ABSTRACT
In this chapter, we examine the use of fuzzy logic systems as rate-based feedback congestion-control mechanisms in Asynchronous Transfer Mode (ATM) networks. The information used by the fuzzy logic systems is based on the short-term observation of the network status and the system’s acquired experience. A feedback value, which is between 0 and 1, is generated by the fuzzy logic systems and is sent back to all the active sources to throttle their transmission rates for every fixed time interval. Two source rate regulation mechanisms are also proposed. In the Rate Reduction (RR) mechanism, the current transmission rate of an active source is reduced by the value of the feedback signal. In the Rate Reduction Rate Increment (RRRI) mechanism, we assume that there is an infinite buffer space at each source terminal, where cells are put with no rate regulation. When congestion is detected at the network, the source buffers transmit cells to the network at the same reduced rate as in the case of the RR mechanism. If the bandwidth of the server is plenty enough, the source buffers transmit cells at the speed of the peak cell rate. An experimental study shows that the proposed fuzzy logic control systems have lower time delay as well as lower cell-loss rate compared with traditional feedback control mechanisms. The RRRI mechanism has better performance than the RR mechanism at the cost of extra buffer space each source.
