Current XML database solutions largely neglect or make insufficient use of available schema definitions. Although XML documents are to some extent self-describing by means of their markup, availability and use of rich structure and type information contained in schema definitions written in languages such as XML Schema is nevertheless essential for their effective management in a database: only with this information at hand, it is possible to ensure database consistency by document validation, to store, index, and access the content of elements and attribute values in an appropriately typed manner and not just as text, and to perform more sophisticated query optimizations. In this paper, we give a system overview of the Persistent Typed DOM (PTDOM), a schema-aware native XML database solution originally developed for the management of MPEG-7 media descriptions. The core of PTDOM is made up of a schema catalog capable of managing schema definitions written in MPEG-7 DDL, a superset of XML Schema. PTDOM exploits the schema information maintained by this catalog in a variety of contexts: for document validation, for typed storage of elements and attribute values, for structural indexing of XML documents providing additional efficient access paths to document contents, and for optimized construction of query execution plans for XPath expressions. This along with its profound extensibility with new datatypes, user-defined functions, and value index structures makes PTDOM a flexible and effective database solution not only for the management of MPEG-7 media descriptions but also for the management of any kind of XML documents for which schema definitions exist.

Top