In database discussions lately I have heard the terms restore and recovery used interchangeably as if they mean the same thing. With regard to relational databases, these are different processes so my goal is to bring clarity to these terms.
The restore process involves copying data from corresponding backup files to a specific location.
In the recovery process, the database log file and data files are brought to a consistent state that involves rolling transactions forward or backward.
It is important to note that a database recovery process can occur independently of a restore process. For example in SQL Server instances there is a feature that allows a user to attach the log and data files to a SQL Server instance bringing a database online. Depending on the state of the database when it was taken offline (cleanly shutdown or I/O was being performed) the recovery process can be used before the database is online to bring the database log file and data files to a consistent state.
The good news is Pure Storage has product offerings that can speed up these processes. FlashBlade can be used as a database backup target with high throughput capability.
If your database uses FlashArray as its primary storage, low latency can speed up the time it takes in recovery.