There are three main types of backup: full, differential and incremental. And the devil, really, is in the names. Full backup means that you are performing backup of all files each time; incremental backup will only do new and changed files or parts of files since the last backup. Differential backup will also upload new and modified files or parts of the files, but it compares itself with the full backup.
We have a separate article on incremental backup, and now we’re going to explain how differential backup works and focus on the difference between incremental and differential backup.
What Is Differential Backup?
It does not take much time to realize that doing full backups each time is not an effective backup strategy. It might be reliable and easy to setup, but it will consume more time and more space than… Than what, actually?
Here comes differential backup - the first thought in optimizing the backup scenario. It uploads only the new and changed files since the full backup. This type is significantly faster than full backup to perform. However, each time you perform differential backup it backs up data that differs from the original source data even if it wasn’t modified recently.
Avoid performing differential backups too often, because the amount of the backed up data grows with each operation. As a result, all differential packs combined could require even more storage space than a full weekly backup. What other problems will arise ahead?
Data recovery will be slower, than recovering from the full backup, because the software needs to combine the last differential and the full backup together.
So, both of these backup types are performing backup of changed and modified files only. However, there is a distinctive difference between incremental and differential backup.
On the scheme below you can see how differential backup works. The setup is given:
- We have 10GB; as an initial dataset to backup.
- Between each consequent backup, our data set grows for 0.5GB.
- We will perform 3 differential backups.
- In the end, our data set will be 13GB.
This is how differential backup essentially works. It only uploads the difference since the full backup.
And here is a scheme for incremental backup. The setup here is a bit different:
- The initial data set is, again, 10GB.
- The data grows on 1GB during the first period, on 0.5GB, during the second period and on 1GB during the third period.
- In that case, in the end of the third period, there will be 12.5GB of data on your storage.
With the very same setup differential backup in the end of the third period would have 15GB worth of data (10GB+1GB+1.5GB+2.5GB).
The difference between incremental and differential backup, hence, is obvious:
- Incremental backup is faster than differential.
- In the long run, differential backup will be more storage consuming, than incremental backup.
It might seem that it is a clear victory of the incremental backup. However if you lose any incremental, except the last one, in the backup chain - you won’t have an ability to recover your files. If you lose any differential in the chain - you will be able to recover any other differential backup (only the difference will be lost).
Both incremental backup and differential backup options might be way more efficient than performing full backup each time. But why might be?
As you’ve already seen, there are several issues that you can face while using these backup types without planning. Left on their own, incremental and differential backups will soon become less effective and more storage and time-consuming than you would want.
The main idea behind any backup processes, really, is planning. There is no “one button solution” in the backup industry. And if any vendor claims otherwise - know that you will be left with few options to choose from and most probably you will have the only option - to adapt to the solution. With CloudBerry Backup we are providing you enhanced and flexible instruments, so you can create a backup, perfectly suitable for your needs.