1. A relational DBMS is a “storage” or “persistance” layer. NO! The relational model is an abstract mathematical model for representing and manipulating data. Whether the data is in memory or on disk is completely irrelevant from the model perspective. In an implementation of the model (an RDBMS) the designer, programmer or user should not need to know how or where the data is represented.
2. SQL is relational. NO! SQL is influenced by the relational model, but contains many troublesome violations of the basic principles. This makes working with it much harder than it should be.
3. NoSQL DBMSs are innovative. NO! NoSQL DBMSs are largely a throw back to methods that the relational model (even in its pale SQL form) has completely superceded: graph, hierarchical and multi-value.
4. Relational DBMSs are not scalable. NO! As one of the central principles of the relational model is the separation of the logical and physical layers, you can use whatever tricks you want on the physical layer to improve scalability and performance without having to make any changes from a designer, programmer or user perspective.
5. Relational DBMSs are inflexible. NO! Changes to the database structure for the most part have no effect on existing applications. Existing applications continue to function when new tables or columns are added to the database. The main exceptions are when the applications are poorly designed and rely on column order rather than column names.