Since the beginning of the project, there has been the need to share metadata among several GeoNetwork nodes. Usually, each node takes care of a region of interest so it is important to be able to perform a search over all these nodes at the same time. This is called distributed search and exploits the Internet connectivity. In our cases, this distributed search can be heavy to perform if there are many maps with associated thumbnails. Furthermore, GeoNetwork is usually employed in countries (like Africa, Brazil) where the connectivity is bad, making the use of distributed search not feasible.
The harvesting is the process of collecting remote metadata and storing them locally for a faster access. This is a periodic process to do, for example, once a week. Harvesting is not a simple import: local and remote metadata are kept aligned. Using some magic, GeoNetwork is capable of discovering metadata that have been added, removed or updated.
GeoNetwork is able to harvest from the following sources (for more details see below):
Another GeoNetwork node (version 2.1 or above).
A web DAV server.
A CSW 2.0.1 server.
An old GeoNetwork 2.0 node.
An OAI-PMH server.