What's a "Generational Backup Scheme" And How Does It Work?
A generational backup is one of the simplest, most effective methods for making and keeping backup copies of your data. Properly done, it combines ease of use and data protection.
The most common generational backup scheme is the three-generation or "grandfather-father-son" method. In its most basic form it involves making a complete copy of the data to be backed up on removable media such as tape or CD. This is the grandfather. At the next scheduled backup period, say the next day, another complete copy of the data is made, which of course includes the changes in the data during that period. This is the father. At the next scheduled backup, the third copy, or son is produced.
The fourth backup is made by recording over (or replacing, depending on the media) the grandfather copy. The new copy becomes the son, the previous son becomes the new father and the father is promoted to grandfather. This continues in rotation so there are always three backups, each of a different point in time.
The advantage of saving the two previous backups as well as the current backup is that if the computer data somehow become corrupted and the problem isn't caught until after the backup is made, there are still two uncorrupted, albeit increasingly out of date, backup copies. Given reasonable care, it's unlikely a problem will corrupt all three backups before it is caught. Similarly, if one of the backup copies is damaged, you still have two more. The three-generation backup also makes it easy to store one of the copies (usually the grandfather) in a more secure and often off-site location.
Note that this approach does not take incremental backups into consideration each backup is a full backup.
Another approach to this is to create a full backup to serve as the grandfather. Let's say this is on a Sunday. The next backup date might be the following Sunday, when the father backup is created, with the son being created a week later. You still can create incremental backups each day between full backups. That way, instead of losing a week's worth of data, at worst you lose a day's worth, or however many days occurred between the corruption of the data and discovering the problem. This further minimizes the potential loss and, since you're using incremental backups, makes those daily backups much faster than full backups.