Object-oriented databases in a client-server environment

In contrast to existing database systems, the evolution of object-oriented databases has accompanied the growing trend towards the use of desktop workstations. They have consequently been designed around a client server architecture from the start. The kind of applications for which object-oriented databases have been associated also tend to suit the client-server approach, in that they typically exploit the powerful client workstations to handle much of the computational load. A key purported benefit of object-orientation is its ability to form classes which bind given attributes of an entity with a clearly defined set of access functions called methods, so called encapsulation. The resulting self-contained objects are well suited as a basis on which to build a client-server model. Individual objects form a compact and convenient unit which can be moved around a distributed system. In practice however this raises a number of issues which are discussed. The author covers some of these issues, paying particular attention to two: performance and access to non-object oriented legacy data