The isolation represents the way of separating the database from the effects of network accesses, thereby maintaining the consistency.
The different levels of isolation are:
? read committed: This level of isolation uses the shared locks and the reads to the database give the constant and consistent values.
? read uncommitted: No locks implemented. This is the least effective isolation level.
? repeatable read: There are lock over the rows and values but the updates are maintained as a separate phantom row which is the next set of values for the specific record. Values can change within a specific transaction of a SQL function.
? SERIALIZABLE reads: This is the implementation of pure lock mechanism where one specific transaction is not allowed access to specific record before another one completes.