A differential backup backs up everything after the last full backup. This differentiates it from incremental backups, which back up anything after any kind of backup is implemented—full, incremental, or differential.
A differential backup only backs up a file that has been changed by a full backup. If a file has been modified by a differential or incremental backup, the differential backup will look the other way and say, “Nope. You’re not for me.”
What are the advantages of a differential backup? It is easier to restore an entire system when differential backups are used.
Let’s go back to File A. It was backed up about six times during the week, all on different days. We still made a full backup on Monday. However, we now do differential backups during the rest of the week instead of incremental backups.
When the server goes down on Friday, how many backups do we need to use in order to restore everything? Why, just two—Monday’s full backup and Thursday’s differential backup.
Another example:
Monday: Files A + B + C are backed up (full backup).
Tuesday: File A is modified. File A is backed up (differential backup).
Wednesday: File B is modified. Files A + B are backed up (differential backup).
Thursday: Files A and B are modified. Files A + B are backed up (differential backup).
As we can see, we know that a differential backup will always have the latest versions of all the files that were modified after the last full backup.
Why not just forget about incremental backups if it’s a whole lot easier to do restores with differential ones?
Differential backups take up more resources (more tape or more space, depending on what kind of media you are using). If you are trying to back up a large network, the number of media required could multiply enormously with differential backups.
Most administrators use some combination of two out of the three types in order to achieve the most efficient use of resources in their environment.
| 
Disclaimer
