Registry Guide: Harvesting

In the Registry framework, a publishing registry is a registry that can create descriptions of new resources and share it with the rest of the VO. The framework further provides a means for any client to discover and collect resource descriptions of all the resources in the world that are known to the VO. This mechanism starts with discovering all of the publishing registries and pulling from them--i.e. harvesting--all of the resource descriptions created and managed there. Typically, the clients that collect the resource descriptions are so-called full searchable reigstries. These are registries that enable users and clients to search for resources, and in that role, they collect together into their database all know resources.

There are four main features of the Registry framework that enable global harvesting:

1. A publishing registry makes its resource descriptions available for harvesting in the VOResource XML format. (This is mandated by the Registry Interfaces v1.0 standard; the core of the format is defined by the VOResource v1.0 standard.)

2. Clients regularly pull the latest resource descriptions from a publishing registry using the Open Archives Initiative Protocol for Metadata Harvesting (OAI-PMH). The recipe for doing this is outlined by the Registry Interfaces v1.0 standard.

3. Clients can discover available publishing registries by querying a special registry called the Registry of Registries (RofR) using OAI-PMH. The recipe for doing this is spelled out in an IVOA Note called The Registry of Registries.

4. A publishing registry administrator lets the world know it has resource descriptions to share by registering itself into the Registry of Registries via a web form. As part of the registration process, the publishing registry is checked for compliance to the IVOA registry standards.