Incremental Backup

Incremental backups are different from differential backups and full backups.

Depending on the incremental backup software you may be using, it may mean two things:

"Incremental" may refer to backing up only files that have changed. The backup software compares the previous copy of a file to the current, often by checking dates and sizes, and determines whether a new copy needs to be made.
"Incremental" may also refer to incremental deduplication, as processed by BackupChain, Incremental in-file delta deduplication means that after the backup software has determined that a file needs to be backed up, it now checks the contents against the last backup.

This comparison process, called deduplication, may find duplicate blocks from backup cycle to backup cycle. Blocks that haven't changed are not backed up. Blocks that have changed are now compressed and stored in the backup folder, as shown below:

Deduplication

To be efficient, incremental backup software needs to use deduplication and remove duplicate data from the backup set. This greatly reduces the load in the backup media, and also reduces bandwidth when the backup needs to be sent over the wire, as with online backups and remote backups. Incremental Windows Server backup software like BackupChain, hence, needs to do more than just copying the same file over and over again. By scanning for and compressing file changes, the backup solution can achieve a backup storage use reduction of up to 95%.

A Hyper-V backup using BackupChain usually reduces the storage usage by 50% for the first full backup, and the nightly backups are typically 5 to 10% of the first backup. A 100GB virtual machine would, for example, back up first using just 50GB space, then 2.5 to 5 GB per day afterwards.