Distributed database
A distributed database management system is a database management system (DBMS ) which has a certain degree of autonomy, heterogeneity, and distribution. Depending on the nature of these three criteria in a distributed DBMS possibilities can be distinguished.
Examples of distributed DBMS are master-slave systems, client-server systems ( low spread ) and peer-to -peer systems (highly dispersed ).
Autonomy
The autonomy of a (distributed) database management system refers to the degree, can operate independently to the various interconnected DBMS. This independent decisions can be made in terms of design, communication and execution. A high degree of autonomy through various decisions usually also to a growing heterogeneity.
One distinguishes the following autonomies:
- Communication autonomy of a distributed database (VDB ) refers to the ability to decide on the communication with other parts of the database.
- Execution autonomy refers to the ability of part of the VDB, without the action to perform local operations by external operations of other parts and to determine the execution order itself.
- The association autonomy of part of the VDB implies that it may decide whether and to what extent it shares its functionality and resources with other parts.
With a high degree of autonomy is also called a federated DBMS or a " Federated database " as opposed to an integrated database. The term " federated " "distributed" instead emphasizes that systems are developed independently from each, while a highly distributed system may have been designed also from autonomous components.
Heterogeneity
With a high degree of homogeneity is known as a homogeneous DBMS, i.e., the systems in use are substantially similar or the same. A trivial example is an asymmetric distributed system (client-server system ) consisting of personal computers, on each of which the same operating system is operated and operate the exception the same DB application.
In contrast, there are also heterogeneous DBMS. This use according to different systems and applications. An example of this are the hospital information systems, as a function of the respective departments (laboratory, patient management, accounting, service planning) results in very diverse data (image, sound, text, etc.). The programs used - especially in university hospitals - are usually specific developments that do not apply via the defined scope of work ( each department uses its own software). For this reason, computer scientists are charged with the planning, development and maintenance of such systems in large hospitals.
Implementations
In the 1980s there were already functioning prototype of distributed databases, but the problem was present at that time, that network protocols were rarely standardized so that the implementations were run only under very specific environments.
Some developments, some of which contain at least functions of distributed database systems are: