Distributed database architectures

The characteristics of distributed database architectures are discussed. It is noted that the advent of powerful modeling facilities, such as object-oriented and knowledge-based data models, has paved the way for a complete reconsideration of several of the long standing assumptions and perspectives that are pervading the field of distributed database architectures. It is now possible to propose novel architectures based on high-level data modeling facilities. The advocated architecture is called the semidecentralized or clustered architecture and combines the positive aspects of both logically centralized and federated databases. This architecture substantiates the substrate which automatically classifies the activities that each preexisting data base management system may undertake and is in the position to describe the entire volume of information that can be supplied by each individual database system in the network.<>