Sp-eprca: an Atm Rate Based Congestion Control Scheme Based on a Smith Predictor

This report presents a feedback control algorithm for ATM congestion control in which ABR/UBR source rates are adjusted according to VC queue levels at intermediate nodes along the path. The goal is to fully and promptly utilize the available bandwidth left by QOS constrained traac (CBR+VBR) for transmitting the less QOS stringent ABR/UBR traac, without incurring in queue overrow and consequently cell loss. In order to obtain this, we propose a simple and classical proportional controller, plus a Smith Predictor to overcome instabilities due to large propagation delays. In the SP-EPRCA model, each queue ideally behaves as a simple rst order dynamic system with a delay in cascade. The delay is out of the feedback loop, and therefore does not aaect stabilitity. Moreover, since the system dynamic is a rst order one, it is not only stable but it has no damped oscillations as well. The real SP-EPRCA implementation, however , may behave as a higher order system due to innacuracies in estimating propagation delays. We show in which conditions system stability is guaranteed under this more realistic scenario. Finally, we propose an eeective SP-EPRCA implementation in which each source computes its input rate based on the maximum queue level along the path, which is compatible with the current EPRCA type UNI speciications. Theoretical and experimental results show max-min fairness of the proposed control scheme, its eeciency under the constraints of the EPRCA implementation, as well as its cell loss free property. Moreover, the relationship between throughput, buuer size, and round trip delay is established. Finally, various queue services disciplines are considered in conjunction with the basic scheme.