(A) Recovery measures
Checkpoints are used in log-based recovery method. Checkpoints are inserted into the log whenever an output to storage is performed. i.e. the database has been updated until this point to the permanent storage. The transactions that were committed before this checkpoint need not be redone in case of failure after this checkpoint. Using checkpoints, we do not need to read the complete log file and perform redundant redo operations in case of a failure.
https://en.wikipedia.org/wiki/Transaction_log
codex.cs.yale.edu/avi/db-book/db5/slide-dir/ch17.pdf