Voice/data transmission in packet form: a study of delay performances

Methods corresponding to two types of service discipline are presented for the computation of the probability distribution function of the queue length in a statistical multiplexer in the presence of a mixed voice/data traffic. The model chosen for the description of the speech patterns on the voice lines was a three-state Markov chain model, yielding hyperexponentially distributed silence periods and exponentially distributed active periods. Although the computations involved in these methods were quite important (due to the use of matrices whose numbers of elements were growing as a function of the fourth power of the number of vice sources), they presented no technical difficulties and could be carried out for the desired sets of parameters. The results obtained for different values of the parameters corresponded to what was expected.<>