Distributed database systems represent collections of nodes connected to the network, each node being a database that communicates and collaborates with the other databases so that the user can access data from anywhere within the network, in a transparent manner. Distributed database systems are mainly composed of distributed databases and the distributed database management system.
Distributed database systems must help solve the problem of “information islands“, which have arisen as a result of geographical separation, incompatible computer architectures, incompatible communication protocols, and so on. Integrating database systems into a whole from a logical point of view could stop this way of thinking. It is not mandatory for each node to have its own database.
The distributed database system comprises the following software components: the data component, the hardware and communication component, the local database management system, the global data dictionary, and the distributed database management system. The data component comprises the distributed database. The hardware and communication component comprises the network of computers and the communications infrastructure that makes the connections within the network. It includes the complete description of the nodes and connections within the network.
The local database management system, together with the local database is a standard database management system. He has his own dictionary for local data. The global data dictionary contains information about the distributed database: location, structure, availability, and how to use the data.
Distributed database management systems are the software through which several local databases are linked through a communication system so that users from other stations can access data from any station. In other words, it comprises a complex system of programs that provides the interface between the distributed database and its users. A distributed database is a database controlled by a central database management system and in which the data storage media are located at the nodes of a computer network. The database is fragmented into partitions at different locations, each partition/fragment being able to replicate or not.
Often there is confusion between a distributed database and a distributed processing database. Distributed processing means accessing a centralized database through a computer network. A distributed database is a collection of logically inter correlated and physically distributed data on a computer network.
The objectives of distributed databases, like traditional databases, refer to centralized control, data independence, minimum and controlled data redundancy, efficient access to data, data integrity and security.
In the case of distributed databases, centralized control is more pronounced than in the case of centralized databases. To these, a hierarchical control structure is identified based on a global administrator who has central responsibility for the entire distributed database and on local administrators who can have a high degree of autonomy that can go until the coordination between stations is achieved.
Distributed databases differ significantly in terms of station autonomy, which can vary from complete autonomy, without the existence of the global administrator, to a complete centralized control.