A Performance Evaluation Of Alternate Numbering Based XML Indexing Techniques

Abstract

Since XML became a standard of representing semi-structured data and exchanging format over the web, the sheer volume of XML data has become larger. While relational database represents data as a structured format, XML represents data in a self-describing way as a hierarchical tree structure. For expediting query processing over XML, many different types of indexing techniques have emerged. We will focus on numbering-based indexing techniques in this thesis. We will present performance comparison according to different XPath queries among three distinct numbering based XML indexing named GENE (Generic numbering based), XISS (Range based numbering), and XACC (Dimension based numbering) over shallow/deep tree structured data generated by ToXgene. By doing experiments, we realized that XACC showed relatively better query response in most of the cases. An analysis goes to three dimensions: varying size of the XML data, distinguished XPath queries having different features, and two different structures of XML data.