For JDBC/SQL programmers...
ozone is a truly object-oriented system. There is no need to describe database schemes or database queries in a different language than the underlying programming language. You can program persistent objects like "normal" objects. No type mismatch, no impedance mismatch - just pure Java with persistent objects in a transactional environment.

For legacy system programmers...
You have probably tested a lot of tools that can store Java objects. The most of these tools use relational databases as back-end and so they have to map objects to tables. Such tools are not really object-oriented and not really relational. They inherit the disadvantages of both worlds.

For users of other Java ODBMSs...
ozone is based on a central activation architecture. Objects never leave the server for any reason. The client applications control their objects in the database via proxy objects.

This architecture has advantages but also disadvantages over systems that activate objects in the client. ozone is better for applications where the data are frequently changed by different clients.