Open source backup software has almost no profile in the market compared with mainstream backup applications, but it does offer a robust alternative selection of products to the fully commercial offerings.
Open source means that the source code is free and can be altered and copied at will as long as all changes are published. While there are variations in practice -- with some open source backup products, including commercially licensed code -- the basis of all open source products remains the same.
Open source software now runs in almost every enterprise. The most visible example is the Linux OS, which is most prevalent in the data centre. Open source applications are less visible than Linux, but available applications include backup.
Main open source backup software packages
Open source backup software can be used by organisations of all sizes, so the criteria for using it are whether a product provides the functionality required, a degree of commitment to the open source model and a willingness to step outside the comfort zone that a commercial vendor offers (although paid-for support for some open source software packages is available).
Tony Lock, programme director at research and analysis firm Freeform Dynamics, said, "Its use tends to be confined to those with sufficient IT skills to not require the addition of support services and who are happy to update things themselves."
Market share information for free open source backup software is not available as it is for fully paid-for products. However, open source apps directory SourceForge tracks downloads, and its stats show that Bacula's 5,523 weekly downloads outstrip both Amanda's 1,536 and BackupPC's 1,102 downloads.
Bacula was developed by the co-founder of Autodesk, Kern Sibbald, who has dedicated himself to developing and maintaining the package. It stores data in its native format and offers enterprise features such as scheduling, automation and backup tracking, as well as media management. It does not include database agents, and it can be complex to set up.
Amanda (Advanced Maryland Disk Archiver), on the other hand, is the oldest open source backup software package. In 2007, a new distributor -- Zmanda -- retooled it into Amanda Enterprise. The new version stores data in a non-proprietary format, automatically determines when full backups are needed and is the only open source backup application that includes agents for a wide range of databases.
BackupPC is a simpler, file-level backup package, suited mainly to smaller organisations. It supports both disk and tape, scheduled full and incremental backups, and can store those backups in a backup format such as Tar. BackupPC stores its backups in a snapshot-like tree structure and stores duplicated files as a links to a single copy of the file.
As an alternative to these more fully featured applications, you can also roll your own script-driven backup application, using open source tools such as Rsync and Tar to do the heavy lifting. This allows you to tailor backup routines to very specific requirements.
Acquisition is a matter of downloading the software either from the developer or support provider's website, in the case of commercially supported applications, or from SourceForge, in the case of free software. For Chris Puttick, chief information officer at Oxford Archaeology, updating Bacula is simple and involves downloads from online repositories.
Open source backup software: Benefits and drawbacks
The core benefits of open source backup software are similar to those of all open source software. There is no licence fee so acquisition costs are zero and the code is freely downloadable so if a product does not include a desired feature you can develop it yourself.
Community development also implies community support, although all major open source backup developers also offer a paid-for support programme, which is often their main revenue stream.
For example, Amanda Enterprise comes with a three-tiered subscription model. Each element, such as backup server, client or database agent, attracts a separate charge. Prices range from $100 per client with basic support to $1,500 for a virtual appliance configured as a backup server with top-level (Premium) support. Quoting Sun figures, the company projects a total cost of about $4,000 for Amanda Enterprise for one Linux backup server, 15 Windows/Linux/Solaris backup clients, disk-based backup support to 1 TB, plus support for a tape library with two drives and 40 slots. This contrasts with costs of between $18,000 and $19,000 for comparable Symantec NetBackup and EMC Networker installations.
Device support is not an issue, especially since the rate at which new backup devices arrive -- requiring new drivers -- is not high. Puttick said, "Bacula is the most broad-minded in terms of backing up remote devices such as those in branch offices."
Open source backup carries a few drawbacks. For instance, an IT organisation might need a higher level of technical knowledge to fine-tune the configuration than with commercial software, where there tends to be greater development effort put into GUIs and usability.
Puttick said, "We have constrained ourselves to using the version of Bacula that comes with the version of Ubuntu (8.04) that we are running. Bacula's GUI is not very good though, so to do any configuration you have to go back into the config files." Puttick does not, however, see this as a major drawback. For him, the stage of planning and scheduling the backup tasks is the major time cost. "I don't think there's a huge cost to not using the GUI. . . . Once you've worked [the schedule] out, the software isn't the hard part," he said.
The feature sets of open source backup can be smaller than those of commercial software too, but, Puttick said, "The feature sets probably look smaller -- but do you really need those features?"
For Lock, one drawback of open source backup software is identical to that of commercial software: "Few organisations are happy with how effectively they can recover data when they need to, irrespective of what software they utilise," he said.
Free open source backup software is a match for mainstream packages provided the technical expertise exists within the organisation to dive into the configuration files and tweak as necessary. Unsupported versions offer the advantage of being free to download, as well the ability to adjust the source code to meet specific requirements. However, the key requirement for using open source backup is arguably your willingness to step outside the comfort zone provided by commercial software vendors.
This was first published in October 2010