What RPO Really Means

In a backup world, Recovery Point Objective (RPO) defines points you can recover to. For example, doing hourly backups will ensure that you can always recover to a state that is no later than an hour before a disaster. This is usually referred to as “one hour RPO.” But this is not enough. What if you need a file that was deleted a month ago? One-hour-old backup will not help: You need at least a month-old backup, or a year-old backup.

Rule #1: RPO should not only define the latest point you can recover to, but also a set of older points to which recovery is possible. 

At the same time, keeping all backups forever is usually not a feasible option. With hourly backups, you would end up having about 8,760 recovery points every year, which becomes unmanageable very soon. It is not possible to keep all the recovery points and some will need to be deleted. However, simply deleting the oldest recovery points, when you reach a manageable limit (usually no more than a hundred), would require deleting all backups older than a week, which is likely not acceptable. A natural effect is that, as time goes by, recovery points do not require the same frequency. Choosing between 2-hours-old or 3-hours-old backup can make a huge difference, but choosing between 268-days or 269-days-old backups likely will not. Therefore, you want to plan to be able to recover to more of recent and to less of older recovery points.

Rule #2: RPO should plan to delete some of your old backups, leaving some other backups live longer. 

Once you determine your RPO, you can then choose the backup scheme that fits it the best. There are several common retention schemes ranging from very simple to quite sophisticated, some of which are natively supported by Acronis Backup & Recovery.

Join me for part two of this series to learn the benefits of different backup schemes.