Usually refers to the random sequence of customers leaving a queueing service center. More generally, it is the random point process or marked point process with marks representing aspects of the departure stream and/or the service center or node from which they are leaving. For example, the marked point process (X^d, T^d) for departures from an M/G/1 queue takes X^d as the Markov process for the queue length process immediately after the departure time and T^d is the actual time of departure.