Tuesday, April 21, 2009

Segregation of Data

It may not be politically correct to use the word "segregation" as it applies to humans, but I am using it to describe how we can separate our data into pieces that change frequently and pieces that hardly ever change.

Whether you use Windows or Mac, the operating system takes up a huge amount of space on your disk, and the vast majority of this operating system data very rarely changes at all. It may change when you apply patches to the operating system, or when you install new software packages like Photoshop, Illustrator, or Lightroom, but otherwise it is pretty stable.

User data (the stuff you put on the computer, like photos, letters, music, etc) changes a lot, however. Every time you use your computer you are potentially changing data that is kept as part of your user profile, and any files you save or delete also cause changes.

So, the issue becomes one of separating (segregating) the user data from the operating system data. Both Windows and Mac make this fairly easy to do, if you are familiar enough with the system layout. On Windows, each user has his/her own area under "Documents and Settings" where the desktop, My Pictures, My Movies, My Documents, etc, etc are all stored. So, if you are careful to keep all of your files either on the desktop or in one of those folders, it will all be easy to locate and back up, simply by backing up the Documents and Settings folder. Incidentally, this will back up everything for all users on the computer, assuming you have different logins and multiple users.

On Macintosh it's almost the same, except each user has a folder under "Users" which stores their desktop and all their other files. Since MacOS X is Unix-based, they call the user directory the "home" directory. Each user will have a separate folder under "Users", and backing up just "Users" will back up all the user home directories.

This takes care of the stuff that comes built-in, but what happens if you install more disks and store more data? Well, that's actually easier, since you know that there is no operating system data on any of the additional disks (under normal circumstances), and you should know what type of data you will be storing on these additional disks, and how often it needs to be backed up.

One thing I should mention here is the concept of redundant storage (mirrors and raid). RAID (Redundant Array of Inexpensive Disks) storage was not designed for backup purposes. It was designed to provide higher reliability, and to be able to withstand the failure of one or more disk drives without data loss. In a mirrored configuration, the system puts copies of your data on two disks, so if one fails there is still a copy that can be retrieved until the bad disk is replaced and the mirror rebuilt. This saves you from a disk crash, but does not prevent you from accidentally deleting your data. This is where a backup will save you. If you never make mistakes, RAID is a good thing, but it's not a complete substitute for backups.

No comments:

Post a Comment