Layered DBMS Architecture

More than two decades ago, DB researchers faced up to the question of how to design a data-independent database management system (DBMS), that is, a DBMS which offers an appropriate application programming interface (API) to the user and whose architecture is open for permanent evolution. For this purpose, an architectural model based on successive data abstraction steps of record-oriented data was proposed as kind of a standard and later refined to a five-layer hierarchical DBMS model. We review the basic concepts and implementation techniques of this model and survey the major improvements achieved in the system layers to date. Furthermore, we consider the interplay of the layered model with the transactional ACID properties and again outline the progress obtained. In the course of the last 20 years, this DBMS architecture was challenged by a variety of new requirements and changes as far as processing environments, data types, functional extensions, heterogeneity, autonomy, scalability, etc. are concerned. We identify the cases which can be adjusted by our standard system model without the need of major extensions or other types of system models.1