Data abstraction is the idea that a database design begins with a high level view and as it approaches implementation level, the level of detail increases. The benefit to using levels of abstraction is the ability to work with and integrate multiple views into a cohesive set.
In 1970, the American National Standards Institute (ANSI) Standards Planning and Requirements Committee (SPARC) established a framework for database design based on the degrees of abstraction. The ANSI/SPARC architecture is composed of four levels of data abstraction; these levels are external, conceptual, internal, and physical.
The External Model Edit
The external environment is the end users' view of the data. The end users view of data usually applies to their specific business needs and those of their organizational unit.
The benefits of representing the design through the external model are:
- It is easier to identify the data needed by the end users.
- It can be easily checked to ensure it is adequate and will support the processes, requirements, and constraints as defined., it easy to use by OOM
- Through good design, security can be increased by only allowing users access to the subset of data that they need.
The Conceptual Model Edit
The conceptual model is created by taking all views and forming a global view of the entire database. The conceptual model is generally represented by E/R Diagrams. Another name for the conceptual model is the logical design of the database.
The Internal Model Edit
The internal model can be created after a specific DBMS platform is chosen. The internal model is the database as seen by the specific DBMS. What sets the internal model apart from the external and conceptual is its reliance on its software platform. The goal in designing the internal model is to acheive logical independence, where the internal model can be changed without affecting conceptual model.
The Physical Model Edit
This is the final and lowest level of abstraction. This is the model which describes such implementation level design as how the data is stored on media and what media to use. This level of abstraction is reliant on software and hardware. This level is created last because designers will know the exact specifications of the database and can from there decide what exact hardware specs are needed. The goal of the physical level is to create a design where the physical model can be changed without affecting the internal model.
If the rules established by the ANSI/SPARC are followed, the database is easily scalable and upgradeable. A common need is for the ease of upgradability in the physical model. As technology improves and as the database grows and needs more processing power and space it is important to be able to upgrade the hardware without worrying about needing to redesign parts or the entire database.