Modelling and Stability Analysis of Binary Abr Flow Control in Atm Network
FY Ren,C Lin,B Wei
DOI: https://doi.org/10.1093/ietcom/e88-b.1.210
2005-01-01
IEICE Transactions on Communications
Abstract:Available Bit Rate (ABR) flow control is an effective measure in ATM network congestion control. In large scale and high-speed network, the simplicity of algorithm is crucial to optimize the switch performance. Although the binary flow control is very simple, the queue length and allowed cell rate (ACR) controlled by the standard EFCI algorithm oscillate with great amplitude, which has negative impact on the performance, so its applicability was doubted, and then the explicit rate feedback mechanism was introduced and explored. In this study, the model of binary flow control is built based on the fluid flow theory, and its correctness is validated by simulation experiments. The linear model describing the source end system how to regulate the cell rate is obtained through local linearization method. Then, we evaluate and analyze the standard EFCI algorithm using the describing function approach, which is well-developed in nonlinear control theory. The conclusion is that queue and ACR oscillations are caused by the inappropriate nonlinear control rule originated from intuition, but not intrinsic attribute of the binary flow control mechanism. The simulation experiments validate our analysis and conclusion. Finally, the new scheme about parameter settings is put forward to remedy the weakness existed in the standard EFCI switches without any change on the hardware architecture. The numerical results demonstrate that the new scheme is effective and fruitful.