Intelligent control techniques for efficient regulation of ABR queue length in ATM switches

Designing effective congestion control strategies for broadband networks is known to be difficult because of the variety of dynamic parameters involved such as link speeds, burstiness of the traffic, the distances between traffic sources and switching nodes. The uncertainties involved cause quantitative modeling of broadband networks almost impossible. Qualitative modeling techniques using the principles of fuzzy logic would lead to effective solutions. As part of our ongoing research program focusing on the development of intelligent control techniques for ATM networks, in this paper we present a regulator based on fuzzy logic control techniques for controlling congestion in broadband networks by keeping the ABR queue length around a reference value. Our simulations show that the controller is quite effective especially under network overload conditions.