Figure 1.
Design of the original vertebrate biodiversity networks, some of which are still active today, used a three-tiered system in which portals are connected to a layer of servers using the DiGIR protocol.
This architecture requires hundreds of individual servers and hundreds more connections between them and the portal. The result is a network in which each element is a potential point of failure. The six portal servers consisted of the four shown and two additional mirror portals for the Mammal Networked Information System (MaNIS). Key: M = MaNIS, O = ORNIS, H = HerpNET, and F = FishNet II.
Figure 2.
This screenshot of IPT shows how users map their local field headings to Darwin Core terms, an essential task for data publishers.
The Darwin Core term names are on the left and terms loaded from a database or spreadsheet on the right, which are selected using dropdown menus. Fields that have the same name string in both Darwin Core and the publisher dataset are matched automatically, while those that do not match must be selected manually (via adrop-down list) by the “data publisher”.
Figure 3.
Example of an IPT summary page displaying some of the metadata provided for the dataset hosted by VertNet for the Cowan Tetrapod Collection of birds
(http://ipt.vertnet.org:8080/ipt/resource.do?r=ubc_bbm_ctc_birds).
Figure 4.
The current workflow for biodiversity data networks has multiple steps that separate the publishing of datasets from downstream aggregation and enhanced discoverability.
The IPT supports the creation and publication of Darwin Core Archives accessible for download, with a publicly available summary web page. Aggregators harvest, process, and upload Darwin Core Archives into systems effective for searching, filtering, visualization, and download.