The rethinkdb itself has persistence, no?
Your web2py server could consume results from a rethinkdb query without saving anything in the RDBMS (postrgres, mysql, etc), just as it could consume information from another website (
dowjones.com,
cnn.com, etc) with or without saving anything locally. I'm not coming up with a use-case for it at the moment, but I have consumed services from a power-switching device, so I won't be surprised if someone else comes up with something.
I would be inclined to have the rethinkdb query come from my client code, in the javascript on a web page or within a mobile app, but I'm sure someone has more imagination that I do.
/dps