Processing XML in LDAP and its Application to Caching

Pedro José Marron

ISBN 978-3-89722-798-9
195 pages, year of publication: 2001
price: 40.50 €

In the past years, the proliferation of XML-based systems in the Internet has dramatically increased to the point of requiring the redesign of existing software and even the creation of new paradigms to properly handle semistructured data. This transition has lead to an increase in the involvement of the database community to reevaluate the classic models and even to create new models and mechanisms to efficiently represent and process semistructured data. Following the same trend, also the Internet community, and more importantly, its standardization bodies, like the World Wide Web Consortium (W3C), have started developing XPath, and a series of related query languages, especially tailored for XML, to allow the querying, processing and translation of XML-based documents in an efficient way.

At the same time, with the release of the Lightweight Directory Access Protocol (LDAP) v.\;3, the popularity of directories to store information about users, networks, etc., has been gaining momentum, reaching a stadium where most universities and research institutes, like AT\&T, and commercial companies, like Netscape or Microsoft, offer LDAP support in their browsers and operating systems.

The similarities between LDAP and the Document Object Model (DOM) developed by the W3C to be used as the canonical representation model for XML, have lead to the design in our thesis of a set of models and algorithms for the generic representation, processing and querying of XML documents on top of an LDAP-based system, showing that it is not only possible to combine the best of both worlds into a single system, but that this system is far more efficient than their DOM-based counterparts.

Additionally, we have implemented our algorithms in \textsc{HLCaches}, a hierarchical, distributed, LDAP-based proxy caching system that enables the storage and retrieval of XML documents, as well as traditionally cacheable documents, like HTML and graphical data, in a common environment. Furthermore, XML-documents can be queried and processed, providing us with a system that is able to very efficiently implement the functions of a proxy cache, thus speeding up the access to common data from the perspective of most users, and to implement the full expressive power of the XPath query language to process XML documents.

The purpose of this thesis is, therefore, to present the set of generic mapping algorithms between the XML and the LDAP world, as well as to fully characterize their efficiency, not only at a theoretical level, but also when implemented as part of a proxy caching system with full XPath query capabilities.