Realizing High Performance Orthogonal Persistence

A rapidly growing demand for networked information has stressed the need for effective ways of managing large amounts of complex data. While more traditional solutions such database technology are already playing a major role, it seems clear that the mismatch between the relational and object oriented paradigms will be exposed by some applications as a source of major inefficiency. The upside project at the ANU has been investigating the application of orthogonal persistence to large information servers—what we refer to as ‘high performance orthogonal persistence’. Central to our work has been exposure to a major application that seems well suited to such technology, the Australian Bureau of Statistics’ Business Register project. This paper discusses our approach to high performance orthogonal persistence, outlines some of our achievements to this end, and looks at some of the key problems that remain.