The Four Methods of Server Backup

Posted by Ben Austin on Jan 6, 2015 6:30:00 AM


The following is an excerpt from our recently published eBook, The Big Book of Backup. If you'd like to download the full eBook click here.

The objective of the vast majority of backups is to save data without errors or corruption in a way that minimizes storage space. Additional objectives also often include reducing the impact on computing resources and bandwidth usage, while making the recovery process as quick and easy as possible.

While individual systems may boast some additional features, all backup methods share many of the same components. The following sections explain the four most common backup methods in use today, and highlights the pros and cons associated with each method.

1. Full Backups

A full backup stores a copy of all files and typically occurs automatically according to a pre-set schedule. Files are usually compressed to save space, however, even when compressed full backups may consume a lot of storage. Additionally, full backups cause heavy access to the backup disk, which shorten disk life and consume network bandwidth.

The advantage of full backups is the ease of restoration. Restoring a file requires only the file name, location, and date from which to restore the data. Restoration is relatively straightforward as long as the backup files from that date or time are available.

Although full backups are certainly comprehensive, they may be more robust than many businesses require. It is important to consider that only a small percentage of files change from one backup to another. Consequently, performing full backups will yield multiple identical copies of files and consume valuable storage space on the backup medium.

2. Incremental Backups

Incremental backups save space by backing up only the files that have been created or changed since the last backup. The advantage of incremental backups is that the volume of data backed up at each iteration is much smaller, which in turn saves space on the backup medium and uses less network bandwidth.

However, incremental backups increase computing overhead, because each source file must be compared with the last full backup as well as the incremental iterations to determine whether data is new or changed. Additionally, it is more complex to locate a specific file to restore as it may require searching several iterations. To completely restore all files requires merging all iterations while taking care to keep only the most recent version of each file.

Many enterprise backup strategies include a combination of full backups and incremental backups. For example, running a full backup once per week—on weekends when network and computing resource demands are lower—and scheduling incremental backups on weekdays. Backing up files with this combination enables a restoration that does not require looking through or merging more than a week’s worth of iterations.

Some strategies limit the impact on disk backup storage by copying older full backups from disk to tapes, which are then stored off-site. Although this approach is more secure than storing both the file system and backup media at the same location, the manual work to change tapes, label them and transport them is time-consuming. Additionally, it creates a difference between the finished backups that are stored off-site and the current state of the live file system. In the event of a disaster, any data changes that occurred since copying the last full backup to tape may be lost.

3. Differential Backups

Differential backups are similar to incremental backups, except that each backup operation stores the new and updated files since the last full backup. For example, if a full backup was performed on Sunday and a file changed on Monday, that file will be part of every differential backup until the next full backup is run.

Using differential backups simplifies recovery because only the last full backup and the last differential backup is needed to create a complete restoration. As with incremental backups, differential backups need to compare current and already-backed-up files to identify any changes. However, differential backups require more space and network bandwidth compared with incremental backups.

4. Virtual Full Backups

Virtual full backups use a database to track and manage backed-up data, which helps avoid some of the pitfalls of other backup methods. A full copy, or replica, is taken only once and does not need to be taken again as long as the storage medium—typically a network-attached storage location—remains unchanged. The virtual full backup periodically synchronizes backup data to the database.

Virtual full backups are generally performed automatically by backup software. The user experience appears the same as that of a full backup. Restoring one file or an entire disk is a matter of choosing a preferred recovery point and the file or files to recover.

Want to read more? Click the button below to download the full eBook.


Download "The Big Book of Backup"

Topics: Backups, virtual backup, server backup

Recent Posts

Posts by Topic

see all