XML Data Management in Object Relational Database Systems

1.1 User Perspective

There is a general trend for users to increasingly prefer to manage all of their data in one platform for the benefits of easier data management, integrated query services, and sharing a common infrastructure for all types of data. This is more so with relational databases that are widely adopted as mature platforms for managing structured data. With data extensibility mechanisms, such as user defined types, user defined functions and extensible indexes that are added into relational databases, the RDBMS has evolved into an ORDBMS with object-relational support (Stonebraker, Brown & Moore, 1999) that is capable of managing more complex data that may not fit into the relational model. Using an ORDBMS, users are able to manage various kinds of non-relational data such as object hierarchies in one platform. XML data management is a continuation of this trend with users looking for a single data management platform to avoid the overheads of partitioning and integrating data across multiple repositories. Major database vendors such as Oracle, IBM and Microsoft, in recognition of this trend have released XML capabilities in their products to support industrial strength applications that span all forms of data including relational data and XML data.

The other aspect is that XML has been used as a data exchange language to exchange data. Organizations have used XML as a portable way of exchanging data, in particular, exchanging data from and to relational systems with well-defined XML schema. This is associated with the need to generate XML from relational data at the data sending side and the need to extract relational data from XML at the data receiving side. To support this, it is critical for an ORDBMS to provide support for bridging the XML and relational data models.