Abstract :
The paper deals with the design of a routing control scheme based on active queue management (AQM). Starting with the packet data network model introduced recently (Fuks, H. and Lawniczak, A.T., Mathematics and Computers in Simulation, vol.51, p.103-19, 1999; Manfredi, S. et al., IEEE Symp. on Circuits and Systems, 2004), we propose a dynamic routing strategy based both on distance and router queue length. The effectiveness, in terms of throughput and packet delivery time, of the proposed routing control algorithm is shown by simulations on a simplified packet data network model. It is given analytical estimates of the critical load leading to the onset of congestion and the worsening of network performance. The estimates are shown to be in agreement with the numerical simulations. Then we validate the proposed AQM-routing in realistic scenarios using the NS-2 network simulator which takes into account also the effects of discretization and the nonlinear nature of communication protocols and traffic.
Keywords :
parameter estimation; queueing theory; telecommunication congestion control; telecommunication network routing; AQM routing control; active queue management; communication network congestion reduction; communication protocols; critical load estimation; discretization; dynamic routing strategy; nonlinear traffic; packet data network model; packet delivery time; router queue length; throughput; Circuit simulation; Circuits and systems; Communication networks; Communication system control; Computational modeling; Computer networks; Computer simulation; Mathematical model; Mathematics; Routing;
