When choosing (and running) backup software, it’s useful to understand the main backup schemes so that you can make an informed decision. Choosing correctly will save significant time and disk space.
The first thing to highlight is that when running a backup (especially a disk image backup), very little will change each time the backup is run. Most of your personal files will remain unchanged, and nearly all of your operating system, program and configuration files will remain unchanged. Only a very small percentage of your overall data will change between each backup run, so it’s a good idea to use this to your advantage.
There are three main backup schemes. Full backup, differential backup and incremental backup.
Full backup is what we normally think of when doing a backup. Essentially copying all your data from where it is to a backup location, e.g. an external disk, network or cloud storage. However, when you run another backup on the same data a day, week or month later, you typically want to avoid copying all the data again. Since most of the data is the same, it’s much faster (and uses much less disk space) if we only copy the difference in the data compared to the previous backup.
Copying only the differences between the data in its current state and the last full backup is called a differential backup.
Copying only the differences between the data in its current state and the last backup (full or otherwise) is called an incremental backup.
For example, you may decide to run a full backup on a Sunday and differential backups each day from Monday to Friday.
Sunday (Full backup)
Monday (Difference: Sunday to Monday)
Tuesday (Difference: Sunday to Tuesday)
Wednesday (Difference: Sunday to Wednesday) …
Conversely, if you run the same backup schedule using incremental backups, it works as follows.
Sunday (Full backup)
Monday (Difference: Sunday to Monday)
Tuesday (Difference: Monday to Tuesday)
Wednesday (Difference: Tuesday to Wednesday) …
Each day, the differential backup becomes larger since each new backup has all the changes from the previous day as well as all the cumulative changes since the last full backup on Sunday. However, the incremental backup does not get larger each time it is run, since it only ever has the changes from the previous day. Differential backups get slower and larger each time they are run, whereas incremental backups remain relatively fast and small. This is why, if you have incremental backup available as an option, you should use it.
When you need to restore backed up files, this works differently with differential and incremental backups. For differential backups, your data is recreated by combining your last full backup and the relevant differential backup. Using the example above, if on Friday you want a file from last Tuesday, it will combine last Sunday’s backup with Tuesday’s backup (which has all the differences from Sunday to Tuesday). However, for an incremental backup to do the same thing, it would combine Sunday’s full backup with the two incremental backups run separately on Monday and Tuesday, since each incremental backup only contains the differences for a single day.
This means restoring data from incremental backups will take a little longer since it needs to merge multiple differences together, but in practice this delay is rarely noticeable, and somewhat inconsequential when compared with the benefits.