When dealing with very to-many large relationships, what are people doing with respect to paging and filtering results? I can see two different approaches:
In the relationship, whether included with a resource or retrieved via a relationship’s
selflink, only provide a
relatedlink that hits a collection that supports whatever filtering and paging is appropriate.
Provide filtering and paging directly at the relationship’s
selflink. This can reduce the request count by one, but doesn’t seem to provide substantial benefits.
What are others doing?