Abstract—The steady-state performance of a bulk transfer TCP flow (i.e., a flow with a large amount of data to send, such as FTP transfers) may be characterized by the send rate, which is the amount of data sent by the sender in unit time. In this paper we develop a simple analytic characterization of the steady-state send rate as a function of loss rate and round trip time (RTT) for a bulk transfer TCP flow. Unlike the models in [7]–[9], and [12], our model captures not only the behavior of the fast retransmit mechanism but also the effect of the time-out mechanism. Our measurements suggest that this latter behavior is important from a modeling perspective, as almost all of our TCP traces contained more time-out events than fast retransmit events. Our measurements demonstrate that our model is able to more accurately predict TCP send rate and is accurate over a wider range of loss rates. We also present a simple extension of our model to compute the throughput of a bulk transfer TCP flow, which is defined as the amount of data received by the receiver in unit time. Index Terms—Empirical validation, modeling, retransmission timeouts, TCP.

Abstract—Several analytic models describe the steady-state throughput of bulk transfer TCP flows as a function of round trip time and packet loss rate. These models describe flows based on the assumption that they are long enough to sustain many packet losses. However, most TCP transfers across today’s Internet are short enough to see few, if any, losses and consequently their performance is dominated by startup effects such as connection establishment and slow start. This paper extends the steadystate model proposed in [34] in order to capture these startup effects. The extended model characterizes the expected value and distribution of TCP connection establishment and data transfer latency as a function of transfer size, round trip time, and packet loss rate. Using simulations, controlled measurements of TCP transfers, and live Web measurements we show that, unlike earlier steady-state models for TCP performance, our extended model describes connection establishment and data transfer latency under a range of packet loss conditions, including no loss. I.

There are considerable efforts within the Grid and high performance computing communities to improve end-to-end network performance for applications that require substantial amounts of network bandwidth. The Atlas project [19] for example, must be able to reliably transfer over 2 Petabytes of data per year over transatlantic networks from Europe to the United States. Recent experience [1, 2] has demonstrated that actual aggregate TCP throughput realized by high performance applications is persistently much less than the end-to-end structural and load characteristics of the network would indicate is available. One source of poor TCP throughput that has been identified is a packet loss rate that is much greater than what would be reasonably expected [20]. Packet loss is interpreted by TCP as an indication of network congestion between a sender and receiver. Packet loss, however, may be due to random factors other than network congestion, such a

The statistical characteristics of network traffic- in particular the observation that it can exhibit long range dependence- have received considerable attention from the research community over the past few years. In addition, the recent claims that the TCP protocol can generate traffic with long rage dependent behavior has also received much attention. Contrary to the latter claims, in this paper we show that the TCP protocol can generate traffic with correlation structures that spans only an analytically predictable finite range of time-scales. We identify and analyze separately the two mechanisms within TCP that are responsible for this scaling be-havior: timeouts and congestion avoidance. We provide analytical models for both mechanisms that, under the proper loss probabilities, accurately predict the range in time-scales and the strength of the sustained correlation structure of the traffic sending rate of a single TCP source. We also analyze an existing comprehensive model of TCP that accounts for both mechanisms and show that TCP itself exhibits a predictable finite range of time-scales under which traffic presents sustained correlations. Our claims and results are derived from Markovian models that are supported by simulations. We note that traffic generated by TCP can be misinterpreted to have long range dependence, but that long range dependence is not possible due to inherent finite time-scales of the mechanisms of TCP.

In this paper, we analyze the results of a seven-month real-time streaming experiment, which was conducted between a number of unicast dialup clients, connecting to the Internet through access points in more than 600 major U.S. cities, and a backbone video server. During the experiment, the clients streamed lowbitrate MPEG-4 video sequences from the server over paths with more than 5,000 distinct Internet routers. We describe the methodology of the experiment, the architecture of our NACK-based streaming application, study end-to-end dynamics of 16 thousand ten-minute sessions (85 million packets), and analyze the behavior of the following network parameters: packet loss, round-trip delay, one-way delay jitter, packet reordering, and path asymmetry. We also study the impact of these parameters on the quality of real-time streaming.

In this paper, we analyze the dynamics of a sevenmonth real-time streaming experiment, which was conducted between a number of unicast dialup clients, connecting to the Internet through access points in more than 600 major U.S. cities, and a backbone video server. During the experiment, the clients streamed low-bitrate MPEG-4 video sequences from the server over paths with more than 5,000 distinct Internet routers. We describe the methodology of the experiment and the architecture of our NACK-based streaming application, study end-to-end dynamics of 16 thousand ten-minute sessions (85 million packets), and analyze the behavior of the following network parameters: packet loss, round-trip delay, and packet reordering. We also study the impact of these parameters on the quality of real-time streaming (i.e., the number of underflow events).

. We propose a prediction-based best-effort real-time service to support distributed, interactive applications in shared, unreserved computing environments. These applications have timing requirements, but can continue to function when deadlines are missed. In addition, they expose two kinds of adaptability: tasks can be run on any host, and their resource demands can be adjusted based on user-perceived quality. After defining this class of applications, we describe a significant example, an earthquake visualization tool, and show how it could benefit from the service. Finally, we present evidence that the service is feasible in the form of two studies of algorithms for host load prediction and for predictive task mapping. 1 Introduction There is an interesting class of interactive applications that could benefit from a realtime service, but which must run on conventional reservation-less networks and hosts where traditional forms of real-time service are difficult or impossible to im...

The use of parallel TCP connections to increase throughput for bulk transfers is common practice within the high performance computing community. However, the effectiveness, fairness, and efficiency of data transfers across parallel connections is unclear. This paper considers the impact of systemic non-congestion related packet loss on the effectiveness, fairness, and efficiency of parallel TCP transmissions. The results indicate that parallel connections are effective at increasing aggregate throughput, and increase the overall efficiency of the network bottleneck. In the presence of congestion related losses, parallel flows steal bandwidth from other single stream flows. A simple modification is presented that reduces the fairness problems when congestion is present, but retains effectiveness and efficiency.

... This paper addresses this issue. We begin by statistically characterizing the TCP throughput on the Internet, exploring the strong correlation between TCP flow size and throughput, and the transient end-to-end throughput distribution. We then analyze why benchmarking fails to predict large transfers, and propose a novel yet simple prediction model based on our observations. Our prototype,

Large scale networked image retrieval systems face a number of problems that are not fully satisfied by current systems. On one hand, integrated solutions that store all image data centrally are often limited in terms of scalability and autonomy of data providers. On the other hand, WWW-based search engines proved to be fairly scalable, and data providers retain their autonomy. However, such engines often confront users with links to servers that are not available or to images that no longer exist, i.e., they are unable to keep their meta-database consistent with the repositories' contents. Furthermore, existing solutions often neglect the cost of image delivery. The considerable variations in the effective bandwidth in today's Internet lead to highly unpredictable response times, which are often intolerable from the user's point of view. This paper presents the architecture of Chariot, a networked image search and retrieval system that tackles these concerns. With respect to scalabil...