Investigation of queue utilization on network routers by the use of dynamic models

Results of investigation of impact factors to queue utilization dynamics on telecommunication network routers were presented. As was shown by the analysis, use of dynamic model, obtained by applying a Pointwise Stationary Fluid Flow Approximation (PSFFA), demonstrated that the average queue length converges to a steady state value for some time, which can range from few to tens of seconds. It is determined that the use of the steady state estimations when calculating the average queue length is possible only after the end of the transient process. Otherwise, it is advisable to use a more accurate differential model. As the results of the analysis, the duration of the transient process is influenced by such factors as the flow rate, capacity of the router interface, service discipline, etc. Within this research it was shown that the use of non-linear differential model can improve the accuracy of the calculation of the average queue length, depending on the state of the interface and the selected service discipline.