Backing up to the cloud is a great strategy, particularly for organizations that do not have multiple data centers.
There is effectively infinite storage capacity without the hassles of tape media management and sending tapes to off-site storage. Your backups are automatically off site and immune to any failure, disaster or malware infection that may disrupt your network.
But backup isn't the whole story, as cloud-based backups can make restores a slow and frustrating process. Through hybrid backups that combine a local backup cache with a cloud backup, restores are less painful, while organizations get all of the benefits of cloud-based backups.
Restore from the cloud
If you back up only to the cloud, then you must restore from that cloud over an internet connection. This will usually be fine for a single Word document or Excel spreadsheet.
What about larger files, design documents, or sound and video media? Transferring a few gigabytes may take a while, especially if the internet connection isn't as fast as you might like. The real problem comes when a lot of data needs to be restored, for example, after someone makes a big mistake and everybody in the office is waiting for that restore. As an example, 20 GB of restores would completely fill a 50 Mbps internet connection for about an hour. Restoring 150 GB of lost virtual machines (VMs) would take a whole working day, which is quite a large loss of productivity if that is the main file server and nobody can work until it is restored.
Hybrid backups: Back up locally, plus to the cloud
The only way to take the internet connection out of the restore path is to have a backup on premises for the data you are most likely to need to restore.
Because most restores are for issues that occurred in the last 24 hours, you don't need to keep nearly as many local copies as you retain in the cloud. Many cloud backup products perform hybrid backups, as they start by making a local copy of the backup data, so the backup finishes quickly. Then the local backup is replicated to the cloud storage to complete the protection. This local copy used by the backup software will only be the changes since the last backup, so not really enough to protect against bulk data loss. The backup software would need to consolidate a few sets of change data to protect all the files you are working with regularly, which are the files you need restored fastest. This would be a good approach for backups that happen entirely on the computer with limited storage capacity, where there isn't an on-premises backup appliance.
A backup appliance that acts as the bridge between the on-premises computers and a cloud repository is the ideal place to keep a full backup for fast restores. A hardened backup appliance will also protect against ransomware that tries to encrypt on-premises backups. Naturally, the appliance should be getting any new backup data up to the cloud as fast as possible for even better protection.
Restore from local
But getting backups to the cloud is not the problem, as restores from the cloud are more likely to cause a business disruption.
To reduce the time to transfer backups, either buy a faster internet connection or restore within your network. That 150 GB VM that would take over seven hours to restore over a 50 Mbps internet connection would take approximately 20 minutes to transfer over a local Gigabit Ethernet network. Most likely, the storage for the backup would limit the restore speed, rather than the network.
Hybrid backups lead to unified restores
While most restores are for recent data, there are plenty of use cases where older files need to be restored. A prime example is ransomware, which will happily encrypt files that you haven't changed in months. Another is when files are inadvertently deleted, and people don't realize for weeks or months.
To recover some files you will need to use the cloud copy. In many scenarios, the local backup cache may satisfy 90% of your restores, but the cloud may be required to recover the remaining 10%. It is important that the restore tool is aware what data is in which location and chooses to maximize the use of local restores for speed of recovery. This is the reason that you want a single product that does hybrid backups, both local and cloud.