University of Konstanz
Database & Information Systems Group
Prof. Marc H. Scholl

XML & Databases


[28.04.2004] PDF-Links
Zur Information: die verlinkten Publikationen sind für externe Zugriffe gesperrt und nur innerhalb der Universität einzusehen.

[24.04.2004] Seminarvorbesprechung
Die Seminarvorbesprechung findet am Dienstag, den 25. April um 12:30 Uhr in Raum C252 statt.


One-time Events


Updating XML Documents

Whereas most commercial XML databases offer ways to update XML the structure and contents of XML documents, the official standard is still on it's way - the Working Draft for XQuery Update was published in January 2006. Because of the different approaches how to store XML documents, the internal mapping of updates can look quite differently. While the Pre/Post encoding is especially efficient for sequentially created filed, updates are already conceptually support in numbering schemes like ORDPATH. - Beside a short introduction to XQuery Update in your talk, you should present and discuss the update approaches for the MonetDB/XQuery encoding and the ORDPATH scheme.

presented by Manuela Grindei

Commercial XMLDBs

X-Hive and Tamino are the two most popular commercial databases that have been especially tailored for XML data. But, naturally, the big three database vendors Oracle, IBM and Microsoft are also following suit now and offer their own XML data platforms. - This talk is about introducing the three systems and discussing their design, architecture and interfaces.

presented by Cordula Bauer

Concurrency in XML

Database processing can be kept simple as long as all requests are sequentially processed, but as soon as the number of transactions increases, concurrency comes into play. Especially for update operations, the parallel execution of read and write operations soon becomes fundamental. First approaches are now developed for XML databases, using multi-granularity locking schemes. - In this talk, a short introduction into transaction and concurrency issues is to be given, followed by possible solutions that have been presented for XML databases.

presented by Daniel Butnaru

XML Compression

Compression algorithms are widely available for all kinds of data, including text, multimedia and generic binary files. In fact XML files have quite a regular structure that is suitable for specific compression algorithms. Some first approaches can be found in the literature, with XMill [Liefke, Suciu] as one of the most popular and XMLPPM [Cheney] as one of the most advanced. - In this talk, you should give an overview on the existing XML algorithms and focus in more detail on the approaches developed by Cheney.

presented by Sebastian Graf

Content Indexing in XML

Indexes are a substantial part in relational databases. In many XML storage systems, however, content indexes are still poorly supported. The first system that fully supported the indexing of values was LORE, based on the DataGuides. Systems like ToXin use similar approaches, distinctly shrinking the query time for selective predicates. - A survey on content index-based systems is to be given in this talk, followed by a discussion on the underlying index structure.

presented by Matthias Röger

Structural Joins, Twig Joins & PathStack

A key operation in processing XPath queries is to efficiently parse single location steps down to the final result set. Research has shown that intermediate context sets can get pretty large, even when input and output sizes are small. The discussed algorithms optimize the calculation time by compactly representing partial results, yielding much faster execution times. - In the talk, the join algorithms are to be introduced and explained.

presented by Irina Andrei

Native XML Storage

Two major approaches have proven to be successful to store XML nodes in databases. One of them chooses a relational encoding to allow XML processing via traditional RDBMS, as it was done in the XPath Accelerator. Alternatively, nodes can also be natively encoded, using a storage format that it optimized for a special XML application. TIMBER is one of the more complex and successful systems of this type. - In the talk, the idea of native storage is to be presented, along with a presentation of TIMBER and one or two other native systems.

presented by Xuan Moc

XML Benchmarks

Benchmarks are popular tools to measure the performance of, e.g., computer software and hardware, file systems and also databases. A number of benchmarks have been developed especially for XML data, with XMark, XBench, X007 and MemBeR being some of the popular ones. - In this talk the most important XML benchmarks are to be presented, including a discussion on their major focus and application domain.

presented by Bok-Him Lin

Seminar Description

XML is still a new and dynamic field of research. Standards for XPath, XQuery Full-Text Retrieval or Updates are still on their way to get confirmed, and for many issues, such as the storage and querying of XML data, no optimal solutions have been found yet. In this seminar some up-to-date research topics will be presented and discussed.