To the best of my knowledge there are probably dozens of various proprietary streaming protocols across the various devices and the resources to seamlessly integrate with all of them, if even technically possible would be a nightmare versus the approach of providing a protocol and api and letting each manufacturer include it.
In one of your points you are using Roon to do the heavy lifting and on the first two you are asking for a mixture of both, ie how would you handle a playlist that includes some supported formats on you Aurender and some that aren’t?
I fully agree with the approach Brian outlined and think you may be envisioning it from the wrong perspective. I see my Roon server as the central hub that feeds outward rather than being directed inward and the various endpoints at locations as the items that should be considered around that. To me this approach goes a long way to providing a universal standard for this type of product, an i eagerly await to seeing which companies hop on board to support it.
My decision for future streamer/endpoint purchases would be based around does this product support the Roon Speakers (i hope this name is changed btw, as the initial assumption when seeing it is it is a hardware product versus what it actually is) protocol or not. to me it makes streamers with local storage redundant as they are not necessary when i can have a dedicated server to all of the processing for anything in the house versus doing to processing at each and every point. This allows for the end devices to be much more interchangeable.