The Corporation for National Research Initiatives (CNRI) is developing the Handle Server and the Repository to work both together and independently.
The handle server can provide address lookup for any item or resource on the Internet that has a registered handle. In the most obvious case, the record associates a handle with the URL for the resource.
If an item has been deposited in a CNRI-style repository, its handle-server record will point to the repository.
When a program presents the handle, the handle server will respond that the item can be retrieved from the named repository using the Repository Access Protocol (RAP). The program will then use the RAP protocol to communicate with the repository and negotiate access subject to applicable permissions and restrictions.
The repository will use the same handle to identify the item. In fact, if the user's program "knows" which repository an item is stored in, it need not consult the handle server at all. It can start the RAP conversation with the known repository immediately.
Restrictions that may apply to objects in the repository
In this article, based on a presentation at the April 1995 meeting of the Coalition of Networked Information (CNI), Bill Arms of CNRI introduces the handle-server and repository in simple terms and diagrams.
This paper provides a much more formal description of the concepts behind the repository architecture and the Repository Access Protocol (RAP).
Previous -- Up -- Intro -- Index
Handle Server:4 --
(1/19/96)