Chances are, you have never heard of Amanda… in the sense of open source that is. And if you have not heard of Amanda, then chances are you have not heard of Zmanda either. I will explain both, and I will give you my view of why it is important for you to at least be aware of these products and their relation to data protection. Whether you should invest in either depends on many factors that will become clear shortly.
Let's start with Amanda. Amanda is the most popular open source data protection product in the market today, at least based on the number of free downloads: 250,000 or more. Like most free downloads, these usually come from universities -- both students and IT folks -- and scientific labs. But, they also include individuals from corporations that are experimenting with open source. In a nutshell, Amanda is a client/server data protection software that runs on a Linux server (backup server) and protects clients that run Windows, Linux or Unix (only a few variants at the moment). It was developed originally at the University of Maryland and then dropped into the world of open source. Since it was distributed to the open source community, hundreds of programmers have contributed to its development, bug fixes and its general care and feeding. As a result, the usage of the product has continued to climb dramatically over the past few years.
But, unlike Linux operating systems (where there are companies like RedHat and SUSE, which is now Novell) or Linux-based databases (where there are companies like mySQL), Amanda did not have a "for profit" sponsor until recently. In late 2005, a newly-formed company was charged with working to make Amanda a more usable product that would be able to support enterprises of all sizes. In keeping with the open source model, Zmanda has grabbed leadership of this space and is feverishly encouraging additional programmers -- some internal to the company, but most belonging to other companies/organisations -- to enhance Amanda so it can effectively compete with Symantec NetBackup, EMC Networker, CommVault Galaxy, Tivoli and others that fall in the enterprise-class data protection software category. Even within the last six months, Amanda has come a long way. But, it also has a long way to go before I would consider it a full member of this class. Should you therefore ignore it? No. However, the reason I am writing this column is to make you aware that, under the right set of circumstances, Amanda is worth considering.
Enter Zmanda. The company has released a specific version of Amanda (two versions, actually) that they support under the classic open source subscription model. You pay only for subscription and support and not for the product itself, just like any other open source product. Of course, the whole idea is to price it such that the total cost of ownership is significantly (as in one-half to one-fourth the cost) lower than other commercial products.
But before you jump into the fray, ask yourself the following questions:
- Does the current product have support for my systems?
- Does it have the features I need?
- Does the product have support for my applications (e.g., Oracle, SQLserver, and DB2?)
- Does it have adequate disk support?
- What about archiving?
I am sure that as you look into these options you will have other questions that are specific to your organisation's needs. Version 2.50 of Zmanda does have support for Windows and Linux, but not for all popular flavours of Unix. It should support databases and other applications in the future but does not right now. It also lacks a GUI and does not yet support all the new innovations that we have seen in the world of disk support (like VTL and CDP). But, it does have disk support. It also has some features that I wish we had in the other commercial offerings, like a non-proprietary data format and like having the ability to do a recovery without requiring the vendor's software. Of course, its Linux support is excellent.
In my view, real innovation occurs when there is a monetary incentive and there is a discontinuity in the technology curve. That is why we have seen the massive transformation in data protection software in the past five years. SATA was the technology that opened up opportunities that just were not available before. But, before that, one could make a pretty reasonable argument that data protection software from all the major vendors had become pretty bloated, and the rate of innovation was very slow. Adding support for a new tape library does not count as innovation in my book. It is precisely at such times, when differentiation between vendors' products is low, that open source starts to make a lot of sense. Thousands of programmers start developing and creating a simpler, less cumbersome product with adequate functionality for many companies that don't need it all. Also, they are cost-sensitive and like the freedom.
That is how mySQL and, of course, Linux itself got going. Now it is Zmanda. But unlike the other segments, data protection is now experiencing phenomenal innovation. So, Amanda's (and therefore, Zmanda's) challenge will be to not only create the old tape-based functionality but also to add all the new juicy disk-based functionality that is coming in waves currently. I suspect it is up for the challenge but at least be aware that there could be a lag before you see all of these features.
It was bound to happen. If database, J2EE, server virtualisation and security tools got an open source counterpart, how far behind could data protection be? If you have simpler needs, cost is a major issue and you crave that freedom from the big vendor -- for whatever reason -- then you should check out this new space. But my advice: Do not run a production environment without the support that comes with Zmanda. Amanda may be free, but she can be trouble without the support.
About the author: Arun Taneja is the founder and consulting analyst for the Taneja Group. Taneja writes columns and answers questions about data management and related topics.
This was first published in July 2006