Consider three systems. A steel plate, a total vacuum (which can be imagined as a volume in the parking lot of the same size and shape as the plate for all practical purposes; i.e., classical quantum mechanics), and a photon.
If there is no steel plate, the context is only that of light. One can speak of curvature as either light on light interaction or gravity; they are words for the same observation.. Einstein rejected Newton's gravity as "action at a distance". Since there is no matter in the volume in the parking lot, the speed of light is a constant, c. If there is matter, than one can model it by a curved line and constant c, or just simply say "it takes longer" by ignoring the path and only considering the source and sensor.
If there is a real steel plate, then for classical physics (light and matter) which is what we're discussing here in, STR is irrelevant; Quantum mechanics introduces Planck's constant to describe photo-equivalent electrons in terms of the deBroglie wave in terms of the Schroedinger equation - the simplest form is:
Consider the quantity (Px - Et) and compare it with its relativistic expression:
Then , which shows that STR and the speed of light (c) as a "Universal Constant" is irrelevant in the case of classical quantum mechanics (which applies to the world in which we try to send photons through steel plates).
For QM, the universal constant in context is Planck's constant h, and only the difference between Px and Et is relevant; if Px = Et (the vacuum) the result is the same as for STR
The DeBroglie wave length then characterizes the system, which will change according to P and E, where x and t have their normal interpretations as space and time.
The steel plate. the volume in the parking lot, and the velocity of light must be explained classically.
In the parking lot and the steel plate, the velocity of light light must be explained classically, not relativistically, and in a vacuum (the parking lot) is given by:
from Maxwell's equations by considering the linear displacement current between two plates of an imaginary capacitor in terms of Coulomb's and Ampere's laws, and models the propagation of light in terms of E and B fields.
However, within the steel plate, the velocity is given by:
where the change in the speed of light is represented by D and H fields instead, due to the changes in permittivity and permeability.
The speed of light therefore slows down in matter in the real world.,..
Your model seems to be of that of an atomic gas, where electrons are absorbed/re-emitted locally at each atom, with the only criteria being the delay in the process. But if no energy is absorbed/emitted (the plate doesn't heat up), the photons and atoms don't interact, and the process from one side of the plate to the other doesn't change (Px - Et) = 0. That is, the steel plate might as well not be there.
Real matter is not modeled in this way. A lattice is modeled in terms of overlap integrals (Bloch functions), where the individual transitions are included the field of atoms plus overlap in terms of Fermi energy levels. Energy is deposited/re-emitted in terms of acoustic vibrations - the actual process depends globally on the arrangement of atoms in the lattice.
Relativistic quantum field theory is a different context. In this case, point interactions are modeled as "events", where the lines are straight lies if there are no events; otherwise, the matter and energy are modeled by Feynman propagators at each interaction site. Again, the only consideration here is the CHANGE in energy at each site, so the probability has to be re-normalized, which conflicts with the idea of the existence of initial and final states (one doesn't have to calculate what happens at the bottom and top of the diagram). (This is the QFT "cutoff" where the whole system is the model of photon-equivalent particle collapse - the Big Bang/Crunch).