Techniques for Storing XML SIhem Amer-Yahia and Mary Fernandez   View Paper (PDF)   Return to Advanced Technology Seminars Abstract XML is the de facto standard for data exchange between applications on the Web. Applications, such as electronic markets, will produce and consume large volumes of data and therefore will require efficient and reliable storage and retrieval of XML data. Many techniques for XML storage have been proposed, including flat files, relational database management systems, object-oriented database systems, LDAP directories, and native XML database systems. To better understand the requirements of XML storage systems, we first review various classes of XML documents including highly structured data as stored in relational databases, "mixed" content from document-processing applications, and "streams-oriented" data from e-commerce and transactional applications. We also consider the types of queries typically applied to these classes of documents. In the second part, we present features of the XQuery and XPath data model that must be supported by an XML storage system and then we describe in detail a variety of storage alternatives from industry and research. We focus on techniques that use relational storage. Typically, these techniques produce a logical relational schema for the XML data and treat the storage system as an "black box". In the last part of the tutorial, we consider new techniques that open the storage system's "black box" so that we can take advantage of physical-layout features. DiSC'03 © 2003 Association for Computing Machinery