They are controls for avoiding inconsistencies in the data when working in a multi-user environment.
To control the concurrency, it is necessary to lock the information.
The programs’ concurring execution must not cause bad performance.