Erik Behrends:
Evaluation of Queries on Linked Distributed XML Data.

Complete Text [
.pdf, 1.42MB]
PhD Thesis, Faculty of Mathematics, Goettingen University, December 2006
In: Evaluation of Queries on Linked Distributed XML Data, December 2006
© Goettingen State and University Library

XML (eXtensible Markup Language) is the de-facto standard for exchanging information and for representing data in the World Wide Web. In contrast to the document-centric perspective given by the well-known language HTML which defines the human-readable content and the layout of web pages, XML offers more flexibility and expressiveness. XML documents are not required to be self-contained but may rather have links to other XML resources. For expressing such links between XML documents, the W3C (World Wide Web Consortium) proposed XLink - but mainly for browsing purposes. If the linked documents are considered from the data-centric viewpoint, it shows that XLink does not specify how the referenced instances should be handled. Especially, it is not possible to query along links though the W3C XML Query (XQuery) Requirements explicitly state that this has to be guaranteed. In order to cope with these issues, an XLink extension "dbxlink" has been proposed. It allows for modeling interlinked XML instances as integrated views where XLinks are resolved in a transparent way. In particular, it is possible to query these instances with XPath and XQuery. In this work, the dbxlink model is described and it is investigated how to query distributed XML instances interlinked with a simple kind of XLinks according to this approach. Different strategies are analyzed and emerging problems like the handling of cyclic instances are treated. It is shown how to extend XPath-based query systems in order to be able to handle queries wrt. dbxlink. Furthermore, optimizing techniques like special caching strategies are proposed. The results of these investigations have been used to conduct a proof-of-concept implementation of the dbxlink approach as an extension to the open source XML database system eXist.



	author = {Erik Behrends},
	title = {Evaluation of Queries on Linked Distributed XML Data},
	journal = {},
	year = {2006},
	note = {PhD Thesis, Faculty of Mathematics, Goettingen University, December 2006},
	url = {}