The Sakai code comes preconfigured to use Hypersonic SQL (HSQL), an in-memory relational database engine. It can also be configured, at deployment time, to use either Oracle or MySQL. The object/relational mapping (ORM) software Hibernate is made available by the Sakai framework to allow TPP tools to effectively define their own database tables and automatically forward generate them. Sakai still, at this point, uses SQL statements directly embedded in the Java source code, in many cases this SQL uses Oracle or MySQL specific syntax. In fact, Sakai releases seem to be tilted more toward Oracle than any other RDBMS, this is probably due to the fact that the University of Michigan uses Oracle for most of its database needs. This is an area of concern if you wish to use another RDBMS like PostgreSQL. Ideally Sakai should be using a Hibernate (or perhaps iBatis' sqlMaps) layer for ALL of its database requirements.