Multidatabase management in Pegasus

Abstract

Pegasus is an object-oriented multidatabase system being developed at Hewlett-Packard Laboratories. The goal of the system is to provide facilities for multidatabase applications for accessing and manipulating multiple autonomous heterogeneous object-oriented, relational and other databases. This paper gives an overview of the architecture, describes the common object model and gives the justification for adapting the model for a multidatabase system. It describes the database environment of the system, discusses how to manage schemas consisting of types and functions from multiple databases. It briefly describes the features of the data language and gives an example. It discusses a method for dealing with object identification issues in a multidatabase system, reviews schema integration issues and briefly mentions schema mapping to relational models.<>