i'm not 100% clear on what you're after, but when i had a similar setup, i wanted to be able to get at my home (NATed) data from work. to do so, i have a script called postroot [1] which, when cpu'd into a server posts the root of the calling terminal as foo.root, where foo is the calling terminal. i could then "mount /srv/foo.root /n/foo" or similar to get at my home box.
it was occasionally useful to be able to use the home network connection from outside, too. for that, i have a script called tun[2] which mounts the network connection from a named root (previously posted with postroot) and connects using that. there's no special server needed; this is all using the standard tools. i'm not clear what you mean about giving terminals an updated namespace. you could stick an entry in cron on the NATed system that calls postroot on the real-world server. in that case postroot ought to do a bit more error checking and probably try the existing connection before removing and replacing it. the structure, and possibly the actual scripts (i don't remember) came from someone in #plan9, possibly maht. [1] /n/sources/contrib/anothy/bin/rc/postroot [2] /n/sources/contrib/anothy/bin/rc/tun
PGP.sig
Description: This is a digitally signed message part
