I've found this, however, it sound suspicius in multcomponent:
"The diffusion coefficient D corresponds to the solution which means that it is the same for both the electrolyte and the solvent. Certainly the diffusion process involves both substances as it consists in a mutual interdiffusion of solute and solvent."
The condition of equivalent diffusivity in liquid systems is only true when dealing with thermodynamically ideal solutions where the density of the system remains constant. Since electrolytes and solvent will often have large differences in their densities, mutual diffusivity generally only applies in very dilute systems. As mentioned above, this does not apply in multi-component systems.
mutual diffusivity as my understanding also referring to the transportation of A to B, thus we can say that diffusivity DAB. it refers and depends on the mechanism of our process.
For two-component systems the relation between mutual diffusion of solvent and solute is trivial. For multicomponent systems, it is possible to extract the solvent diffusivity from multicomponent diffusion and volumetric data. It is a sort of change of reference frame. In such a way you can evaluate the extent of counterflow.
A general reference is:
Miller, D. G.; Vitagliano, V.; Sartorio, R. J. Phys. Chem. 1986,
90, 1509-1519.
Whereas, a more specific reference on the counterflow is: