Reconsidering microscopic mobility modeling for self-organizing networks