Mirroring the IRI Data Library using DODS and THREDDS

March 2003

We have set up a mirror of the IRI Data Library using DODS/OpenDAP and THREDDS to transfer the data, metadata, and collection structure between the two sites. We are using standard HTTP software to cache the metadata, and custom code to incrementally cache the data on the mirror site. Once the code is more fully developed, a similar system should be able to cache from the entire collection of THREDDS servers.

A mirroring scheme has three key parts: a server, a client, and a cache.

Server	-	serves datasets using THREDDS and DODS
Client	-	views, manipulates, and re-serves a THREDDS/DODS dataset collection
Cache	-	maintains a local copy of the information
Suggestions	-	Our experience leads to some suggestions for both the THREDDS and DODS/OpenDAP standards.

Note: both the links above and the links on the left hand menu let one go from page to page. The links on the left-hand menu will continue to appear on subsequent pages.