System Communication and Connectivity

VegBank Server System.
The three component databases (plot, community, and plant taxon entities) have been designed and implemented on the Oracle 8i, Postgresql 7.1, and HypersonicSQL RDBMS. In addition to the back-end database implementations, software has been written to allow web access. This software allows a user to query vegetation data for multiple attributes stored in the databases. Queries are stored in the user’s profile records for future reuse or modification.

Embedded VegBank Client System.
We have developed a Java desktop client tool that has two purposes: (1) to connect with the application server (using http) in order to perform operations more extensive than those provided to browser-clients (such as data-validation, and data transformation); and (2) to manage vegetation data locally, at the workstation level and in the field. Users of this application client tool can store personal plot data as well as plot data extracted from VegBank in an embedded database identical in design to the database resident on the VegBank server. The embedded client solution is the pure-Java HypersonicSQl RDBMS, which allows it to be run on any operating system.

System architecture.
This web-based system is based on a typical three-tiered architecture that is comprised of a client, application-framework, and a back-end database. The middle-tier was implemented using Java Servlet technology, XML (eXtensible Markup Language) for communication, and Java DataBase Connectivity (JDBC). Due to the open and flexible nature of the implementation tools, we can and do deploy the VegBank enterprise on a number of platforms and database engines. The data loader framework currently being used leverages a Java plug-in framework that allows a flexible interface for legacy data to be inported into the VegBank database. We have written loader plug-ins for the USGS-NPS Vegetation Mapping Program Plots Database, the TurboVeg database system (Hennekens & Schaminee 2001) and the Carolina Vegetation Survey protocol (Peet et al. 1998). In addition, we are designing a more flexible class to load data stored in a variety of ASCII formats. Simple form-driven data entry will also be supported. Data can be extracted from the VegBank server in XML, and we will soon implement a set of convenient alternative formats to facilitate use with common analytical tools.