What is 'Optimistic Locking' in a data model?

1. A Explicitly locking every row in the target table before a read operation begins to ensure that no concurrent writes can occur during the query lifecycle.
2. B Utilizing a `version` column to verify that a record has not been modified by another process in between the initial read and the final commit phase.
3. C Assuming that the underlying storage hardware will automatically resolve all concurrent write conflicts at the physical disk-controller level.
4. D Implementing a centralized, global mutex lock within a secondary Redis cluster to strictly serialize all incoming data update requests.