Understanding the types of backup and why you need them

Backups are safety copies of your data, and they come in several kinds. Specifically, backups are the data on a computer or server kept on another system or medium (another disk, a NAS, the cloud), so that you can restore it if you lose it: a failed disk, an accidental deletion or, worse still, data encrypted by ransomware such as Cryptolocker. We’ll look at the types of backup one by one, and then at why every company needs them.

Types of backup

1. Full backups
2. Incremental backups
3. Differential backups
4. Synthetic full backups
5. Cloud backups
6. Other types of data protection (RAID)

Full backup

A full backup is the best-known way to back up data: it copies all the selected data. As a rule, every backup scheme starts with a full backup. If you choose to run only full backups every time, keep in mind that the method is inefficient, because it uses a lot of space, hardware resources and time, which can slow down the computer or server. After the first full backup, you can choose an incremental, differential or synthetic full backup.

Incremental backup

An incremental backup copies only the data that has changed since the last backup, whether that was a full backup or another incremental one. Each such backup therefore contains only the changes since the backup before it.

Incremental backups take up less storage space and put less load on the network, which is why they can run more often than full backups. To restore, however, you need the last full backup and every incremental backup made after it.

Differential backup

A differential backup is the type of backup that copies new and changed files since the last full backup. This kind of backup compares data only with the last full backup.

What to know about differential backups:

• They’re much faster than a full backup, and two sets are enough for a restore: the last full backup and the last differential backup.
• Each differential backup keeps growing until the next full backup and can get almost as big as it, which is why a new full backup is needed periodically.

Differential backups are very common for Microsoft SQL Server databases and other applications that use databases.

Synthetic full backup

A synthetic full backup is a full backup built on the backup server from the last full backup and the incremental backups made after it, without reading all the data from the source again. A synthetic full backup reduces the amount of data transferred and speeds up the creation of a full backup.

It’s the ideal kind of backup for large files or when the connection between the source and the backup is slow.

Cloud backup

Cloud backups are kept with an external provider and can be accessed remotely, online. When we wrote this article, in 2020, they seemed like the future; today they’re a natural part of any backup strategy, because they provide the off-site copy required by the 3-2-1 backup rule. Many companies offer cloud storage where users can safely keep copies of their personal or office data. As a rule, such services aren’t free.

RezolvIT’s Backup as a Service backs up data both on the network and in the cloud.

Is RAID a type of backup?

No. RAID protects data only against a disk failure; an accidental deletion, ransomware or a fire affects all the disks at once. Even so, many technicians confuse RAID with backup, especially RAID 1 and RAID 5, two of the most widely used variants. The full explanation is in the article why RAID isn’t backup.

RAID 1 (mirrored disks) needs 2 hard disks, and RAID 5 at least 3. Both are often used in NAS devices (Network Attached Storage).

As “mirrored” suggests, RAID 1 means an exact copy of the first hard disk on a second hard disk. Advantages: if one of the disks stops working, the data can still be accessed from the second one, and it’s copied automatically to the new disk when the faulty one is replaced. Disadvantages: two disks offer the storage capacity of just one (unlike RAID 0, where the capacities of the two disks add up, but without any protection).

Advantages of RAID 5: parity information is distributed across all the disks, and the array keeps working if a single disk fails: the missing data is calculated from the parity, so nothing is lost. Reads are fast, and only the equivalent of one disk is lost from the total capacity. Disadvantages: writes are slower than with RAID 0, because of the parity calculation, and if a second disk fails before the first one is replaced, the data is lost.

Why does a company need backup in 2026?

Because data can be lost at any time, and a company without backup can lose days of work, or even data it can never recreate:

• Ransomware: attackers encrypt the data and go after the backups first, to destroy them. That’s why you also need an offline or immutable copy.
• Human error: files deleted or overwritten by mistake.
• Hardware failures: failed disks and servers, lost or stolen laptops.
• Cloud applications: even data in Microsoft 365 or Google Workspace isn’t fully protected without a separate backup. See why you need backup for SaaS applications.
• Legal obligations: for companies that fall under NIS2 (in Romania, GEO 155/2024), backup management is mandatory, and the GDPR requires that personal data can be restored in a timely manner after an incident.

Whatever type of backup you choose, test restores regularly: a backup that can’t be restored won’t help you.

If you’re not sure what backup you have and how quickly you could restore your data, request a free IT audit.