Section 24.3. Backup and Archive

24.3. Backup and Archive

The first two elements of data protection are backup and archive. These are related but very different activities. Backup is copying data from one place to another in case the original is damaged. Archive is copying or moving data to long-term storage for quick retrieval of logical components for a specific business purpose. The following comparison between backup and archive clarifies the important distinction between these two activities:

Backups are the secondary copy of primary data.

The purpose of backups is to recover primary data if it's damaged, deleted, or corrupted. Therefore, a backup is a secondary copy of primary data.

Archives are the primary copy of secondary data.

In contrast, an archive is created to be the primary copy of data that is important, but not important enough to be placed on primary storage. This includes old files that are being moved to secondary storage to save space as well as second copies of primary data that are created for a secondary purpose, such as content searching.

Backups recover data that was damaged, deleted, or corrupted.

This is the only purpose for backups. If a file is deleted or damaged, you can restore it from backup copies. Note that in order to perform a restore, you usually need to know the server and filesystem the file resided in and possess the application that created it.

Archives retrieve data from secondary storage.

While the words restore, recover, and retrieve can be used as synonyms in certain contexts, they have different meanings in the context of backup and archive. Restore and recover refer to putting something back in its original condition whereas retrieval refers to obtaining it from an alternate storage location. This is why we restore or recover from backups, but we retrieve from archives, as in "I retrieved the folder from the file drawer."

Archives retrieve data in a manner other than that in which it was stored.

Many applications create numerous completely disconnected pieces of data that are stored in files, emails, and databases. Very few applications are able to search across information stored in multiple applications, filetypes, and files stored over different points in time, but modern archiving applications are meant to do just that. A dynamic archiving system, such as an email system, can monitor and archive all incoming and outgoing email and allow you to search for content across multiple emails, email servers, and points in time. Some filesystem products can do the same for files. In other words, archives can retrieve data in a manner other than which it was stored.

Backups are stored only long enough to cover the usage pattern of the data.

Backups need to be stored only long enough to be able to recover deleted files. How long they need to be stored is based on the usage pattern of the data. If some files are only accessed once a quarter, you should be keeping backups longer than a quarter. If a file is only accessed once a quarter, and you keep backups for only one month, you cannot restore a file that was deleted three months ago and not accessed until yesterday.

Archives can be stored for many years or decades.

Sometimes archives contain information that's been deleted from primary storage and archived in case the information is ever needed again, such as the design plans for a product a company no longer makes. Someone needs to decide when that information can be deleted from the archives. Sometimes archives contain information governed by regulations, and those regulations dictate when that information can be deleted. The result is that archives can be stored for years or even decades.

Document Your Changes

At a large oil company, prior to a company holiday, a backup operator decided to change the next full to an incremental because there would not be a large rate of data change. The admin promptly forgot about the change. Three months later when a restore was needed, it was discovered that they needed to pull from three months of incrementals to do the restore.

David Bregman