Archived: 11 July 2007
Supervisor: Paul Coddington
The Australian Herbaria hold about 6 million records in 8 databases across the country in each state and territory. The demand for querying all this data from a single interface results in the requirement of building a federated database system, which is an integrated and coherent view for a logical collection of independent databases that work cooperatively. The current system, Australia’s Virtual Herbarium (AVH) Version 2, has been implemented based on this architecture. However, some problems have been found during its utilisation.
This project involves designing and implementing the new version of AVH system, Version 3. The new system aims to solve the problems that occurred in the old system, in order to provide a shorter waiting time for the query result, more reliability, a more functional web interface, and use standards that are being developed for biodiversity informatics systems. The new system is built upon a new architecture, which is based on a central index database and using a standard interface to harvest records from each herbarium database with an XML-based data exchange mechanism. In order to accomplish this, some new technologies are adopted in this project, such as the XML technologies for data exchange and data storage, advanced Java web development framework for implementing a flexible web interface and a sophisticated database-class mapping tool for relational database management.
Although the system has been developed for a specific application, the system design and implementation has wider applicability, and the technologies and methodologies can be adopted in other areas as well, such as other scientific applications with large-scale distributed data sets. So this project provides a useful demonstration of the use of federation architecture and relevant technologies, particularly XML-based technologies.