D. L. Jagerman, B. Melamed, and W. Willinger, "Stochastic modeling of traffic processes," in Frontiers in Queueing: Models, Methods and Problems (J. Dshalalow, ed.), CRC Press, Boca Raton, 1995.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....have also been widely studied. Examples include [22] 23] 24] These studies inform our work with respect to the structural characteristics of end to end Internet paths. A variety of methods have been employed to model network packet traffic including queuing and auto regressive techniques [25]. While these models can be parameterized to recreate observed packet traffic time series, parameters for these models often do not relate to network properties. Models for TCP throughput have also been developed in [26] 27] 28] These models use RTT and packet loss rates to predict ....

D. Jagerman, B. Melamed, and W. Willinger, Stochastic Modeling of Traffic Processes, Frontiers in Queuing: Models, Methods and Problems, CRC Press, 1996.

....tractable models but is still subject of current research interest for analytically tractable models. Non analytically tractable models, e.g. fractional Gaussian noise (fGN) and fractional autoregressive integrated moving average (fARIMA) naturally capture burstiness as well as self similarity [10]. Various research papers have subjected these models, e.g. Ledesma and Liu reported the effective construction of fGN in [12] For analytically tractable models, e.g. the Markov modulated Poisson process (MMPP, 7] recent work has been proposed that utilizes the MMPP in order to mimic ....

D.L. Jagerman, B. Melamed, and W. Willinger, Stochastic Modeling of Traffic Processes, in: Frontiers' in Queuing: Models, Methods' and Problems, CRC Press, 1996. -27-

....of data, and a weeMy period, M 4 so the correction factor is a not insignificant 4 3. We should note that not only is a ideal for our purposes, but even when the model (1) breaks down, for instance when there are outliers in the data, then parameter a is a useful and meaningful measurement (see [16,17] and Section 3.4) Furthermore, the metric may be easily adapted to deal with missing data, a feature we use below. 3.3. Simulation results In the previous section we present an estimator called the empirical peakedness, but this estimator is not an unbiased estimator of peakedness, because the ....

D. L. Jagerman, B. Melamed, and W. Willinger, "Stochastic modeling of traffic processes," in Frontiers in Queueing: Models, Methods and Problems (J. Dshalalow, ed.), CRC Press, Boca Raton, 1995.

....H 2 (0:5; 1) where c is a constant [3] The degree of this slow decay is determined by the Hurst parameter (H) Traffic models (fractional Brownian motion (fBm) models, FARIMA models, Cox s M G 1 models, on off models, etc. to capture LRD and self similar properties have also been developed [37, 18, 27, 26]. Among these models the fBm [28] was found to be a popular parsimonious and tractable model of traffic aggregation. It was shown that the fBm is an accurate model if the traffic is aggregated from a large number of independent users whose peak rates are small relative to link capacity and the ....

D. L. Jagerman, B. Melamed, and W. Willinger. Stochastic Modeling of Traffic Processes. In Frontiers in Queueing, pages 271--370. CRC Press, 1997.

....on how to detect accurately the LRD property and how to estimate the Hurst parameter [9] 10] A large group of traffic models (fractional Brownian motion (fBm) models, FARIMA models, Cox s M G # models, on off models, etc. to capture LRD and self similar properties have also been developed [1] [11], 12] 8] Among these models the fBm [13] was found to be a popular parsimonious and tractable model of traffic aggregation. It was shown that the fBm is an accurate model if the traffic is aggregated from a large number of independent users whose peak rates are small relative to link capacity ....

D. L. Jagerman, B. Melamed, and W. Willinger, "Stochastic modeling of traffic processes," in Frontiers in Queueing: Models, Methods and Problems, J. Dshalalow, Ed., pp. 271--370. CRC Press, 1997.

....to model such user feedback, the proposed methodology is able to support studies of user satisfaction as we will demonstrate in a case study. 2 Related Work In the past, network traffic has mainly been modelled by capturing the char acteristics of the stochastic process of the packet arrival [8]. Here, first and second order characterstics as well as long range phenomenons like self similarity (e.g. in Ethernet, VBR encoded video, and web traffic) have been captured. In contrast to those resource oriented modeling approaches, user behavior models try to catch the sequence of user ....

D. Jagerman, B. Melamed, W. Willinger, Stochastic modeling of traffic processes. In Frontiers in Queuing: Models, Methods and Problems (J. Dshalalow, Edr.), CRC Press, 1996.

....3i) R [i] 2m 1 i=m i) R [i] R [0] where m =2 . It is clear that the variances of the wavelet coefficient have included the autocorrelation of X. A simple example to verify Equ. 11) is as following. If X is from discrete Fractional Gaussian Noise (DFGN) it is known that [26] X m 2H 2 , 12) X 2 j(2H 1) 2 2H , 13) where H (0.5 H 1)istheHurst parameter. Through some algebraic manipulation, it is not difficult to verify that Equ. 11) holds for the case of DFGN. To show how accurate Equ. 11) is when used to estimate variances of the ....

David L. Jagerman, Benjamin Melamed, and Walter Willinger, "Stochastic modeling of traffic processes," in Frontiers in queueing, chapter 10, pp. 271--320. 1997.

....only stochastic processes for creating packets but also to model the behavior of users running applications which send traffic over a network. II. Related Work In the past, network traffic has mainly been modelled by capturing the characteristics of the stochastic process of the packet arrival [8,5]. Here, first and second order characterstics as well as long range phenomenons like self similarity (e.g. in Ethernet, VBR encoded video, and web traffic) have been captured. In contrast to those resource oriented modeling approaches, user behavior models try to catch the sequence of user ....

D. Jagerman, B. Melamed, W. Willinger, Stochastic modeling of traffic processes. In Frontiers in Queuing: Models, Methods and Problems (J. Dshalalow, Edr.), CRC Press, 1996.

....real time traffic measurements from working packet switched networks that packet traffic fluctuates over a number of time scales [26] This behaviour is called burstiness. However, the unique definition and characterization of burstiness have not been established yet in the teletraffic literature [11, 18]. A very promising approach to capture this burstiness phenomenon in a parsimonious manner is to use fractal traffic models [16, 26] These models have dynamics governed by power law distribution functions and hyperbolically decaying autocorrelation [26] The important characteristics of these ....

.... phenomenon in a parsimonious manner is to use fractal traffic models [16, 26] These models have dynamics governed by power law distribution functions and hyperbolically decaying autocorrelation [26] The important characteristics of these models are self similarity and long range dependence [11, 16]. Self similar stochastic processes have been defined in a number of ways in the literature [11, 16, 26] From a practical point of view the long range dependent (LRD) processes constitute one of the most important classes of these processes [11, 16] In this paper we consider this class of ....

[Article contains additional citation context not shown here]

JAGERMAN, D. L. -- MELAMED, B. -- WILLINGER, W.: Stochastic Modeling of Traffic Processes. In Frontiers in Queueing, pp. 271--370. CRC Press, 1997.

.... traffic in a parsimonious manner is to use fractal traffic models [13] 19] The dynamics of these models are governed by power law distribution functions and hyperbolically decaying autocorrelation [19] The important characteristics of these models are self similarity and long range dependence [8], 13] Self similar stochastic processes have been defined in a number of ways in the literature [8] 13] 19] In practice the most important class of these processes is that of long range dependent (LRD) processes [8] 13] LRD has been detected as a widespread property of packet network ....

.... models are governed by power law distribution functions and hyperbolically decaying autocorrelation [19] The important characteristics of these models are self similarity and long range dependence [8] 13] Self similar stochastic processes have been defined in a number of ways in the literature [8], 13] 19] In practice the most important class of these processes is that of long range dependent (LRD) processes [8] 13] LRD has been detected as a widespread property of packet network traffic, e.g. Internet traffic [12] 18] In this paper we consider this class of self similar ....

[Article contains additional citation context not shown here]

D.L. Jagerman, B. Melamed, and W. Willinger, "Stochastic Modeling of Traffic Processes," in Frontiers in Queueing, pp. 271--370. CRC Press, 1997.

.... in data traffic is that these clustering activities are present over several time scales [19] This phenomenon triggered a new modeling approach, called fractal modeling, which can offer parsimonious models (e.g. self similar or long range dependent (LRD) traffic models) to capture this behavior [19, 10, 12, 11]. A number of studies have reported that aggregated LAN traffic is consistent with exact self similarity and aggregated WAN traffic is asymptotically self similar (long range dependent) 13, 7] Moreover, it has been found that the scaling structure in measured WAN traffic can be categorized into ....

D.L. Jagerman, B. Melamed, and W. Willinger, "Stochastic Modeling of Traffic Processes", In Frontiers in Queueing, pages 271--370. CRC Press, 1997.

....tractable models but is still subject of current research interest for analytically tractable models. Nonanalytically tractable models, e.g. fractional Gaussian noise (fGN) and fractional autoregressive integrated moving average (fARIMA) naturally capture burstiness as well as self similarity [7]. Various research papers have subjected these models, e.g. Ledesma and Liu reported the effective construction of fGN in [9] For analytically tractable models, e.g. the Markovmodulated Poisson process (MMPP, 4] recent work has been proposed that utilizes the MMPP in order to mimic ....

D.L. Jagerman, B. Melamed, and W. Willinger, Stochastic modeling of traffic processes, in: Frontiers in Queuing: Models, Methods and Problems, CRC Press, 1996.

....systems is presented. It is meant for the simulation of network traffic, stand alone computer systems and any mixture thereof. RELATED WORK Models for generating realistic computer and network workload have been under study for many years. Many models are represented by stochastic processes ([8], 6] only, yielding the time points where workload arrives at low level computer components such as the Ethernet physical layer or a WWW server, as well as the size of the arriving workload, e.g. the amount of traffic to be transferred over the network. User behavior models try to catch the ....

D. Jagerman, B. Melamed, W. Willinger, Stochastic modeling of traffic processes. In Frontiers in Queuing:Models, Methods and Problems (J. Dshalalow, Edr.), CRC Press, 1996.

....for the simulation of network traffic, stand alone computer systems and any mixture of these types. 2 Related Work Models for generating realistic computer and network workload have been under study for many years. Many models are represented by stochastic processes (Jagerman, Melamed, Willinger [10], Hlavacs, Kotsis, Steinkellner [9] only, yielding the time points, where workload arrives at low level computer components such as the Ethernet physical layer or a WWW server, as well as the size of the arriving workload, e.g. the amount of traffic to be transfered over the network. User ....

....an appropriate QoS action is triggered and performed. QoS actions and user actions are at the same level. They can start or stop services, or change parameters for running services. 3. 8 Service Layer Service layer models consist of parameterized traffic generators (Jagerman, Melamed, Willinger [10], Hlavacs, Kotsis, Steinkellner [9] Some parameters can be changed by higher level models and influence the QoS level that the higher level model chooses. Other parameters are fixed and describe a situation that can not be changed by users, like the distribution of file sizes at a web site. ....

D. Jagerman, B. Melamed, and W. Willinger. Stochastic modeling of traffic processes. In J. Dshalalow, editor, Frontiers in Queuing:Models, Methods and Problems. CRC Press, 1996.

....modeling and analysis. Historically, traffic modeling has its origins in conventional telephony, and has been based almost exclusively on Poisson (or, more generally, Markovian) assumptions about traffic arrival patterns and on exponential assumptions about resource holding requirements (e.g. [144,209]) However, the emergence of modern high speed packet networks combines drastically new and different transmission and switching technologies with dramatically heterogeneous mixtures of services and applications. As a result, packet traffic is generally expected to be more complex or bursty than ....

D. L. Jagerman, B. Melamed, and W. Willinger. Stochastic modeling of traffic processes. In J. Dshalalow, editor, Frontiers in Queueing: Models, Methods and Problems. CRC Press, 1996. To appear.

No context found.

D. L. Jagerman, B. Melamed, and W. Willinger, "Stochastic modeling of traffic processes," in Frontiers in Queueing: Models, Methods and Problems (J. Dshalalow, ed.), CRC Press, Boca Raton, 1995.

No context found.

D. L. Jagerman, B. Melamed, and W. Willinger, "Stochastic modeling of traffic processes," in Frontiers in Queueing: Models, Methods and Problems (J. Dshalalow, ed.), CRC Press, Boca Raton, 1995.

No context found.

D. L. Jagerman, B. Melamed, and W. Willinger. Stochastic modeling of traffic processes. In J. Dshalalow, editor, Frontiers in Queueing: Models, Methods and Problems, pages 271--370. CRC Press, 1997.

No context found.

D. Jagerman, B. Melamed, and W. Willinger, Stochastic Modeling of Traffic Processes, Frontiers in Queuing: Models, Methods and Problems, CRC Press, 1996.

No context found.

D.L. Jagerman, B. Melamed, and W. Willinger, "Stochastic modeling of traffic processes," invited chapter in Frontiers in Queueing: Models, Methods and Problems J.H. Dshalalow (Eds.), CRC Press, 1996.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC