Backup Storage

There is no immediate benefit to companies from a backup - it is only when critical data has been deleted or corrupted that a backup is worth its weight in gold.

This allows them to quickly fix errors, maintain operation, and avoid data loss.

The loss of business-critical data places a financial burden on companies, damages the corporate brand and, in the worst case, leads to a company having to close down permanently. It is not uncommon for data to have to be restored in a time-consuming and cost-intensive manner, which in some cases results in high claims for damages from customers or suppliers.

Data loss or data corruption is a risk in every company despite high protection measures.

What are the most common causes of data loss?

• Human error by employees due to accidental deletion or overwriting
• Hardware defects, for example, due to damage, short circuits, or a natural disaster
• Cyber-attacks, which today primarily aim to encrypt backups before a ransom demand is made
• System failures due to a power outage, for example

Data is essential for businesses to run smoothly. A backup with the right protection mechanisms saves companies from losing essential data and enables a quick return to normality after an unplanned event.

A backup consists, simply put, of two basic steps. First, the backup software manages which data from which applications should be copied when and how often. In the second step, this data is written to a storage target where the data is stored separately from ongoing business operations.

However, a simple copy of the primary data is not sufficient as a secure backup. The risk is high that data will also be lost here, encrypted by ransomware, or modified.

How to keep backup data safe?

Software-based storage solutions which flexibly adapt to future technology developments are ideal for this purpose. They protect data on the storage medium from changes, maintain access rights, replicate backup data and thus offer all-round protection.

A proper data backup takes place at regular intervals - all backup types have this in common. Which backup type is best for a company depends on various factors:

• How much time can backup storage take?
• How high can the level of complexity be in order to save storage space or increase security?
• Which backup type disrupts operations as little as possible while providing high data security?
• How often does the data change and to what extent is new data added to the existing?
• What storage capacities are available?

There are three basic ways to perform a backup: full backups, incremental backups, and differential backups.

Full backup

In a full backup, all the data of a system is copied. The great advantage of a full backup is that data can be restored quickly, because it only needs to be restored from a backup copy.

Advantages:

• Fast and easy data recovery
• Better memory management, because the entire data set is stored in a single backup file

Disadvantages:

• The implementation takes a lot of time
• Large storage capacities are required for storage
• Storage space requirements increase with each full backup

Differential backup

In a differential backup, all changes since the last full backup are backed up. The basis is a full backup, then only data which has been changed or newly created since the last full backup is backed up. Only two files are required for recovery: the current full backup and the current differential backup.

Advantages:

• Smaller amount of data compared to full backup
• Time and storage space required for data backup are reduced
• Easy data recovery, because only two backups are required - the full backup and the last differential backup
• Backup states can be deleted independently of each other, because each differential backup refers to the last full backup

Disadvantages:

• For data recovery, the full backup and the last differential backup must be imported. The time required for recovery increases.
• Both parts, the full backup and the differential backup, must be present and error-free for a full restore.

Incremental backup

In an incremental backup, after an initial full backup, only the files which have changed or been added since the previous backup are backed up. The previous backup can be a full backup or an incremental backup.

Because an incremental backup always refers to the previous backup copy, the backup files are closely chained together. After a specified number of incremental backups, the cycle concludes with a full backup. Restoration is only possible in blocks (full backup and all subsequent incremental backups).

Advantages:

• The amount of data is less than with differential backups
• It takes less time to perform than full or differential backups and uses less disk space

Disadvantages:

• The data must be imported chronologically for a successful recovery
• Because the last full backup and the entire chain of successive incremental backups must be restored, the recovery takes longer
• If an incremental backup in the chain is missing or corrupted, a full restore is not possible

How often data changes and how quickly it needs to be restored determine the intervals at which backups should be created. The goal is to minimize the amount of data which can potentially be lost and to have as few backups as possible from an economic perspective. To achieve this, companies set two target parameters for backups:

• How much data loss can a company risk from a system failure? Recovery Point Objective (RPO) refers to the maximum amount of data a company can lose. This value determines how much time may elapse between two data backups. The longer the period between two backup copies, the greater the data loss which is possible.

• Recovery Time Objective (RTO) defines how long a company can manage without IT system and infrastructure after an emergency without endangering business operations (business continuity). RTO thus refers to the time span between the time of damage and the recovery of business processes. The RTO value determines how quickly the systems and infrastructure must be restored.

To prevent extortion attempts from missing the mark, cyber criminals often target the last line of defense of businesses and organizations - the backup. Today, ransomware is personalized, professional, and targeted, primarily targeting companies' backup data.

On the company side, an attack on the IT system can never be ruled out. It is impossible to find and close all security gaps within the IT infrastructure.

Backups are not used to detect and directly defend against ransomware. They are insurance in the event of an attack. Backups can make life difficult for ransomware and, in the event of an attack, restore the valid data to replace or restore infected data. This is why it is important to store backup data on a secure storage target.

But how can companies prevent ransomware from infecting their backup storage? What measures help against the increasing threat of ransomware?

WORM storage

WORM (Write Once Read Many) storage is a highly effective strategy against ransomware - data cannot be modified once written (even by employees) and malware cannot encrypt or delete the data.

WORM is a data backup method in which data is written unalterably to a storage medium. Reading is possible as often as required.

A common method to implement WORM storage is S3 Object Lock. The S3 Object Lock API provides an implementation which enables granular WORM policies and adds additional protection against object modification and deletion. This is not only possible in the cloud, but also S3 compliant in your own data center.

Versioning is a possible addition to Object Lock. Versioning is the storage of file states (versions) after a file has been modified. Each new version of a file is stored so that users can restore previous versions of a file at any time. With Object Lock, these versions are protected according to the WORM principle as soon as they are placed on the backup storage target.

3-2-1-1-0 backup

Because even the backup itself is not immune to data loss, a solid backup strategy is fundamental. One backup copy is not enough. Even the backup can be damaged or destroyed - companies always need more than one copy and must protect them especially.

Why?

If there are multiple copies in different locations, it is much less likely that all backup copies will be lost, corrupted, or encrypted by ransomware at the same time.

The 3-2-1 backup rule is a well-known and simple rule, and still forms the basic framework of a solid backup strategy:

• At least three data copies are to be kept. In addition to the originals of the productive data on the primary storage, the other two data sets serve as backups.
• Data should be stored on at least two different media.
• One backup copy should be kept in an external storage location (offsite or geographically separated location).

The important thing with this strategy is that at least two storage targets are physically separated from each other in case of a fire or natural disaster, for example. Nowadays, this can even be fulfilled within a storage system, such as in a scale-out cluster with asynchronous replication.

For higher protection, two items can be added to the 3-2-1 backup rule: an offline, airgap, or immutable copy and zero errors in the backup process when restoring the data.

Offline means that there is no longer a connection to the infrastructure. If a hacker has access to the infrastructure, all data with online connection can be affected. An offline backup copy means that a hacker no longer has access to it. Such an offline backup is also called an airgap backup.

This category also includes immutable backups, which are protected against changes or deletion thanks to the WORM principle, for example. A common method for this is the use of S3 Object Lock.

The 3-2-1-1-0 strategy goes one step further and requires that backups are stored and restored without errors. There should be zero errors during recovery - backups are too important for this. To achieve this, companies must perform recovery tests at regular intervals and check that everything is working as it should during the recovery process.