Object-oriented databases (OODBs) are an important technology for advanced data-intensive applications, as they have emerged as the principal alternative to relational databases in the commercial world. Such systems are used, or being considered for use, in a wide range of applications, including network management, spatial data handling and scientific data analysis. The Object Data Management Group (ODMG) standard for OODBs is beginning to be supported by products, and promises to provide improved product compatibility in terms of the data model, query language and programming facilities supported. However, OODBs are generally associated with advanced applications, in which performance requirements are strict, and experience shows that the performance of such systems often deteriorates rapidly when active data set sizes exceed the size of main memory.

Parallel database servers are now widely accepted as an effective mean of supporting high performance query processing on relational databases, but little work has been done on the design and development of parallel systems to support OODBs, or the ODMG model in particular.

The talk will describe the design of a parallel ODMG compliant object-oriented database system which we are developing. The key issues and design choices will be highlighted. We will also compare the design with that parallel relational database servers. This will highlight the major differences in the design of parallel servers to support the two main database models.