Full, Differential and Incremental Backups : What Do These Mean?

Those of us who work in the Information Technology industry should be familiar with the terms Full Backup, Differential Backup and Incremental Backup, but for many people outside of the computer industry, these terms are often confusing, as is the term Backup Cycle.

I hope to clarify all of this here, and to help anyone who is looking to backup their computers at home find the best way to do this and to help avoid losing precious data.

A Backup is a copy of a set of files from one location to another. This could be a copy of a whole Hard Drive, or just specific Folders on a hard Drive, or indeed specific File Types (avi, doc, jpg etc) on a Hard Drive.

The different types of backup rely on what is called the Archive Attribute, which is a flag that is set when a file is modified, and which is cleared when a backup is performed.

Good backup software (I highly recommend Cobian Backup) will allow you to state whether you want to perform a Full Backup, a Differential Backup or an Incremental Backup.

Full Backup

This is simply a backup of all files in the set of folders that have been selected to be saved.

Full backups can take a long time to run, especially if a lot of files, or large files are being saved.

Once a full backup has completed, the archive attribute is reset for all files, to indicate that a full backup has been taken.

Differential Backup

A Differential backup only saves those files that have changed since the last Full backup was taken, and is usually fairly fast to execute.

In other words, only those files that have been modified will be backup up. These could have been edited, copied, moved, or could be new files, and therefore their archive attribute will have been set.

Multiple Differential backups are usually taken between Full backups, and as more files are modified, the Differential backup will take longer to execute.

Incremental Backup

Incremental backups work in a similar way to Differential backups, except that after each one, the archive attributes are reset for the files that were included in the backup.

This means that each Incremental backup will only include those files that changed since the last Incremental backup, however, should you choose to save your data using Incremental backups, restoring the data could be time consuming.

Restoring Data From Backups

If the last backup taken was a Full backup, then only the Full backup needs to be restored.

If Differential backups have been taken, then the data needs to be restored from the last Full backup, and then updated with the data from the last Differential backup.

To restore data using Incremental backups, you need to restore the last Full backup, then each of the Incremental backups in turn, beginning with the earliest, since each of these only contains those files which changed since the previous Incremental backup. Tedious eh!

The Backup Cycle

Traditionally it is recommended that three generations of backups be maintained, known as the Grandfather, Father and Son principle.

The latest Full backup is known as the Son. This is the most recent save of a set of data, and in most cases, should data need to be restored, it can be done from this.

Often Full backups will be taken weekly or monthly for example. When the next Full backup is taken, that becomes the Son, and the previous backup becomes the Father.

The cycle repeats, and the Father then becomes the Grandfather. On the 4^th Full backup, the media (disk, tape, flash drive etc) that was used for the Grandfather backup can be re-used.

The process can be expanded to retain as many full sets as is necessary, and companies usually keep monthly backups throughout the year, and annual backups going back a number of years for accounting, tax and audit purposes. This is not necessary for personal computer backups, but having at least two copies in case of problems is useful.

So why keep multiple full backup copies?

Well, assume you save your family photographs to a hard disk or DVD and then find that it doesn't work and have to restore them from your backup. If the backup doesn't work, it's faulty, or if you deleted or corrupted a file before the Full backup was taken, you can't restore it. However, if you retain a cycle of backups, you may find that the previous Full backup has a good copy of the file that you need.

A good motto is to never assume that your computer or hard drive will never crash.

Another good motto is to never assume that your latest backup will be perfect.

What Is A Good Backup Cycle For A Home Computer?

A good backup cycle is to have at least two Full backup copies of your data.

I for example, schedule my backup program to take a Differential Backup weekly for most of my data, since this does not change that often (photographs, music, documents).

I keep 5 Differential backups, and then make every 6^th backup a Full backup.

That means every six weeks I take a full backup of my data, and I also keep 2 Full backup copies. The number of copies is limited by the available space on the external hard drives in my case.

For my email, I save this daily, since the files change every day. I keep 6 Differential copies and on the 7^th day take a Full backup.

Summary

I hope this information is useful and clear to understand.

One of the worst thing that can happen to anyone these days is to lose data on their computer, since much of our lives is retained on there - photographs and memories, financial documents, address book etc.

So don't forget to take backups, and don't forget to maintain adequate backups either.

If you are wondering what backup software to use, I highly recommend Cobian Backup, which is free to use.

Disclaimer: I have no connection with Cobian Backup software. I use it to save my data, and am pleased to recommend it as a comprehensive backup solution.

More Information:

Backup Rotation Scheme (Wikipedia)