Adding XML to Databases; XML Europe '99

One of the split plenary sessions on Wednesday focused on two leading database providers and how they intend to add XML capabilities to their databases. Steve Munsch, the XML Evangelist for Oracle Corporation and Paul Nerger, Executive Director of Marketing for Informix provided us with radically different approaches to the integration of XML and databases.

Steve Munsch began by discussing not only the explosion of the Internet but of wireless devices as well. According to Munsch the demand for access to enterprise information is growing rapidly. Today we want to process, store, query and exchange all sorts of enterprise information. According to Munsch, it is clear that XML provides the mechanism to do so. XML is strategic to Oracle because it can be used to represent all data. Oracle is interested in extending the datatypes that can be handled by their products and XML can help them do so. In addition to XML, itself, Oracle sees XML Schemas, XSL style sheets, and an XML query language as critical.

Munsch discussed the Oracle XML parser. It is written in Java and runs in the database VM Kernal. It is integrated with an XSLT engine and transforms into and out of XML on the server side, using either the tree-based DOM or the event-based SAX interface. Dynamic generation of valid XML documents from an Oracle database is Oracle's end goal. Oracle will deliver XML or can provide a transform to HTML for viewing. In the future, Oracle plans to incorporate a native XML datatype into its database and to be able to issue XML-based queries.

Paul Nerger from Informix provided a different perspective on how a database should support XML. Informix is based on an Object/Relational model with a universal data option that turns a relational database into an object database. Using this technology, Informix allows the end user to create new data types to meet business requirements. To Informix, XML is simply one of many interesting new datatypes. Nerger believes this is important because although we know XML is a standard, standards can change. The user-defined datatype mechanism allows for infinite extensibility to follow XML as this standard evolves over time.

The Informix technology that enables users to define new data types and associate their own business rules to these datatypes is called DataBlades. Informix does include an XML parser. However, Nerger stressed that Informix is designed so the user can integrate their own XML parser. Right now the Informix focus is to develop basic DataBlades and provide a mechanism for the user to extend these to define their own. DataBlades currently defined include text, geospatial, rich media, and time-series.

Return to TOC