Drift correction and framework ions in DiffusionAnalyzer

The drift correction in DiffusionAnalyzer is automatically applied when calculating ionic diffusivity.
(1) Is it always required to include the drift correction for a trajectory?
(2) It makes sense to include it if all the ions moving in the same direction, but what if the ions are randomly moving around, like in an amorphous system?
(3) If I want to calculate the diffusivities of different species in a same trajectory, should I use a same framework? For example, I have an AIMD calculated trajectory for LiAlO2. When using DiffusionAnalyzer to calculate Li diffusivity, Al+O will be the framework and the drift correction will be calculated based on Al+O displacements. Now, I also want to calculate the O diffusivity. Should I use the same framework (Al+O), or use Li+Al as the framework ions and recalculate the drift correction? Using the same framework makes more sense to me as the calculated Li and O diffusivities are based on the same framework.