> The pair of nodes was chosen randomly at each single test. And, we > didn't collect enough measurements to obtain the same set of nodes > for all the protocols. Oh, so nothing to worry about. This data is clearly invalid. -- Juliusz