Diffusion Model of RED Control Mechanism

We present a diffusion model of a network node controlled by RED mechanism used to indicate congestion but not to delete packets. Diffusion approximation allows us to study the dynamics of flow changes introduced by this mechanism in a more efficient way than simulation. After introducing some basic notions on diffusion approximation and on our approach to solve diffusion equations analytically or numerically, we present a closed loop model of flow control and investigate the influence of delay and of control parameters on performance of the system. Also FECN/BECN scheme is considered: flow remains constant within an interval of fixed length and is changed in next interval if the number of marked packets during the interval is above a certain threshold. Diffusion results are validated by simulation.

[1]  Tadeusz Czachórski,et al.  Markovian model of RED mechanism , 2001, Proceedings First IEEE/ACM International Symposium on Cluster Computing and the Grid.

[2]  B. Jouaber,et al.  Modeling the sliding window mechanism , 1998, ICC '98. 1998 IEEE International Conference on Communications. Conference Record. Affiliated with SUPERCOMM'98 (Cat. No.98CH36220).

[3]  D. Zwillinger.  Handbook of differential equations , 1990 .

[4]  Van Jacobson,et al.  Random early detection gateways for congestion avoidance , 1993, TNET.

[5]  Erol Gelenbe,et al.  On Approximate Computer System Models , 1975, JACM.

[6]  Ioannis Lambadaris,et al.  Performance Evaluation of Explicit Congestion Notification (ECN) in IP networks , 2000, Multimedia Information Systems.

[7]  Tadeusz Czachórski,et al.  A diffusion model of the dynamic effects of closed-loop feedback control mechanisms in ATM networks , 1999 .

[8]  Gordon F. Newell,et al.  Applications of queueing theory , 1971 .

[9]  Martin May,et al.  Analytic evaluation of RED performance , 2000, Proceedings IEEE INFOCOM 2000. Conference on Computer Communications. Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies (Cat. No.00CH37064).

[10]  Tadeusz Czachórski,et al.  Diffusion models of leaky bucket and partial buffer sharing policy: transient analysis , 1996, Modelling and Evaluation of ATM Networks.