The Service Level Agreement (SLA) is driven by the needs of the customer. While assessing the existing environment (or building a new environment), the solution provider will examine the size of the backups, type of backups (file systems, database, etc), and how the backups need to be performed (online, or backup window). Once the environment is understood, an SLA can be written.
The SLA will address a maximum recovery time based on the type of data and the size of the file. A sample SLA might be to backup a 25GB file system and will be made available for recovery within two hours.
Recovery methodologies vary. Some companies that might have people onsite with root access can restore and mount a file system in the customers environment. This may work fine for a simple file system but when a database is involved, a different approach might work better.
Database recovery should have a DBA involved. The DBA knows his or her environment better than anyone. The solutions provider in this situation will locate the backup (tape or disk) and make this data available for the DBA to restore in the most effective manner. Typically, this is a database restore and applying logs for a full forward recovery up to the point of the error (in the case of DB corruption).
With regard to restoring in a "GB per minute" fashion, this is definitely possible. Again, by analyzing the environment and understanding the backup needs, an SLA can be created around these parameters. In the example SLA that I stated before, this would equal 0.208 GB a minute.
This was first published in July 2001