Fuzzy based Tuning Congestion Window for Improving End-to-End Congestion Control Protocols

Transmission Control Protocol (TCP) is the transport-layer protocol widely used in the internet today. TCP performance is strongly influenced by its congestion control algorithms which limit the amount of transmitted traffic based on the estimated network capacity to avoid sending packets that may be dropped later. In other words Congestion Control is Algorithms that prevent the sender from overloading the network. This paper presents a modified fuzzy controller implementation to estimate the network capacity which reflected by congestion window size. Fuzzy controller use Round Trip Time "RTT" as network traffic indication as well as current window size and slow start threshold "ssthresh" as currently occupied bandwidth indicator. NS2 used as a simulation tool to compare proposed fuzzy approach with most widespread congestion control protocols including; TCP-Tahoe, Reno, New Reno, and Sack. Simulation results show that the proposed mechanism improves the performance against throughput, packet drop, packet delay, and connection fairness.