Having a proper backup strategy is one of the most important aspects of SAN management. In the SAN environment, backups have two goals. The first goal is to archive online data to offline media. This process is repeated periodically for all online data on a time schedule. The second goal is to provide access to offline data for recovery from a problem. For example, database recovery often requires a retrieval of archived log files that are not currently online.

Scheduling a backup depends on several factors:

  • Identification of critical applications that require more frequent backup cycles within a given period.
  • Recovery point and recovery time goals. Consider how precise your recovery point must be, and how long you are willing to wait for it.
  • The rate of change (RoC) associated with the data. For example, if you are using synchronous/asynchronous replication, the RoC affects the amount of bandwidth required between the primary and secondary storage devices.
  • Overall impact on a SAN environment, storage performance, and other applications.
  • Identification of peak traffic periods on the SAN. Backups scheduled during those peak periods can slow the applications and the backup process.
  • Time to schedule all backups within the data center.
  • Time it takes to back up an individual application.
  • Resource availability for archiving data, such as offline media access.

Include a recovery-time objective for each application when you design your backup strategy. That is, consider the time and resources necessary to perform a backup. For example, if a scheduled backup stores so much data that recovery requires a considerable amount of time, examine the scheduled backup. Perform the backup more frequently, so that less data is backed up at a time and the recovery time decreases.

If an application requires recovery within a certain time frame, the backup process must provide a time schedule and specific data processing to meet the requirement. Fast recovery can require the use of recovery volumes that reside on online storage. This process helps to minimize or eliminate the need to access slow offline media for missing data components.