ACID compliant data is required to have these abilities:

• To be accessed at the lowest level of the dataset (Atomicity)
• For the same data to be returned from all nodes at all times (Consistency)
• To isolate concurrent transactions from each other (Isolation)
• Once the data is stored, it is stored permanently and can be recovered even in the case of system failures (Durability)

Relational databases are usually designed to be ACID compliant and enable the same data to be stored in the database by only one writer at a time. Relational databases introduce a locking mechanism to prevent data from being written at the same time by multiple writers (Isolation), and that feature means that we can keep a lock on newly written data until it is replicated across the cluster (Consistency). This locking also enables us to only change atomic parts of each piece of data by locking a table, index, or part of an index and writing changes to it (Atomicity). The data in relational databases is stored durably and there are special provisions usually put in place that enable us to make sure the data that is stored will be recoverable and restorable in case of failures (transaction logs, snapshots, backups, and so on).