Recovery Point Objective
To define a recovery point objective, determine how much data loss your system can sustain in the event of a disaster. This value can be used to determine how frequently checkpoints must be taken, and how frequently a mirror grid must request updates.
The frequency of checkpoints determines how often updates are available, and the frequency of mirror requests determines how quickly a mirror grid begins fetching new checkpoints once they are available. The available bandwidth between sites and the expected update rate must also be taken into consideration to determine how long it takes to fully mirror a checkpoint.
Consider the following example: User A can tolerate 1 hour of data loss, expects to write 400 MB/hour, and can transfer the data in about 6 minutes over a 10Mbps WAN link. If the user takes periodic checkpoint every hour, the mirror data is at least 66 minutes older than the primary, their recovery point objective is never satisfied. Therefore, they must take checkpoints more frequently to guarantee that they can meet their objective.