Data backup primer
It is essential to understand the importance of data backup. A key to developing this understanding is to learn the terminology and criteria used to define critical data that needs to be backed up.
There are three main options for backup:
- Local tape backup
- LAN backup to a central server
- E-vaulting backup (or online backup)
LAN backup solutions allow all systems attached to a LAN the ability to backup to a central server on the LAN.
Electronic Vaulting (e-vaulting) can be described as the process of electronically transferring data from your local system (a personal computer, server or workstation) to a remote computer system. Communication lines are used to transmit the backup data to the remote location. If the e-vaulting product operates over the internet (or an intranet) then the TCP/IP protocol, made popular by the internet, is the method (protocol) that is used to send the data from the client system to the remote system.
Often the terms "network backup" and "network recovery" are used to describe LAN backup or e-vaulting solutions, therefore it may not be clear which category a specific product falls into without further analysis of the features offered.
Factors to consider
Developing a backup strategy is the only sure way to protect your data from loss or corruption. There are many factors which help to determine an effective strategy for regular backup of your important data files. If you are new to developing backup strategies, there are a number of factors which need to be considered carefully. In no particular order these are: data sensitivity, frequency of modification and cost.
Consider first of all the sensitivity of a particular piece of data or, more precisely, the critical nature of the data. The more sensitive data on your system will require more frequent backups. Think about including these files in your daily backup schedule. It is important that data which is critical to your business operation be backed up as often as possible, even more than once a day.
Consider next the frequency of modification of your data. All the modifications made to the computer data since the last backup have the potential to be lost forever. With many computer systems the amount of data change is typically quite small in comparison to the amount of overall disk capacity.
But remember, you must be prepared to re-apply any modifications on your data since the time of your last backup. Industry studies indicate that an average of four percent of your data changes between backups - typically daily.
The retention period is a backup parameter related to the frequency of change of your data and possibly the legal nature of the data. If the data does not change rapidly then less frequent backups could be stored for longer periods of time. The retention period may be specified in days or generations.
You use the generation retention method when you do not know how often the backup will take place, but you want to ensure that a certain number of copies of the data are available.
Certain data types must be kept for at least seven years for legal or regulatory reasons. This would require specifying a retention of seven years on a periodic backup such as a monthly backup. Usually you only have so much time available for backup. If your business operation restricts the availability of the system for backup then you need to get the data off the system as fast as possible. The fastest tape drives today are DLT format, but the cost of using this type of drive is high.
And finally, you must consider the cost of backing up your data. If you are performing local tape backups then the costs involve the time for performing the backups and the number of tapes required to implement your backup strategy. Don’t forget, the cost of some tapes can be significant !
If you are sending your data off-site via an e-vaulting service then the total amount of data stored remotely is the primary cost factor. Archiving data to tape, as opposed to keeping older data on-line when using an e-vaulting service will reduce costs.
By organizing your backups into different categories with different backup schedules you can help to reduce your costs considerably. For example, it may not be necessary to backup your applications if you are able to restore these components yourself from their original installation media. Customizations made to application initialisation files and licensing information might have to be restored manually but this will vary with each application.
Most sophisticated computer users create separate DATA directories on their systems where all non-application program files are stored, such as databases, spreadsheets, word processing documents, and graphic images. This type of organisation makes it easy to define a backup strategy for critical data.
Only perform backups of the operating system after you have upgraded or made changes to its configuration. You should save a couple of generations of your system data. Old operating system versions are generally not going to be of much value to you.
Organize the rest of your files based on their sensitivity and frequency of change, and construct a monthly, weekly and daily backup schedule for these files. Try to keep your retention periods to a minimum as saving multiple copies of your data will cost you more.
But always remember, there will be times when you need a file restored to a state other than the last backup. In these instances saving multiple generations of a backup can be very important. An example of this requirement would be to recover from a virus - you may want to go back a couple of days to restore a clean system state.