Three layered architecture of DBMS is a framework that helps in structuring a database management system from three aspects- users, DBA and storage. The whole purpose of three layered structure is to implement abstraction in Database systems. Change in one schema must not affect the other schema. This architecture allows the DBA to design the schema and  implement secure access mechanism on stored data.

The three layers are also called schema which refer to representation of database at these three levels.

three layered architecture of DBMS

  • The external schema is used by the users to interact with database. It deals with the way the users visualize the database according to their need. This level allows multiple views of same data for different users without changing the underlying two schema.
  • The Conceptual schema is the domain of a Data Base Administrator. It is designed to represent the organizational data as relations and records. If a change is to made in the conceptual schema it should not affect the existing external schema. This level too allows multiple views of different conceptual records.
  • The Internal schema is how the data is actually stored in the storage. Any change in the internal scheme in terms of hardware or file structure should not affect the above two schema

Detailed Three layered Architecture of DBMS

detailed three layered architecture of DBMS

Mapping among these schemas allows creation of relationship among records in the three levels. Three layered Architecture of DBMS defines mapping between external and conceptual views and conceptual and internal views. This mapping implements the correspondence among the records and defines the relationship of the conceptual and external view.

The external view is the abstraction of conceptual view which in turn is the abstraction of internal view. It describes the contents of the database as perceived by the user or application program of that view. A mapping between conceptual and internal schema describes relationship of conceptual schema and the physical database.