Actually, that would be 3 engines then. The one that does cfdocument is neither DDX nor the add-on service and its PDFG web app (which serves the cfhtmltopdf), which runs as a separate process from CF.
Instead cfdocument processing is done via an embedded "ice browser" feature within CF--that runs literally in process with CF. Not big-timing or showing off (Mark knows that about me), just offering further clarification--to anyone who may care to know. :-)
And the "limiting" (allowing only one request at a time, across a set of just several tags) is indeed a surprising gotcha on Standard, for folks who hit it (and realize it's what's causing a hangup for such requests). FWIW, I discuss it more in a preso I did on the many (many) differences between CF Std and Enterprise:
https://www.carehart.org/presentations/#differences_std_ent
Hope that's helpful to someone.