So you’ve decided on some servers that will meet your hardware requirements.  What else do you need to think about?  There is more to storing and interacting with data than just dumping it onto a server.  One major consideration is data backup.  Servers are not fool-proof, and while incidences of major data loss are not common, they do happen.

The first step is determining what your need is regarding backup.  This is ultimately derived from the data you’re working with and how valuable you perceive it to be.  If you’re working with critical data that can’t be obtained again (and you don’t retain the original data), you may need a sophisticated system of near-real-time backups so that you always have a backup of your latest data.  However, if losing data wouldn’t be the end of the world for you and you’re willing to tolerate taking several days to recover lost data, then a very simple, or perhaps even non-existent backup solution is called for.

Once you know what kind of backup you’re looking for, you can start shopping around for options.  In some cases, hardware providers have backup services they can provide (for a fee).  Otherwise, datacenters – companies that will house servers in special facilities – may offer such services if you locate your servers in their facilities.  Additionally, 3rd party backup solutions exist in which vendors might install software on your servers that will back up your data.  Finally, if you have the expertise and time in-house, it is perfectly feasible to establish your own solution.

So what do you look for in a backup solution?  Every backup solution comes with different options and functionality.  At a basic level, a backup system might simply write your data from one hard drive to another one in the same machine.  If one hard drive fails, the other can be put in place to carry on.  An even older method, yet strangely somewhat more secure, is a good old-fashioned tape backup.  This writes data on a regular basis from hard drives to tapes (think cassettes).  This is a cheap method of backup and the tapes can be carried to a safe place to be stored in case they’re needed one day.  The most sophisticated backup is an off-site backup in which all your data is stored in at least two locations – the site of your server and another place often hundreds or even thousands of miles away.  The idea being that if floods, fires, earthquakes, or other such events should destroy the building that houses your server, you won’t have lost your data.   The other major concern of backups is recovery time.  If you’re storing tapes in a safe-deposit box in a bank, it could take days to recover data as someone has to retrieve them, get them into the hands of your system administrator, and he has to load data off them and put everything in the right place.  More advanced methods may use swappable hard drives or pull copies from the internet allowing recovery times to be as low as a few hours, minutes, or even seconds.

Of course, all these options come with a price tag.  The more sophisticated and secure your storage is, the more expensive it’s going to be.  In some cases, depending on how much data you’re storing, the costs could be even MORE expensive than the server itself.  Most backup solutions by 3rd parties are determined based on a per gigabyte rate.  The more data you backup, the more you get charged.  Fortunately in many cases, you can compress backups, reducing the total storage space necessary, but none-the-less, the expense can be pretty great.