XStorM: A Scalable Storage Mapping Scheme for XML Data

With the increasing ubiquity of XML, an eXtensible Markup Language, the industry is racing to provide XML infrastructure for e–commerce, information interchange, effective query of diverse sources and yet more integration of diverse data. It is anticipated that large volumes of XML data will be created manually from HTML documents or generated using some WWW tools and electronic data interchange (EDI). In this paper, we examine how large amounts of XML data can be stored in a relational database. Our scheme considers the unique irregular features of XML, including missing elements or multiple occurrences of the same element, and elements which may have atomic values in some data items and structured values in others. A detailed experimental study demonstrates good query performance, effective space utilization and scalability.