In case of a disaster, a traditional cloud backup application needs to transfer all data across limited internet bandwidth and back to the data center or the new location. The bandwidth restrictions make cloud recovery a time consuming process that most organizations can’t endure. The technology that helps the backup process like incremental backups and deduplication won’t help. DRaaS attempts to solve the cloud backup recovery challenge by eliminating the time pressure associated with the transfer. The organization can, at least temporarily, use the cloud provider’s compute resources to recover applications in the cloud. In-cloud data transfers complete much faster than internet transfers do.
Many DRaaS vendors claim to deliver push button disaster recovery but few explain what actually happens when the button is pressed. What occurs in the cloud after the organization declares a disaster dramatically impacts the speed and quality of the recovery effort. IT planners need to understand that push-button DR does not mean the recovery is complete, it just means that a series of steps have triggered.
Step 1: Data Preparation
How the data protection application sends data to the cloud directly impacts how the cloud provider recovers it. Storing data in a proprietary backup format requires recovering data from cloud backup storage to cloud primary storage. The internal transfer, within the cloud, completes much faster than a WAN transfer does but it still consumes valuable recovery time.
The data and virtual machines in the backup may also require transformation to run under that provider’s hypervisor. Transferring the data to production quality storage and transforming it to run in the DRaaS vendor’s hypervisor adds to the total recovery time.
Some backup applications and most replication applications store data in their native format and cloud providers may store backup data on production quality storage, which reduces the transfer times but increases costs. Cloud providers may also run VMware or Hyper-V, the exact hypervisor that the organization uses, which eliminates transformation times. Requiring the provider to run the same hypervisor as the organization will limit the number of potential cloud providers from which the organization can choose.
The organization needs to decide how quickly they need to recover, typically prioritizing certain applications. In many cases, it makes sense to store the applications and their data in different cloud storage types based on recovery expectations.
Step 2: Networking
With the organization’s applications running in the provider’s cloud, IT can enable users to login to the applications and get back to work. Since the application now runs in the provider’s cloud, the network routing needs to change. IT must carefully plan network failover and re-routing. The likely difference in network and backup team members means the team in charge of data protection and the team in charge of the network. The network team may not have experience with this type of failover. The organization needs to make sure that either they have that expertise or they need to work with a cloud provider that does.
Step 3: Exit
In most cases, the organization will want to resume operations in their original data center or in a new one post-disaster. They will not likely choose to continue running applications in the cloud. Exiting a cloud provider often means incurring egress charges while data transfers to the primary data center.
DRaaS takes the initial pressure off cloud restoration since applications start in the cloud and data transfers within the cloud. However, IT needs to plan for a cloud exit. IT planners should look for providers with low or no egress fees when recovering from a disaster. They should also look for providers that have ability to bulk transfer recovered data via an appliance or tape.
DRaaS, on paper, looks like the ultimate cloud use case. It eliminates secondary DR site costs and facilitates better testing. As IT planners examine it more closely though, they see challenges to overcome. When deciding on a cloud data protection solution, organizations need to understand what happens after pressing the disaster recovery button to ensure that the provider meets the organization’s expectations.
Understanding the cloud provider’s capabilities is only one step in creating a great disaster recovery plan. To learn more about creating great disaster recovery plans join us live on August 22nd at 1:00pm ET / 10:00am PT for “How to Create a Great Disaster Recovery Plan.” During the webinar experts from Storage Switzerland, Veeam and KeepItSafe discuss the various components of a disaster recovery plan and how to fine-tune them to meet DR expectations.