That's a big question. Here's my attempt at a generic answer.
Level 1: Loss of disk or server
Protection: Mirroring, backups, snapshots, etc. which aren't stored on the server in question.
Level 2: Loss of data center or building
Protection: Same as #1, just do to a greater degree. You may consider site-to-site replication to be ready to recover easily and quickly.
Level 3: Loss of multiple data centers, a campus or a city
Protection: Again, this is all a matter of degrees. If you want to protect against a level 3 disaster, your "backup" has to be really far away. If you don't want to lose any data in such a disaster, it needs to be asynchronous replication. If you don't mind downtime in the event of a level 3 disaster, then shipping tapes very far off-site can also be a good solution.
All of these answers are basically the same. Basics are a good last line of defense. Replication is a much better backup as the "backup" exists in restored form all the time as opposed to traditional backups that must be restored from tape before they are useful.
There are basically two questions:
1. Are you going to use backup, replication or both?
2. How far away are you going to put your backup/replication/mirror?
After you've provided for data recovery, you also have to look at how you're going to recover your infrastructure, people, processes, etc. It's no good if you've got computers with data on them that can only be accessed by people who can't get to them. It's no good if they can use the computers but can't make or receive phone calls.
Editor's note: Do you agree with this expert's response? If you have more to share, post it in one of our .bphAaR2qhqA^0@/searchstorage>discussion forums.
This was first published in February 2003