Time-Based Media Preservation Repository Use Cases:
Ingest of large audio and video files
Title (goal) | Ingest of large audio and video files |
|---|---|
| Primary Actor | Audio/video digitization workflow scripts |
| Scope | |
| Level | |
| Story |
2. Fedora projects over content and metadata stored in HPSS and on filesystem in order to manage fixity and mediate access. |
Dissemination of large audio/video files
Title (goal) | Access to / dissemination of large audio/video files |
|---|---|
| Primary Actor | Collection manager/curator |
| Scope | |
| Level | |
| Story | Collection manager finds identifier for audio/video object for which he/she wants to download master or mezzanine file (managed in Fedora via projection onto HPSS) and makes request for file through locally-developed web application, having authenticated via CAS. Application and/or Fedora are able to authorize download of file based on user's identity, user/group permissions on object, and role/group membership information obtained via LDAP from Active Directory. File (may be as large as 200GB) is downloaded from HPSS to user-designated location. Example of authorization: Master files from Lilly Library collections may be downloaded by users who have the role of Lilly Library staff. This role is determined based on the user's membership in a LillyStaff group in Active Directory. |
Fedora Triplestore Use Cases:
Easy traversal against the triplestore
Title (goal) | Easy traversal against the triplestore |
|---|---|
| Primary Actor | |
| Scope | |
| Level | |
| Story | a. bottom up: Given an object, list all the parents of all levels b. top down: For a collection, we would like to know hierarchical structure of the object, even the relationship stored on its children. would like to know how many children of the object and how many grandchildren for each of its children... c. statistical information for the objects with particular datastream on repository level: for example, we need to know how many items have a PDF file for a particular collection. A great many example queries that have proved useful in the past are documented here. |
Support Flexible data model change
Title (goal) | Support Flexible data model change |
|---|---|
| Primary Actor | |
| Scope | |
| Level | |
| Story | Easily insert/delete one layer of the existing object hierarchy. |
2 Comments
Scott Prater
Concerning the dissemination of large audio/video files: can you provide a concrete example flow? I.e., does authentication happen through Fedora and LDAP, or somewhere else? How would the username (principal) come into the Fedora session? How do you imagine expressing permissions on a Fedora object? Would it be role-based (this object allows anyone who has a "reader" role to read the object)?
You can look at Access Roles Module for an example flow written there, and for some ideas of where LDAP would fit in a role-based access scheme.
Greg Jansen
I'm not sure we need quite that much detail to understand the use case. However, it sounds like group memberships are coming from LDAP. It also sounds like users and groups need to be assigned permissions within the repository, at least permission to view and download large media streams.
To me it sounds as if the LDAP group principals are the key missing ingredient to support this case. Once we have that principal in our enforcement context (via principal factory), then we can enforce roles that are assigned to LDAP groups.