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.

In this paper, we present a model for TCP/IP congestion control mechanism. The rate at which data is transmitted increases linearly in time until a packet loss is detected. At this point, the transmission rate is divided by a constant factor. Losses are generated by some exogenous random process which is assumed to be stationary ergodic. This allows us to account for any correlation and any distribution of inter-loss times. We obtain an explicit expression for the throughput of a TCP connection and bounds on the throughput when there is a limit on the window size. In addition, we study the effect of the Timeout mechanism on the throughput. A set of experiments is conducted over the real Internet and a comparison is provided with other models that make simple assumptions on the inter-loss time process. The comparison shows that our model approximates well the throughput of TCP for many distributions of inter-loss times.

Today's universal communications increasingly involve mobile and battery-powered devices (e.g. hand-held, laptop) over wired and wireless networks. Energy efficiency, as well as throughput, are becoming service characteristics of dominant importance in communication protocols. Although standard TCP versions lack the functionality to efficiently adjust their error-control strategies to distinct characteristics of network environments and to specific constraints of communicating devices, the wide range of TCP-based applications have rendered TCP the de facto standard for reliable end-to-end communications. In this work we propose "grafting" two components of strategic significance onto standard TCP: a Probing mechanism and an Immediate Recovery strategy. Our results show that these enhancements yield higher throughput while maintaining lower levels of energy expenditure, and thus have the potential of promoting TCP's congestion control to a universal error-control schema for heterogeneou...

We discuss the design principles of TCP within the context of heterogeneous wired/wireless networks and mobile networking. We identify three shortcomings in TCP's behavior: (i) the protocol's error detection mechanism, which does not distinguish different types of errors and thus does not suffice for heterogeneous wired/wireless environments, (ii) the error recovery, which is not responsive to the distinctive characteristics of wireless networks such as transient or burst errors due to handoffs and fading channels, and (iii) the protocol strategy, which does not control the tradeoff between performance measures such as goodput and energy consumption, and often entails a wasteful effort of retransmission and energy expenditure. We discuss a solution-framework based on selected research proposals and the associated evaluation criteria for the suggested modifications. We highlight an important angle that did not attract the required attention so far: the need for new performance metrics, appropriate for evaluating the impact of protocol strategies on battery-powered devices.

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 three quantities. The first is the send rate, which is the amount of data sent by the sender in unit time. The second is the throughput, which is the amount of data received by the receiver in unit time. Note that the throughput will always be less than or equal to the send rate due to losses. Finally, the number of non-duplicate packets received by the receiver in unit time gives us the goodput of the connection. The goodput is always less than or equal to the throughput, since the receiver may receive two copies of the same packet due to retransmissions by the sender. In [9], we presented a simple model for predicting the steady state send rate of a bulk transfer TCP ow as a function of loss rate and round trip time. In this paper, we extend that work in two ways. First, we analyze the performance of bulk transfer TCP flows using more ...

Emerging technologies such as Bluetooth are expected to become a ubiquitous solution for providing short range, low power, low cost, pico-cellular wireless connectivity. Bluetooth is a Master driven Time Division Duplex (TDD) system that supports an asynchronous channel for data traffic as well as synchronous channels for voice traffic. Data applications running over Bluetooth such as http, ftp and real audio will need transport layer protocols such as TCP and UDP to send packets over the wireless links. In this paper we study several schemes designed to improve the performance of asynchronous data traffic over a Bluetooth piconet that supports multiple active slaves. We propose and compare a number of SAR policies and MAC scheduling algorithms with a view towards enhancing the performance of transport layer sessions. We investigate the effect of different FEC and ARQ schemes at the baseband level, using a two-state Markov channel model for the Bluetooth RF link. We also study how the presence of circuit-switched voice impacts the performance of data traffic. Keywords--- Medium Access Control (MAC), Scheduling, Time Division Duplex (TDD), Segmentation and Reassembly (SAR), Forward Error Correction (FEC), Automatic Repeat Request (ARQ), TCP, UDP. I.

In this paper, we outline a methodology that can be applied to model the behavior of TCP flows. The proposed methodology stems from a Markovian model of a single TCP source, and eventually considers the superposition and interaction of several such sources using standard queueing analysis techniques. Our approach allows the evaluation of such performance indices as throughput, queueing delay and packet loss of TCP flows. The results obtained through our model are validated by means of simulation, under several topology and traffic settings. I. INTRODUCTION According to recent estimates 95% of the traffic carried today over wide-area IP networks uses TCP as transport protocol, which amounts to 80% of the overall end-to-end flow count. These figures alone highlight the key role that TCP plays in delivering a reliable service to the most common network applications such as email programs and Web browsers. During the many years of "honorable" service, TCP has undergone several alteration...

In this paper, we investigate the performance of TCP Tahoe and NewReno on wireless links with a FEC/ARQ protocol at the link layer. The effect of link layer (LL) parameters like the LL packet size, number of LL transmission attempts, and FEC code rate on the TCP throughput is evaluated. It is shown that the retransmissions at the link layer must persist long enough to outlast the average bad state duration of the Rayleigh fading process in order to achieve performance better than TCP without an ARQ at the link layer. It is further shown that with a suitable link layer in place, TCP Tahoe offers as high a throughput as offered by other enhanced versions like NewReno in the considered wireless environment. I. INTRODUCTION Wireless data applications like e-mail/web browsing, mobile computing, etc., are gaining increased attention due to rapid advances in the areas of wireless communications and the Internet. Transmission Control Protocol (TCP) is an end-to-end transport protocol in the ...

An analytical framework for modeling the performance of a single TCP session in the presence of random packet loss is presented. A Markovian approach is developed that allows us to study both memoryless channels (i.i.d packet loss) and channels with memory (correlated packet loss) modeled by a two state continuous time Gilbert model. The analytical results are validated against results using the ns simulator. It is shown that the model predicts throughput for LANs/WANs (low and high bandwidthdelay products) with good accuracy. Further, throughput for the i.i.d loss model is found to be relatively insensitive to the probability density function (p.d.f) of the loss inter-arrival process. For channels with memory, we present an empirically validated rule of thumb to categorize the channel transition frequency. Keywords---Transport control protocol, wireless networks, performance analysis. I. INTRODUCTION M OST of today's Internet traffic is carried by networks using TCP. While initial...

FEC is widely used to improve the quality of noisy transmission media as wireless links. This improvement is of importance for a transport protocol as TCP which uses the loss of packets as an indication of network congestion. FEC shields TCP from losses not caused by congestion but it consumes some bandwidth that could be used by TCP. We study in this paper the tradeoff between the bandwidth consumed by FEC and that gained by TCP.