Dynamics of TCP traffic over ATM networks

Investigates the performance of transport control protocol (TCP) connections over ATM networks without ATM-level congestion control and compares it to the performance of TCP over packet-based networks. For simulations of congested networks, the effective throughput of TCP over ATM can be quite low when cells are dropped at the congested ATM switch. The low throughput is due to wasted bandwidth as the congested link transmits cells from “corrupted” packets, i.e., packets in which at least one cell is dropped by the switch. The authors investigate two packet-discard strategies that alleviate the effects of fragmentation. Partial packet discard, in which remaining cells are discarded after one cell has been dropped from a packet, somewhat improves throughput. They introduce early packet discard, a strategy in which the switch drops whole packets prior to buffer overflow. This mechanism prevents fragmentation and restores throughput to maximal levels     

Evaluation and comparison of cell loss and delay models for ATM multiplexers

Among several models that are available to represent the aggregate cell flow generated by on–off sources at an ATM multiplexer (either at an access or a switching node), the Interrupted Bernoulli Process (IBP) is characterized by particular simplicity and analytical tractability. The superposition of sources individually modeled as an IBP, whose cells enter a common buffer, is considered in this paper. The main goal is to compute approximations of two basic Quality of Service (QoS) indicators, namely, cell loss rate, whose analytical computation has been already presented in previous works, and the rate of cells exceeding a specified delay, whose presentation constitutes the theoretical novelty of the paper. Analytical expressions of these two quantities are given for homogeneous sources, i.e., possessing the same statistical parameters and QoS requirements. The analytical formulation is carefully evaluated by comparing the results obtained with others presented in the literature and with simulation results; in the latter, the actual cell arrival process is generated by means of a Markov Modulated Deterministic Process (MMDP) model of the on–off sources. Several comparisons are performed for different offered loads and by varying the buffer length, which show the effectiveness and the limits of the technique under investigation.     

Max-utility wireless resource management for best-effort traffic

Due to the characteristics of wireless channels, utility-based resource management in wireless networks requires a set of mechanisms that are different from those for wireline networks. This paper explores in detail why and how the requirements are different. In particular, we analyze the wireless network performance to find out the scheduling algorithm that maximizes total utility of the system. Unlike previous studies, this paper focuses on scenarios in which wireless networks are not fully loaded and all of the users are best-effort data users, i.e., there is no streaming user. Our first key conclusion is that Kleinrock's Conservation Law provides a valuable means to accurately capture the perceived rates of best-effort users in such systems. The queueing analysis further indicates that, within periods during which channel conditions are stable for each user, albeit differ from user to user, the max-utility scheduling algorithm can be derived using queueing theorem and can be readily implemented in actual systems for utility functions that are of exponential or log format. When further taking into account the time-variant nature of wireless channel conditions, our simulation results demonstrate that dynamic weighted fair queueing, with weights adjusted according to the channel conditions, can achieve highly desirable performance with great flexibility.     

An efficient method for computing cell loss probability for heterogeneous bursty traffic in ATM networks

In this paper, we present an approximation for the probability of cell loss of heterogeneous bursty traffic in broadband integrated packet networks based on the asynchronous transfer mode. The sources considered here alternate between active and silent periods and are characterized by their peak and average transmission rates. The cell loss probability is obtained by considering only the number of active sources at a given time and computing the amount of traffic that exceeds the link capacity. Since the amount of buffered excess traffic is not considered in this computation, this approximation is actually an upper bound of the cell loss probability. The numerical efficiency of this bound enables it to be used as a measure based on which robust and simple resource allocation strategies can be developed for bursty sources. Comparison of this bound with the simulation results has shown that the bound is close to the actual loss probabilities especially for large burst lengths and high utilizations.     

ABR performance in presence of bursty TCP traffic

The available bit rate (ABR) service class is a solution for the integration of data traffic in asynchronous transfer mode networks. Many algorithms have been proposed to implement ABR services. The authors present simulation results showing poor performance by a common ABR algorithm when supporting TCP bursty traffic. As a solution to this problem, the authors propose time averaging of the parameters calculated by the ABR algorithms     

On self-similar traffic in ATM queues: definitions, overflowprobability bound, and cell delay distribution

Recent traffic measurements in corporate local-area networks (LANs), variable-bit-rate video sources, ISDN control-channels, and other communication systems, have indicated traffic behaviour of self-similar nature. This paper first discusses some definitions and properties of (second-order) self-similarity and gives simpler criteria for it. It then gives a model of self-similar traffic suitable for queuing system analysis of an asynchronous transfer mode (ATM) queue. A lower bound to the overflow probability of a finite ATM buffer is obtained, as also a lower bound to the cell loss probability. Finally, the stationary distribution of the cell delay in an infinite ATM buffer is obtained     

Weighted proportional loss rate differentiation of TCP traffic

Relative service differentiation refers to service models that provide assurances for the relative quality ordering between classes, rather than for the absolute service level in each class. An example is the proportional differentiation model which provides a way to control the quality spacing between classes locally at each hop. In this model, certain forwarding metrics are ratioed proportional to the class differentiation parameters that the network operator chooses. In this paper, we propose a new proportional loss rate differentiation mechanism that integrates relative loss rate differentiation directly into active queue management using random early detection. Copyright © 2004 John Wiley &Sons, Ltd.     

Congestion avoidance for supporting a best-effort service in an ATM LAN

In this paper, after having presented some issues related to atm within a lan, we introduce the besteffort service. We then discuss different concepts of flow control to support such a service. After having proposed a definition of fairness, we focus on one flow control mechanism, namely a hop- by hop backpressure scheme. We describe the mechanism and show that it can protect the network against excessive traffic while insuring fairness among the different sources. We also explain how such a mechanism could be implemented at the uni by using current technology and by relying on the standards currently being drafted in the itu and the atm Forum.     

Flow control schemes and delay/loss tradeoff in ATM networks

Performance and flow control mechanisms, which represent ATM (asynchronous transfer mode) flexibility, are discussed. To control performance, delay- and/or loss-sensitive service classes, and two mechanisms to realize these classes, are proposed. It is shown that it is possible to have better performance than with other mechanisms, such as no-class distinction or simple priority methods. It is further suggested that this performance controllability results in the provision of multiple logical services, including quasi-STM (synchronous transfer mode; compatible with circuit switching), by an ATM network. ATM flow control is based on a call-oriented resource allocation mechanism similar to circuit switching. The concepts of call/line bit rate ratio and multiplexing degree are seen to be significant for efficient use of resources. When the network handles calls with large call/line bit rate ratios, user-specified flow control parameters at the call setup phase are important for resource assignment. The definition of two types of maximum throughput of each call and its usage for resource management are proposed     

An ATM queue manager handling multiple delay and loss priorities

The asynchronous transfer mode (ATM) technique has been widely accepted as a flexible and effective scheme to transport various traffic over the future broadband network. To fully utilize network resources while still providing satisfactory quality of service (QOS) to all network users, prioritizing the user's traffic according to their service requirements becomes necessary. During call setup or service provisioning, each service can be assigned a service class determined by a delay priority and a loss priority. A queue manager in ATM network nodes will schedule ATM cells departing and discarding sequence based on their delay and loss priorities. Most queue management schemes proposed so far only consider either one of these two priority types. The queue manager handles multiple delay and loss priorities simultaneously. Moreover, a cell discarding strategy, called push-out, that allows the buffer to be completely shared by all service classes, has been adopted in the queue manager. We propose a practical architecture to implement the queue manager by using available VLSI sequencer chips     

TCP throughput and fairness performance in presence of delay spikes in wireless networks

Wireless networks are likely to experience delay spikes exceeding several times the typical round‐trip‐time figures, which can cause spurious timeouts that lead to unnecessary retransmissions and reduction of the TCP sender's transmission rate, and thus, the throughput of the TCP is degraded. This paper presents some research results on the effect of delay spikes caused by handover on TCP performance by using three different mobility models. It is shown that the throughput of TCP connection over a single bottleneck link is decreased in the presence of delay spikes significantly. Furthermore, it is shown that the fairness feature of TCP is also severely affected in the presence of delay spikes. Copyright © 2005 John Wiley & Sons, Ltd.     

A new burst assembly scheme based on the average packet delay and its performance for TCP traffic

We propose and evaluate a new burst assembly algorithm based on the average delay of the packets comprising a burst. This method fixes the average delay of the packets belonging to an assembled burst to a desired value $T_{\mathrm{AVE}}$ that may be different for each forwarding equivalence class (FEC). We show that the proposed method significantly improves the delay jitter experienced by the packets during the burst assembly process, when compared to that of timer-based and burst length-based assembly policies. Minimizing packet delay jitter is important in a number of applications, such as real-audio and streaming-video applications. We also find that the improvement in the packet delay jitter yields a corresponding significant improvement in the performance of TCP, whose operation depends critically on the ability to obtain accurate estimates of the round-trip times (RTT).     

A Markovian model for TCP over ATM

This paper focuses on the development of an approximate Markov chain model and a detailed simulation model for the performance analysis of TCP connections supporting long file transfers within highspeed ATM networks. The Markov model construction is first illustrated, emphasizing the simplifying assumptions introduced to maintain an acceptable complexity in the performance study. Then the model validation is described, comparing the analytical results with those produced by a very detailed simulation setup in which the officially distributed BSD 4.3 TCPreno code was adapted to run on top of a software tool for the simulation of ATM networks. Numerical results indicate that the analytical model is able to capture the overall behavior of the system with respect to the performance measures we consider, with a cost which is orders of magnitude less than the one necessary to obtain reliable estimates from the detailed simulation model.     

Measurement-based simulation model for TCP over ATM

Cell level analysis of ATM networks by means of simulation requires an accurate model for traffic sources. We present a simple model for TCP over ATM traffic sources in an ATM LAN, which captures the fundamental characteristics of the behaviour of tcp in this environment. The model was developed by extensive statistical analysis of numerous traffic traces recorded in an atm testbed. Simulated traffic generated by our model has roughly the same properties as observed in real traffic, even on the time scale of milliseconds. This makes it suitable for use by the analysis of various scheduling and congestion management algorithms.     

Encrypting ATM traffic over the ACTS ATM internetwork

The US government has long been passing encrypted traffic over commercial/public communications systems. With the proliferation of ATM networks, there is a vital need for government agencies to utilize the high bandwidth these new ATM networks have to offer. This article presents a review of a prototype ATM encryption device that was employed on the ACTS ATM internetwork during a government-sponsored demonstration. The article also describes this device, the FASTLANE ATM encryptor and its fielded prototype, the Interim FASTLANE     

Per-source mean cell delay and mean buffer contents in ATM queues

Most performance studies of ATM cell buffers so far have been limited to global performance characteristics such as the total mean buffer contents and the global mean cell delay. The authors present an analytical technique which allows the derivation of performance measures for single sources     

Accurate approximation of cell loss probability for self-similartraffic in ATM networks

Recently, self-similar (or fractal) stochastic processes have been regarded as more accurate models of certain categories of traffic (e.g. LAN traffic) which will be transported in ATM networks. The authors propose a method for estimating the cell loss probability in an ATM multiplexer fed by a self-similar arrival process. The packet arrivals are generated by a fractional Brownian motion process and the service process is deterministic. The approach is based on theory for large deviations, and simulation tests show that it is more accurate than the existing approximation result     

A traffic measurement method and its application for cell lossprobability estimation in ATM networks

Focuses on how to determine a set of parameters describing cell arrival stream characteristics and how to evaluate the accuracy of these measurements. Since a cell arrival stream is both correlated and bursty, the autocorrelation characteristics between the number of arriving cells in two adjacent intervals are analyzed, and then the time interval is determined for which the autocorrelation may be neglected. It is also shown that this set of parameters can represent the first and approximately second moment statistics during any interval. A method of estimating cell loss probability through the use of a queuing model that has an input process is presented. The input process is determined from the set of parameters that describe the characteristics of the source. In order to verify the accuracy, a brief simulation study is carried out     

Bandwidth-dependent scheduling algorithm to limit cell delayvariation for CBR traffic in ATM networks

A multiplexing control scheme is proposed for limiting the cell delay variation (CDV) caused by multiplexing of constant bit rate (CBR) traffic in asynchronous transfer mode (ATM) networks. The control scheme is based on a bandwidth-dependent scheduling algorithm. The limitation of the CDV is proved     

Reducing cell loss in banyan based ATM switching fabrics

In this paper, we propose a new technique for reducing cell loss in multi‐banyan‐based ATM switching fabrics. We propose a switch architecture that uses incremental path reservation based on previously established connections. Path reservation is carried out sequentially within each banyan but multiple banyan planes can be concurrently reserved. We use a conflict resolution approach according to which banyans make concurrent reservation offers of conflict‐free paths to head of the line cells waiting in input buffers. A reservation offer from a given banyan is allocated to the cell whose source‐to‐destination path uses the largest number of partially allocated switching elements which are shared with previously reserved paths. Paths are incrementally clustered within each banyan. This approach leaves the largest number of free switching elements for subsequent reservations which has the effect of reducing the potential of future conflicts and improves throughput. We present a pipelined switch architecture based on the above concept of path‐clustering which we call path‐clustering banyan switching fabric (PCBSF). An efficient hardware that implements PCBSF is presented together with its theoretical basis. The performance and robustness of PCBSF are evaluated under simulated uniform traffic and ATM traffic. We also compare the cell loss rate of PCBSF to that of other pipelined banyan switches by varying the switch size, input buffer size, and traffic pattern. Copyright © 2001 John Wiley & Sons, Ltd.