You need ODS in order to use FHIR with Information Exchange.  However, the storage impact can be minimized by using the production settings in ODS, so that data is only placed in ODS when requested, and gets purged after it is no longer being fetched.

Please check the user documentation, at (in my instance):

http://localhost:57776/csp/docbook/DocBook.UI.Page.cls?KEY=HETRM_ch_gen…

You can just search for the word "gender"

We try to not operate on an entire container, since containers can be arbitrarily large.  The API that you are using is one that should be combined with GetNextSDA, to process one streamlet at a time.