Congestion Control in TCP/IP Using Fuzzy Logic

Congestion occurs in the network when arrival rate to a router is greater than its departure rate. In this paper, using fuzzy logic approach, we have proposed a modified TCP delay-based congestion avoidance mechanism which is based on traditional TCP-Africa algorithm. Here we present our fuzzy controller which is expected to act as a congestion controller in the routers. The proposed fuzzy controller uses queue delay and link capacity as input linguistic variables. The output of fuzzy controller is the window size to which the sending window must be adjusted. Both the current standard and most of the experimental congestion control methods are fundamentally loss-based, that is they rely on packet loss to detect that the network is above full capacity.