The initial step in any distributed system for data browsing, visualization, and access is data discovery. In small systems the process of data discovery may be to pick data sets of interest from a scrolling list of those available. However, as the number and type of data sets available becomes large a completely manual selection process becomes unwieldy. We have developed a data discovery service that enables users to find data sets based on geographic region, observed quantities, or keyword. The data discovery service obtains its metadata directly from the data sever, thus insuring that the information in the service is complete and up-to-date.

We have based our data discovery service on the Netscape Directory Server a Lightweight Directory Access Protocol (LDAP) server. LDAP servers provide the robustness and scalability, via server replication and distributed servers, required of a data discovery service. Initial clients for this service are a Java application and Web server that access the Climate Data Portal (http://www.epic.noaa.gov/cdp).