You can, however, prepare a disaster recovery plan, which will help you to restore operations to normal quickly following a disaster.
What Is Cloud Disaster Recovery?
There are multiple ways to implement a disaster recovery plan. One approach is to rely on on-premise infrastructure for backing up your workloads.
The method that provides you with the greatest flexibility and recovery speed, however, is a cloud disaster recovery plan. A cloud disaster recovery plan is one that makes use of a public cloud -- such as AWS, Azure or Google Cloud Platform -- to back up data, applications and other resources. Then, when disaster occurs, those resources can be restored from the cloud back to their original locations -- whether those locations are on-premise infrastructure or the cloud.
Another way to describe cloud disaster recovery is to call it offsite disaster recovery because your workloads are backed up to a remote site and can be recovered from there.
It’s important note that a cloud disaster recovery plan can be used to back up and restore workloads that run on-premise as well as those hosted in the cloud. You don’t have to run your production systems in the cloud in order to take advantage of cloud disaster recovery.
Why Use Cloud Disaster Recovery?
A cloud disaster recovery plan provides several key benefits, as compared to other types of disaster recovery strategies:
- It is more scalable, since you can easily increase the amount of resources that you back up in the cloud by purchasing more cloud infrastructure capacity.
- You can pay as you go. In other words, you pay for cloud disaster recovery infrastructure as you use it; there is no need to invest upfront in hardware or to pay for more infrastructure than you actually use at a given time.
- Cloud-based disaster recovery makes it possible to leverage geographic redundancy features. This means that you can spread your backed-up resources across multiple geographic regions in order to maximize their availability, even if part of the cloud that you use fails.
- Recovery from the cloud is fast because most cloud infrastructure offers high bandwidth and fast disk I/O.
Building a Cloud Disaster Recovery Plan
When it comes to building a cloud disaster recovery strategy, there are two main questions that you need to answer.
Choosing a Cloud DR Provider
The first is which cloud provider to use. The big three public cloud platforms available today include:
- Amazon Web Services (AWS), which offers a diversified system of cloud facilities. Alongside storage and virtual machines, AWS offers database platforms, data transfer facilities, cold and archive data repositories and more. AWS is one of the oldest players at the cloud services market, and it offers significant discounts for long-term subscriptions. AWS’s virtual environment service, EC2, is scalable and cost-effective.
- Google Cloud Platform also offers various virtualization, storage, and big data processing tools, which were developed alongside with such services as Gmail and Google Docs. It allows creating virtual machines with custom resource configurations, thus helping to save on the recovery time.
- Microsoft Azure is another complex service, which has different infrastructures, platforms and software as a service.
Cloud Disaster Recovery Approaches
The second question is which approach to take to offsite disaster recovery. While all cloud disaster recovery plans involve the cloud in some way, the specific types of cloud resources that are used, and the manner in which they are deployed, varies depending on the DR approach you take.
There are four approaches to consider:
- Backup and recovery. This is the most straightforward offsite disaster recovery strategy. It involves backing up data to the cloud and recovering it from the cloud when a disaster occurs. (As noted above, keep in mind that the data could be recovered from your backup location to a local on-premise environment, or to another cloud infrastructure, depending on your needs.) For a simple backup and recovery strategy to work, you need to ensure that it meets your RTO and RPO requirements. In cases where RPO and RTO needs are very high, a backup and recovery approach to offsite disaster recovery may be difficult to achieve.
- Pilot light. Under this approach, you keep a copy of your production virtual servers and databases stored in the cloud at all times. You keep their data and configurations synced with those of the production systems, but the cloud-based backup resources are spun up and activated only in the event that your production systems fail due to a disaster. Thus, the cloud backup environment functions like a pilot light in your furnace. It is always ready to fire up on demand; however, because it will take some time to spin up the cloud-based backup resources, recovery following a disaster is not instantaneous. The trade-off for the delay is that you don’t need to pay to have your backup resources running in the cloud constantly.
- Warm standby. Warm standby is similar to the pilot light approach, except that your backup virtual servers and databases are actually running at all times. This enables you to put the backup resources into production almost instantaneously whenever a disaster strikes. However, this approach will be more expensive, because you have to pay to keep backup copies of your infrastructure running in the cloud at all times.
- Multi-site. This approach entails using the warm standby technique, but instead of having only one copy of your workloads running in the cloud at all times, you run multiple copies that are spread across different geographic zones of the cloud. This strategy provides the greatest level of availability because it ensures that you can restore your workloads very quickly even if part of your cloud provider’s infrastructure has failed. It is also, however, the most expensive approach.
The specific approach that you take when designing a cloud disaster recovery architecture will depend on your needs as well as your budget. If your databases and infrastructure configurations are relatively constant and your RTO demands are not too great, a simple cloud backup and recovery strategy might work. A more costly strategy might be required if you need more availability or the ability to recover more quickly.
Whichever approach you take, however, you will gain a number of benefits by choosing an offsite disaster recovery strategy. You’ll gain scalability and availability, while also likely reducing your overall costs as compared to a disaster recovery strategy that relies only on on-premise infrastructure.