Analysis of a kind of 2 parallel queueings with precedence for multimedia network communication

In multimedia network communication we should guarantee that the control information can be transmitted in time and as accurately as possible. First, we propose a parallel queueing model with precedence, which makes sure that the high priority packets are dealt with within constant time no matter how the traffic load is. On the other hand, it ensures that the lost probability of the high priority packets is also very low. Next, we use MMM (mathematic model method) to analyse the system and use SOR (successive over-relaxation method) to compute the state probability of the system. Then we compare the precedence queueing scheme and the conventional queueing scheme where all packets are serviced at the same priority level. Then a conclusion is drawn that the former one takes obvious advantage than the latter one in the fields of queue length distribution, average blocking probability, and average waiting time. We also give out how to realize this proposal in the real application and that is by looking up a table constructed before. Finally we provide simulation results to help verify the validity of analysis. It is shown that we can get very good performance of high priority packets while the performance of low priority packets will not be very bad