This paper examines the performance of TCP/IP, the Internet data transport protocol, over Wide Area Networks (WANs) in which data traffic could coexist with real-time traffic such as voice and video. Specifically, we attempt to develop a basic understanding, using analysis and simulation, of the properties of TCP/IP in a regime where (1) the bandwidth-delay product of the network is high compared to the buffering in the network, and (2) there may be transient congestion due to fluctuations in real-time traffic, modeled here as producing random losses among the packets of the TCP connection of interest. The following key results are obtained. First, random loss leads to significant throughput deterioration when the product of the loss probability and the square of the bandwidth-delay product is larger than one. Unless network resources are specifically reserved for data traffic, data traffic will inevitably incur random losses due to transient fluctuations in higher priority real-time traffic when the network is highly utilized. Second, for multiple connections sharing a bottleneck link, TCP is grossly unfair towards connections with higher round-trip delays. This means that a simple First In First Out (FIFO) queueing discipline might not suffice for data traffic in WANs. Finally, we observe that, while the recent Reno version of TCP produces less bursty traffic than the original Tahoe version, it is less robust than the latter when successive losses are closely spaced. We conclude by indicating modifications that may be required both at the transport and network layers to provide good end-to-end performance over high-speed WANs.

This paper presents a control-theoretic approach to reactive flow control in networks that do not reserve bandwidth. We assume a round-robin-like queue service discipline in the output queues of the network’s switches, and propose deterministic and stochastic models for a single conversation in a network of such switches. These models motivate the Packet-Pair rate probing technique, and a provably stable rate-based flow control scheme. A Kalman state estimator is derived from discrete-time state space analysis, but there are difficulties in using the estimator in practice. These difficulties are overcome by a novel estimation scheme based on fuzzy logic. We then present a technique to extract and use additional information horn the system to develop a continuous-time system model. This is used to design a wuisnt of the control law that is also provably stable, and, in addition, takes control action as rapidly as possible. Finally, practical issues such as correcting parameter drift and cmmlination with window flow control are described.

This paper uses a variety of mathematical models to explore some of the consequences of rapidly growing communications capacity for the evolution of the Internet. It argues that queueing delays may become small in comparison with propagation delays, and that differentiation between traffic classes within the network may become redundant. Instead, a simple packet network may be able to support an arbitrarily differentiated and constantly evolving set of services, by conveying information on incipient congestion to intelligent end-nodes which themselves determine what should be their demands on the packet network.

Dynamics of Internet load are investigated using statistics of round-trip delays, packet losses and out-of-order sequence of acknowledgments. Several segments of the Internet are studied. They include a regional network (the Jon yon Neumann Center Network), a segment of the NSFNet backbone and a cross-country network consisting of regional and backbone segments.

We present a new technique for streaming real time video on today's Internet, based on dynamic rate shaping and TCP congestion control. Dynamic rate shaping is a signal processing technique that adapts the rate of compressed video (MPEG-1, MPEG-2, H.26x) to dynamically varying bandwidth constraints. This provides an interface (or filter) between the source and the network, with which the encoder's output (either live or stored) can be perfectly matched to the network's available bandwidth. We couple this adaptation capability with the use of a new semi-reliable protocol that uses the TCP congestion window to pace the delivery of data into the network, but without using other TCP algorithms that are poorly suited to real time media. Use of TCP congestion control ensures that the protocol competes fairly with all other TCP data, and that it optimally shares the available bandwidth. It also avoids the latency problems commonly associated with TCP. In addition, we describe a rea...

The congestion control algorithm embedded in the 4.3-Tahoe BSD TCP implementation has dramatically improved congestion control over the Internet. However, several recent simulation studies on the dynamics of this algorithm has revealed that the algorithm exhibits clear oscillatory patterns in sending window size, round trip delay and bottleneck queue length. In this paper, we present a new congestion signal scheme and a dual traffic adjustment strategy. Simulation results show that our modifications can eliminate the periodic packet losses and substantially reduce the traffic oscillation.

In ATM networks congestion control plays an essential role to support different quality of service guarantees for a large variety of traffic types. While for real-time communications, such as CBR and VBR services, a preventive open-loop congestion control mechanism is applied, a reactive closed-loop mechanism has been suggested for the ABR service class, which was introduced aiming at data communication applications. This paper presents a performance analysis of the ratebased congestion control mechanism developed by the ATM Forum. Using a differential equation approach for the computation of the evolution of the allowed cell rate and the buffer content during steady state, we derive closed-form expressions to estimate the maximum buffer lengths. These expressions can be formulated for steady-state conditions as well as for transient phases. In our approach, scenarios with and without priority for the resource management cells are considered. From numerical examples we conclude that an...

This paper gives a class of flow control algorithms for the adaptive allocation of bandwidths to virtual connections (VC) in high speed, wide area ATM networks. The feedback rate to the source from the network is parsimonious, with each feedback bit indicating whether the buffer at a distant switch is above or below a threshold. The service discipline at the switch is First-Come-First-Served. The important goal of adaptability aims to make all of the network bandwidth available to the active VCs, even though the number of such VCs is variable over a given range. Each VC has two parameters, one giving its minimum guaranteed bandwidth and the other is the weight for determining its share of the uncommitted bandwidth. Judicious selection of these parameters defines distinctive services, such as Best Effort and Best Effort with Minimum Bandwidth. We derive design rules for selecting the parameters of the algorithms such that the appropriate guarantees and fairness properties are exhibited ...

We present an approximate analysis of a queue with dynamically changing input rates that are based on implicit or explicit feedback information. This is motivated by proposals for adaptive congestion control algorithms [RaJa 88, Jac 88], where the sender's window size at the transport level is adjusted based on perceived congestion level of a bottleneck node. We develop an analysis methodology for a simplified system; however, it is powerful enough to answer the important questions regarding stability, convergence (or oscillations), fairness and the significant effect that delayed feedback plays on performance. Specifically, we find that, in the absence of feedback delay, the linear increase/exponential decrease algorithm of Jacobson and Ramakrishnan-Jain [Jac 88, RaJa 88] is provably stable and fair. Delayed feedback, on the other hand, introduces oscillations for every individual user as well as unfairness across those competing for the same resource. While the simulation study of Zhang [Zha 89]...

We consider a flow control mechanism that dynamically regulates the rate of data flow into a network based on feedback information about the network state. Such mechanisms have been introduced recently in a variety of networks including the Internet, and have been advocated for future high-speed networks. We first model the flow control mechanism by a discrete-space stochastic process and define appropriate performance measures for transient and steady-state regimes. However, the model does not appear to be analytically tractable and we study it through simulation. We then simplify it to a continuous-space deterministic (or fluid) model for which we can easily derive closed-form solutions. We find the analytical results for the fluid model to agree well with the simulation results obtained using the discrete-space model. Both models explicitly consider delay of the feedback information, thus making them relevant for high-speed networks. 1 Introduction In a computer network, packets g...