There are different ways to test mobile timing – the model from MicroFocus tries to describe the different contributing factors to latency/bandwidth & offers different possibilities for testing them. My recommendation is to ask engineers about the range of reasonable values for each stage & whether or not there are circumstances that cause the variation of those values to be correlated (if correlation is present, then especially bad combinations are more likely than if if they are independent). Thorough testing should examine the worst cases that are within the target range of service. Knowing what is normal & how bad things can be is perhaps more relevant than an empirical, but limited sampling of some practice cases.