The ANSI/SPARC architecture, which dates from 1975, defines abstraction levels for a database management system:
- Internal (or physical) level: defines how data is stored, and how it is accessed.
- Conceptual level: also called the DRM (data relationship model) or LDM (logical data model). It defines how information is arranged within the database.
- External level: defines user views.
Characteristics of a DBMS
The three-level architecture defined by the ANSI/SPARC model keeps data and processing separate. Generally speaking, a DBMS must have the following characteristics:
- Physical independence: The physical level may be modified independently of the conceptual level. This means that the user cannot see all the hardware components of the database, which is simply a transparent structure for representing the stored information.
- Logical independence: The conceptual level must be editable without disrupting the physical level. In other words, the database's administrator must be able to make improvements without affecting the users' experience.
- Ease of use: People who are not familiar with the database must be able to describe their query without referencing the database's technical components.
- Speedy access: The system must be able to reply to queries as quickly as possible. This requires fast search algorithms.
- Centralised administration: The DBMS must enable the administrator to manipulate data, add elements, and verify their integrity in a centralised manner.
- Limiting redundancy: The DBMS must be able to avoid redundant information whenever possible, both to minimise errors and to prevent wasting memory.
- Integrity verification: The data must be internally coherent, and when some elements reference other elements, the latter must be present.
- Data sharing: The DBMS must allow for multiple users to simultaneously access to the database.
- Data security: The DBMS must be capable of managing each user's data access rights.
Niveles de datos
Les niveaux de données
I livelli di dati
Os níveis de dados