Making a backup is a critical task for any technology team or department (and should be an important task for all computer users). However, for new technicians or server/systems administrators the range of options to choose from when creating a backup solution for an office or organization can often be bewildering in its options. This three-part series aims to give you a general background of the basic concepts related to server and network backups. In this article, we'll look at the various types of backups that can be performed using most commercial backup applications as well as the question of whether or not to back up data stored on other computers using a centrally managed backup server.
Backup Types
Although a backup is easily defined as a second copy of data to be used in case the original fails, there are three basic types of backups that determine which pieces of data from a backup selection (the hard drives, folders, or selection of files chosen to be backed up) are copied in a given backup operation. Each has situations in which it may be more appropriate; each has its own advantages and disadvantages. This section provides an overview of the three in addition to concept of data archiving.
Complete Backups
A complete backup , sometimes called a full backup or an archival backup , is a backup that creates a copy of all files within the backup selection. Complete backups are the least commonly performed backup types because they are the most time-consuming and because they include all files regardless of whether the files have been updated since a previous backup (or if they have ever been updated). You can perform complete backups by using backup software, by using scripts, or simply by manually copying files—making them the simplest of all backup types.
You typically perform a complete backup following the installation of a server or creation of a share point to establish a baseline copy of all the data it contains. Beyond that, you perform complete backups only on a very limited basis and rely on incremental or differential backups because they are less time-consuming and resource-intensive. They are also preferable because most backup software enables you to append them to a complete backup, meaning that you have access to the original files as well as subsequent backups.
Generally, you perform complete backups only if there is a major change in the data source or the backup schedule. If you use tape media, you might perform backups more regularly when replacing or recycling tapes. The only other times you might want to consider a complete backup is prior to a server upgrade (to give you a snapshot of the server and data before the upgrade in the event of a problem) or if you find that appended data has maxed out a fixed media drive or tape set and there is no need to maintain the various states between the current data and earlier backups.
Incremental Backups
An incremental backup (sometimes referred to as a cumulative backup ) copies all files within the backup selection that have changed since the last complete backup. This saves time and resources because only files that have changed are copied in each backup operation. Incremental backups can be performed so they overwrite the original backup data or append the changes to existing data. Generally, the latter is preferred because it gives you access to the original state of the data, if required. Although it depends on the backup tools you are using, an incremental backup often overwrites previous incremental backups, leaving you with the data from the current incremental backup and original data from the last complete backup, with none of the data from the incremental backups between them. Incremental backups are generally less common than differential backups because each backup operating compares data only to the original data state, requires more resources, and leaves out the data states between the current and original states.
Differential Backups
A differential backup (sometimes confusingly called a differential-incremental backup or erroneously called an incremental backup ) copies only the data that has changed since the last backup operation of any kind (complete, incremental, or differential). This backup uses the fewest number of resources and is more efficient than other backup types. Differential backups typically append changed data to existing backup data, giving you multiple "snapshots" of the data contained in a backup selection at each backup operation.
The fact that differential backups offer you these snapshots might be an even greater advantage than the conservation of resources. It enables you to locate a file that existed in the backup selection at any point between the current backup and the last complete or incremental backup. You can thus locate and restore files that have been deleted, altered, or corrupted at any point during the backup history. Although it can be required because of a system problem, it is often required because users have accidentally deleted or modified files, and need to recover them (occasionally because another user might have done something to their files, which should have been prevented by proper file permissions and security). You don't want users to get in a habit of relying on your backups instead of keeping track of their own files. But if something happens to a critical file (or to files belonging to the president of you company, for example), it can be a great advantage to be able to recover those files.
Although differential backups offer some advantages of convenience, they also require more storage space than complete or incremental backups. As such, you might want to consider planning your backup strategy to include other backup types periodically. In an educational setting, for example, you might want to perform complete backups at the end of each semester or school year, thus reducing the storage space for your backups at a point when the downtime required for a complete backup will have little or no impact on users.
Daily Backups
Some backup tools offer an option called a daily backup , which is typically a variation on the concept of a differential backup and, as the name implies, it is a backup operation that backs up only files modified on a given day. In most applications, a daily backup is simply a short-handed way of configuring a daily differential backup to run. As such, the advantage is in the ease of configuration for the technician or administrator configuring the backups more than in the actual operation of the backup (although it might vary depending on the backup tool used).
Data Archiving
Strictly speaking, data archiving is not the same thing as a backup. When you archive data, you remove data that users no longer need to access from a server (or workstation), but you create a copy of it for later reference in case it is ever needed. This frees up storage space on the server that was hosting the data, but still enables you to have access to it if a user does require it at some point. Data archiving can be done on a periodic basis or as part of a cleanup of storage space on a server (in which case you might provide a copy of the data to the user or users that created it). One common example of data archiving in an education environment is archiving of student projects for students who have graduated. Typically, you use very reliable media for a data archive because you might make only one or two copies of the data, which will not be further backed up in the future. As a result, optical media is often chosen for data archiving.
Workstation Backups
Although backup types are appropriate to back up applications that run on servers as well as workstations, there is a distinction for workstation backups in many server/network environments. Typically and ideally, in a server/network environment, users rely only on a network home directory or share points for storage. This is a good approach because backing up the servers that host those share points and directories ensures that all user data is safe from potential data loss. In the real world, however, there are many situations in which users will store data on the hard drives of workstations (media professionals needing faster access to data than is available from a share point or users with mobile accounts, for example). In these situations, there can be a blurred line about whether it is your responsibility to ensure the backup of this data.
On the one hand, your users should take responsibility for any data not stored on a server. On the other, the powers that be may feel it's the responsibility of the IT department to make sure that users either have perfectly reliable computers (which we all know will never happen) or take provisions to secure files on the not-so-perfectly reliable computers. If you can set up a situation in which users are responsible for data on individual computers (either backing that data up themselves or for copying it to a server for backup), by all means do so. Whether you can or not, encourage users to copy data to their network home directory or a share point as much as possible.
If you are in a situation in which technicians or administrators have to take responsibility for workstation backups, however, it can be reassuring to know that there are client-server backup tools out there. Such products include a backup client that installs on a workstation, which can then perform backups of either the entire workstation of specific folders or hard drives. Being server-based, you can configure automated backup sets with these tools, which can greatly reduce the headache of workstation backups because the process is managed from a central backup server and doesn't rely on user or technician interaction at the workstation level.
There is also the option to use backup software that is installed on each workstation to back up to a share point hosted by a server. This is as valid as using a client/server backup tool, but it has two downsides. First, it requires configuration of the backups at each workstation rather than from a central tool. Second, depending on your choice of backup application, it can be significantly more expensive to license the application for each workstation you need to back up than to purchase client/server licenses.
Next Up
In part two of this series, we'll look at the various backup media that can be used to store backups. There is a wide range of options out there that can serve the needs of a small office or a corporate data center—and everything in between. We'll take an overview of all the options and discuss the most commonly used, particularly those that can be confusing for new technicians and administrators.