The wavlengths available are dependent, not the wavelength itself. A photon between barriers has fewer modes available, not more - this effect is seen in cavity QED experiments.
The size of the universe is not infinite, AFAIK - the Hubble constant and WMAP, etc. point to a finite age and extent. If you have a references that say otherwise, please provide a link.
Yes the phton between two barriers has fewer modes but surely those modes are independent of the modes avaidable to a unbounded photon meaning that they may be in the range of ‘forbidden’ modes. Though I take back what I said about non-locality as thinking on it more the expansion is non-local.
The universe appears to be flat from observation, as you should know (I see you work at the USNO) a flat universe in most standard cosmological models is infinite in extent (though there is a 3-torus shape that produces a flat finite universe, but an infinite universe is seen as more likely).