Efficient XML Data Management: An Analysis

With XML rapidly gaining popularity as the standard for data exchange on the World Wide Web, a variety of XML management systems (XMLMS) are becoming available. The choice of an XMLMS is made difficult by the significant difference in the expressive power of the queries and the performance shown by these XMLMS. Most XMLMS are legacy systems (mostly relational) extended to load, query, and publish data in XML format. A few are native XMLMS and capture all the characteristics of XML data representation. This paper looks at expressive power and efficiency of various XMLMS. The performance analysis relies on the testbed provided by XOO7, a benchmark derived from OO7 to capture both data and document characteristics of XML. We present efficiency results for two native XMLMS, an XML-enabled semi-structured data management system and an XML-enabled RDBMS, which emphasize the need for a delicate balance between the data-centric and document-centric aspects of XML query processing.