Backups are the Foundation of Disaster Recovery
No disaster recovery (DR) plan works without backups. If you don’t have a copy of data, you may be able to recover applications, but you’re back to day one of your business. So even before you write your disaster recovery plan, you need a strategy and a process for implementing backups.
Defining Your Backup Strategy
Backup strategies require several major decisions: what to backup, how to backup, when to backup, and where to backup to.
What to backup.
There’s one simple answer: back up everything. You may want to prioritize certain systems as critical and back them up more frequently, but fully recovering from a total outage means you need a copy of everything. Your backup strategy should make copies of data in the cloud, too. Although the cloud provider will keep copies of the latest data, their retention policy may not keep older data around long enough to support non-disaster recovery uses, such as analytics projects.
How to backup.
There are several methods of creating backups. Full backups make a complete copy of all data. This can be time consuming as well as utilize a lot of storage. A differential backup saves only data that’s changed since the last full backup. An incremental backup saves only data that’s changed since the last incremental backup (or the last full backup, if no incremental backup has been done yet). The full backup is the fastest to restore. Both differential and incremental backups depend on restoring the most current full backup first, and then layering the changes on top.
When to backup.
How often you backup has a significant effect on how much data you will lose if you need to restore from backup after a disaster. Each application should have a recovery point objective defining how much data that specific application can afford to lose in an outage. Once you have that number, you can schedule backups that match that requirement. This isn’t always as simple as it sounds, because creating usable backups often requires applications to be shutdown.
Where to backup.
Backups often need to satisfy two needs. They need to be immediately available, so you can access them quickly when they’re needed. They also need to be protected, so they remain available even if the data center is completely inaccessible. The recommendation used to be to keep one copy onsite to meet the immediate need, with a second copy offsite to handle the major disasters. Today, backing up to cloud may handle both needs, though it’s worth considering keeping a copy elsewhere in case network or cloud problems make the cloud inaccessible.