PhD Thesis Proposal Defence


"Efficient Processing Techniques for XML Queries"

By

Mr. Kaiyang Liu


Abstract

XML has become the de facto standard for data representation and exchange
format on the Internet. With the rapidly increasing number of XML data,
efficient XML data management systems are required to handle routine
tasks, such as storage and query. Query processing is an important
component in a XML data management system, in the sense that it greatly
influences the performance of the retrieval of the stored XML data. Two
key characteristics differentiate the query processing techniques for XML
data from those for traditional databases. Firstly, data in traditional
databases strictly conform to a pre-defined schema, whereas XML allows for
irregularity, as it does not enforce a rigid schema on its underlying
data. Secondly, the core part of XML querying is navigation, in the form
of regular path expression, which is different from the traditional SPJ
queries. Our research focuses on efficient query processing techniques for
XML data. Specifically, a spatial access method R-tree has been utilized
to improve the query performance.  Furthermore, we have identified an
important new kind of XML query, ASJ - aggregate structural join, which
has not been investigated before. A simple and efficient index structure,
XA-tree, has been proposed, along with a few query evaluation techniques,
to evaluate ASJs efficiently.  Finally, by extending the case for single
query to the one for multiple queries, we discuss the directions for
future work that addresses several important issues related to query
processing techniques for multiple queries.


Date:				Friday, 4 April 2003

Time:				3:30p.m.-5:30p.m.

Venue:				Room 2303
				Lifts 17-18

Committee Members:		Prof. Frederick Lochovsky (Supervisor)
				Prof. Dik-Lun Lee (Chairman)
				Dr. Qiong Luo
				Dr. Dimitris Papadias


**** ALL are Welcome ****