There is strong evidence that the current implementation of TCP will perform poorly in future high speed networks. To address this problem many congestion control protocols have been proposed in literature which, however, fail to satisfy key design requirements of congestion control protocols, as these are outlined in the paper. In this work we develop an Adaptive Congestion Protocol (ACP) which is shown to satisfy all the design requirements and thus outperform previous proposals. Extensive simulations indicate that the protocol is able to guide the network to a stable equilibrium which is characterized by max-min fairness, high utilization, small queue sizes and no observable packet drops. In addition, it is found to be scalable with respect to changing bandwidths, delays and number of users utilizing the network. The protocol also exhibits nice transient properties such as smooth responses with no oscillations and fast convergence. ACP does not require maintenance of per flow states within the network and utilizes an explicit multi-bit feedback signalling scheme. To maintain stability it implements at each link a novel estimation algorithm which estimates the number of users utilizing the network. Using a simple network model, we show analytically the effectiveness of the estimation algorithm. We use the same model to generate phase portraits which demonstrate that the ACP protocol is stable for all delays.

No abstract

Many congestion control protocols have been recently proposed in order to alleviate the problems encountered by TCP in high-speed networks and wireless links. Protocols utilizing an architecture that is in the same spirit as the ABR service in ATM networks require estimates of the effective number of users utilizing each link in the network to maintain stability in the presence of delays. In this paper, we propose a novel estimation algorithm that is based on online parameter identification techniques and is shown through analysis and simulations to converge to the effective number of users utilizing each link. The algorithm does not require maintenance of per-flow states within the network or additional fields in the packet header, and it is shown to outperform previous proposals that were based on pointwise division in time. The estimation scheme is designed independently from the control functions of the protocols and is thus universal in the sense that it operates effectively in a number of congestion control protocols. It can thus be successfully used in the design of new congestion control protocols. In this paper, to illustrate its universality, we use the proposed estimation scheme to design a representative set of Internet congestion control protocols. Using simulations, we demonstrate that these protocols satisfy key design requirements. They guide the network to a stable equilibrium that is characterized by high network utilization, small queue sizes, and max-min fairness. In addition, they are scalable with respect to changing bandwidths, delays, and number of users, and they generate smooth responses that converge quickly to the desired equilibrium.

No abstract

IntroductionThis papcr prcscnts a ncw activc qucuc managcmcnt (AQM) schcmc that providcs congcstion control in TCPAP networks using a fuzzy logic control approach. The proposed scheme is implemented within the differentiated services (Diff-Serv) framework, providing quality of service (QoS). It is based on the Fuzzy Explicit Marking (FEM) controller [ 11 proposed recently to provide congestion control in TCPAP best-effort networks. The provision of QoS in a Diff-Sew environment requires an adequate differentiation between assured and best-effort classes of service in the presence of congestion, giving priority-preference to assured-tagged traffic. For this purpose, a two-class FEM controller, called FEM IdOut is presented. The proposed fizzy logic approach for congestion control allows the use of linguistic knowledge to capture the dynamics of nonlinear probability marking functions, uses multiple inputs to capture the (dynamic) state of the network more accurately, and can offer effective implementation. The application of fuzzy control techniques to the problem of congestion control in networks is suitable due to the difficulties in obtaining a precise mathematical model using conventional analytical methods, while some intuitive understanding of congestion control is available. A sirnulation study over a wide range of traffic conditions shows that the FEM IdOut (FIO) controller outperforms the Random Early Detection (RED) implementation for Diff-Serv control (FED IdOut -RIO [2]) in terms of link utilization, packet losses, and queue fluctuations and delays. Also, the proposed scheme can offer adequate differentiation among assured and best-effort traffic.