My understanding is that if you a are building a Connector that
produces a Content-Feed, then the GSA will need to use both the
Authentication and Authorisation SPIs. For a Metadata-and-URL feed,
neither is required because the GSA doesn't require the Connector to
be able to log in as the end-user, it makes those requests directly to
the webserver using the URL you supply it.
The connector will still probably still need to log in as some sort of
system user against wherever the documents are coming from, but that
isn't what those API's are doing.
- Brad