Estimation of TCP Retrasmission Timer Using Fuzzy Logic Controller

The TCP protocol is still the dominant transport protocol in the internet. One of the key responsibilities of T.C.P is network congestion control. One of the fundamental transport estimation problems is, setting the retransmission timer (RTO) to a rational value. In this paper we propose an optimization of current RTO estimation algorithm. Bt adding a detection mechanism for bad timeouts and undoing their side effects we are willing to have an aggressive estimator, by using of fuzzy system to adapt the K parameter (of RTO estimation formula) to network internal state, we try to answer the following questions: how much and when should our estimator act aggressively? We implemented our algorithm in NS2 and compared the simulation results of the proposed fuzzy aggressive RTO estimator (FARE) to results of current estimation algorithm. The simulation show the superiority of TCP using FARE over a normal TCP.