I'm going to of course begin with what Clem and others suggested and do it on a simh 11/45 initially to try to get the needed tcp stuff compiled in.  In looking at  https://minnie.tuhs.org//cgi-bin/utree.pl?file=2.9BSD/usr/net/sys/net/NOTES , it's mentioned that 

>>> The 4K bytes of in address malloc space for dynamic structures is ok for an

>>> I/D machine, but may be tight on a /34 or /23.  Not sure yet whether this

>>> will squeeze in.

, so it feels like there's hope that it can be done.  Lots of rereading and research for me to get to the point of completely understanding that NOTES file.

Anyway, I'm going to try and get a simh instance up somewhere publicly accessible (will provide creds to those curious / interested) and see where I get stuck.  Will be back with more questions!  Thanks again for the initial engagement on this!

jake

P.S.  I do have a Xylogics Annex "terminal server" that'd be a great front end to the real 11/45's serial lines as Clem suggested, but for me, the romance of having the machine truly speaking tcp as intended is one of the goals.  I'll keep the Annex handy for when it's running SysV and other things that definitively can't speak tcp.