Fuzzy TCP: Optimizing TCP Congestion Control

Abstract

the slow-start algorithm begins with sending one segment, it takes many round-trip times to reach the Efficient implementation of TCP for the Internet optimal operating point, thus resulting in poor requires a precise determination ofcongestion window utilization of the available bandwidth for short by the source TCP agent. This paper proposes a fuzzy transfers which are small compared to the bandwidthimplementation of TCP (Fuzzy TCP), instead of delay product of the path. Second, since a TCP sender current slow-start /congestion avoidance approachfor has no knowledge about the capacity of the available efficiently determining size ofthe congestion window at resources on the networks and uses default parameters each time a new acknowledgement receives. The at the beginning of transmission, the exponential scheme uses the current size ofcongestion window and growth of the congestion window often misleads the the slow-start threshold (ssthresh) as well as the last sender to send too many packets too quickly, thus and previous round trip times. The proposed fuzzy causing a severe buffer overflow at the bottleneck link. controller updates the size of congestion window using This buffer overflow results in multiple packet loss a fuzzy system. The approach is applied to the source from a window of data. When multiple packets in the agent and the destination TCP remains unchanged. same window are lost, only one of the packet losses may be recovered by each fast retransmit; the rest are

Fuzzy Controller based TCP-Vegas Enhancement for Congestion Control

“…Research results have shown that singleton fuzzifier, triangular and trapezoid input membership functions, center average defuzzifier and product inference engine works best for the fuzzy controller [9,11,13,14].…”

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

“…The proposed fuzzy controller has two input parameters sc and dr, and a fuzzy controller output parameter dcwnd. Where, sc = (ssthresh-cwnd) / ssthresh and dr = (average RTT-RTT) / (0.5(average RTT + RTT)) [13]. The sc parameter is the first FLC Member Ship Function (MSF).…”

“…The proposed fuzzy controller uses singleton fuzzifier, product interference engine, centre average defuzzifier [13] and the proposed fuzzy rules is shown in Table 1.…”

Esquema de controle adaptativo de tráfego de redes baseado em um algoritmo de predição fuzzy

“…Várias técnicas de controle de congestionamento em redes de computadores têm sido propostas na literatura (Wang et al, 2007;Durresi et al, 2006;Nejad et al, 2006;Karnik and Kumar, 2005). Dentre as propostas de controle de congestionamento utilizando lógica nebulosa, algumas utilizam modelos fuzzy sem adaptação dos parâmetros como em Hu and Petr (2000) e outras são baseadas em protocolos ou tecnologias de rede específicos (Nejad et al, 2006;Chen et al, 2003).…”

“…Dentre as propostas de controle de congestionamento utilizando lógica nebulosa, algumas utilizam modelos fuzzy sem adaptação dos parâmetros como em Hu and Petr (2000) e outras são baseadas em protocolos ou tecnologias de rede específicos (Nejad et al, 2006;Chen et al, 2003). No primeiro caso, muitos dos esquemas não são suficientemente precisos em prever o comportamento variante do tráfego gerado por aplicações em tempo real devido a não adaptação de seus parâmetros (Pitsillides and Lembert, 1997).…”