What does “backup” mean to you?

I was at LISA’07 in Dallas in mid-November, and attended Jacob Farmer’s training session, titled “Disk-to-Disk Backup and Eliminating Backup System Bottlenecks.”

Early in the talk, he brought up a question that’s fundamental to this business: what does “backup” really mean? His breakdown was particularly clear, in six categories:

  1. Object recovery — recovery of specific files, rows, tables, etc.
  2. Application checkpoint, e.g., a snapshot of an application prior to an upgrade.
  3. Server hardware failure protection, which includes bare-metal restores.
  4. Protection from storage device failure, which generally has tighter recovery-time requirements than server hardware failure.
  5. Site failure: floods, fires, etc.
  6. Historical archive, usually for regulatory compliance.

I thought this was a particularly insightful way to break down this problem domain. These categories form a nice basis for differentiating backup products, and for seeing where backup products will complement one another nicely. An application that does a great job of object recovery may operate entirely independently of an application that manages historical archives. There’s no reason that a single application should serve both of those needs.

Amanda is an interesting application in that it can be used to achieve any of these aims. As we build in more flexibility and modularity, we will try to balance the attention to these categories to keep Amanda well-rounded.

Comments are closed.