What is the advantage of concurrency control?

By having concurrency control you avoid dirty writes (or inconsistent data). It is like your traffic signal - allows one set of users (drivers) to pass through at a time. In computer terms, if a user wants to update some data then the following steps are taken: (a) the data is fetched along with the timestamp (the date, time when it was last modified) (b) the data is modified by the user who then submits the data for update / storage in the system (c) the system verifies the original timestamp with the current timestamp. (i) If both are same, the new data (including the new timestamp) is updated/stored in the system. (ii) If they are different, then obviously the data had changed between the time it was fetched and now - meaning that someone else had updated in the meanwhile. The system typically informs the user that the data had been modified by a different user and asks him/her to submit the modifications once again. The system would implicitly do step (a) above allowing the user to start from step (b) onwards.