Self-Similar Network Traffic and Performance Evaluation / Edition 1 available in Hardcover
Self-Similar Network Traffic and Performance Evaluation / Edition 1
- ISBN-10:
- 0471319740
- ISBN-13:
- 9780471319740
- Pub. Date:
- 09/04/2000
- Publisher:
- Wiley
Self-Similar Network Traffic and Performance Evaluation / Edition 1
Hardcover
Buy New
$209.95Overview
Product Details
| ISBN-13: | 9780471319740 |
|---|---|
| Publisher: | Wiley |
| Publication date: | 09/04/2000 |
| Pages: | 576 |
| Product dimensions: | 6.38(w) x 9.43(h) x 1.24(d) |
About the Author
Walter Willinger is the editor of Self-Similar Network Traffic and Performance Evaluation, published by Wiley.
Table of Contents
Self-Similar Network Traffic: An Overview (K. Park & W. Willinger).Wavelets for the Analysis, Estimation, and Synthesis of Scaling Data (P. Abry, et al.).
Simulations with Heavy-Tailed Workloads (M. Crovella & L. Lipsky).
Queueing Behavior Under Fractional Brownian Traffic (I. Norros).
Heavy Load Queueing Analysis with LRD On/Off Sources (F. Brichet, et al.).
The Single Server Queue: Heavy Tails and Heavy Traffic (O. Boxma & J. Cohen).
Fluid Queues, On/Off Processes, and Teletraffic Modeling with Highly Variable and Correlated Inputs (S. Resnick & G. Samorodnitsky).
Bounds on the Buffer Occupancy Probability with Self-Similar Input Traffic (N. Likhanov).
Buffer Asymptotics for M/G/
Input Processes (A. Makowski & M. Parulekar).
Asymptotic Analysis of Queues with Subexponential Arrival Processes (P. Jelenkovi).
Traffic and Queueing from an Unbounded Set of Independent Memoryless On/Off Sources (P. Jacquet).
Long-Range Dependence and Queueing Effects for VBR Video (D. Heyman & T. Lakshman).
Analysis of Transient Loss Performance Impact of Long-Range Dependence in Network Traffic (G.-L. Li & V. Li).
The Protocol Stack and Its Modulating Effect on Self-Similar Traffic (K. Park, et al.).
Characteristics of TCP Connection Arrivals (A. Feldmann).
Engineering for Quality of Service (J. Roberts).
Network Design and Control Using On/Off and Multilevel Source Traffic Models with Heavy-Tailed Distributions (N. Duffield & W. Whitt).
Congestion Control for Self-Similar Network Traffic (T. Tuan & K. Park).
Quality of Service Provisioning for Long-Range-Dependent Real-Time Traffic (A. Adas & A. Mukherjhee).
Toward an Improved Understanding of Network Traffic Dynamics (R. Riedi & W. Willinger).
Future Directions and Open Problems in Performance Evaluation and Control of Self-Similar Network Traffic (K. Park).
Index.
What People are Saying About This
"The primary objective of the book is to present a comprehensive yet cohesive account of some of the principal developments..." (IEE Signal Processing, Vol. 18, No. 1, January 2001)
Preface
PREFACE
The recent discovery of scaling phenomena in modern communication networks involving self-similarity or fractals and power-law or heavy-tailed distributions is yet another realization of Benoit Mandelbrot's vision of order in physical, social, and engineered systems characterized by scaling laws. Since the seminal paper by Leland, Taqqu, Willinger and Wilson in 1993 which set the groundwork for considering self-similarity an ubiquitous feature of empirically observed network traffic and an important notion in the understanding of the traffic's dynamic nature for modeling analysis and control of network performance, an explosion of work has ensued investigating the multifaceted nature of this phenomenon.
Despite the fact that data networks such as the Internet are drastically different from legacy public switched telephone networks, the long held paradigm in the communication and networking research community has been that data traffic--analogous to voice traffic--is adequately described by certain Markovian models which are amenable to accurate analysis and efficient control. This supposition has been instrumental in shaping the optimism permeating the late 1980s and early 1990s regarding the ability of achieving efficient traffic control for quality of service provisioning in modern high-speed communication networks. The discovery and, more importantly, succinct formulation and recognition that actual data traffic may, in fact, be fundamentally different in nature from the hereto accustomed telephony traffic has significantly influenced the networking research landscape, necessitating a reexamination and revamping of some of its basic premises.
This book is a collection of chapter contributions which brings together relevant past works spanning a cross-section of topics covering traffic measurement, modeling, performance analysis, and traffic control for self similar network traffic. The primary objective of the book is to present a comprehensive yet cohesive account of some of the principal developments and results concerning self-similar network traffic across its various facets, with the aim of serving as a reflective milestone that captures the state-of-the-art in the field. The book is organized around three main subtopics--traffic modeling, queueing-based performance analysis, and traffic control. By and large, the chapters reflect how research in these areas has reacted when faced with the new scientific discoveries involving self similarity and ubiquitous presence of heavy-tailed phenomena in networked systems.
The spectrum of reactions ranges from evolutionary--holding on to traditional frameworks and tested concepts, and trying to extend, generalize them in the presence of unfamiliar characteristics that, in many ways, contradict conventional wisdom--all the way to revolutionary, which embrace the novel and, at times, surprising features giving rise to new questions, research problems, and challenges both on theoretical and practical fronts of relevance to the future Internet. Overall, the reader may find the majority of book chapters to be of an evolutionary rather than revolutionary nature: Many of the problems that have been considered in the past and have been assumed to fit into the powerful, but also mathematically convenient, framework of Markovian analysis are being reformulated and analyzed to incorporate the slowly improving understanding of data traffic. More fundamental issues such as whether or not these problems are still relevant in light of the stark contrast between hereto assumed properties of network traffic and observed reality has attracted less attention to date. In this sense, the book chapters give a sense of how science, in many instances, works when faced with new discoveries and realities, and they also illustrate how a ``give and take'' between traditional approaches, on the one side, and unconventional thinking on the other side can lead to progress, thus advancing our overall understanding in the various subtopics covered in this book. It will be interesting to observe if, and when, future developments in these areas will require more concentrated focus on revolutionary ideas and approaches to networking research and practice, especially as far as network performance analysis and traffic control are concerned.
The chapter contributions have been organized into three parts: (i) estimation and simulation, (ii) queueing with self-similar input, and (iii) traffic control and resource provisioning. The threefold categorization is not strict in the sense that some chapters encompass subject matters that cross the set boundaries. Chapter 1, in addition to serving as an introductory chapter which provides the necessary background and technical know-how to understanding self-similar traffic that is common to many of the chapters, also gives a bird's eye view of each chapter, how they fit into the overall picture, and comments on the role and potential relevance for future advances. The remaining two chapters in Part I deal with traffic characterization, estimation, and modeling issues. Wavelet analysis is introduced as a powerful technique for both modeling and estimation in self-similar traffic. Augmenting the theme of traffic modeling are issues surrounding simulations such as those arising in the generation of self-similar traffic and workloads which entails, in many instances, sampling from heavy-tailed distributions requiring special considerations.
The second part of the book consists of ten chapters and focuses on traditional performance evaluation issues, in particular, queueing behavior of finite and infinite buffer systems when fed with long-range dependent input. Due to the breakdown of Markovian assumptions which are key to achieving tractable analysis in traditional queueing analysis, the technical challenges encountered with self-similar input are great, and this part of the book exposes what is known about queueing with self-similar input, above and beyond the phenomenon that queue length distribution decays polynomially and not exponentially. The traffic models employed, to a large extent, can be viewed as variants of on/off renewal reward processes where session arrivals are allowed to be Poisson, however, on-or off-periods which correspond to busy and idle transmission times, respectively, are heavy-tailed. Starting with Chapter 4, many of the chapters employ asymptotic techniques to investigate tail behavior in queueing systems which, in turn, are related to buffer overflow or packet drop probabilities. Chapters 8, 9, and 10 provide asymptotic bounds on the tail probability. Chapter 12 discusses a traditional, Markovian view of modeling and analyzing variable bit rate video traces which represents a form of extreme adherence to conventional techniques and world view which has its roots in telephony traffic. Chapter 13 provides a form of transient analysis which, in spite of its elementary nature, is a useful exercise and points toward the need for nonequilibrium analysis.
A total of six chapters make up the third part of the book which is mainly concerned with traffic control and dynamic resource provisioning issues that arise under self-similar traffic conditions. There are two aspects to the question, one centered on the problem of resource provisioning/dimensioning and ensuing tradeoff relations, and the other based on the traditional traffic control framework of feedback control and its implementation in network protocols. With respect to resource provisioning, due to the amplified queueing delay incurred when employing buffer dimensioning, an alternative resource provisioning strategy based on bandwidth dimensioning as the central control variable has been advanced. A high-level discussion is provided in Chapter 16. Chapter 17 provides analysis of bufferless systems and long-range dependent processes whose future behavior is conditioned on past behavior which are relevant to on-line resource provisioning and traffic control. Chapter 19 describes a concrete resource provisioning architecture based on framing. Feedback traffic control presents a more subtle challenge to traffic management where the central idea revolves around exploiting correlation structure at multiple time scales, as afforded by long-range dependence and self-similarity, to affect traffic control decisions executed at smaller time scales. Chapter 14 discusses the influence of the protocol stack and network traffic, and Chapter 15 gives a detailed characterization of TCP based connection arrivals and network traffic which constitutes the bulk of current Internet traffic. Chapter 18 introduces the multiple time scale congestion control framework and its use in self-similar traffic for throughput maximization.
We conclude the book with two overview chapters which seek to take stock of known results, and point toward research avenues and open problems that may benefit from concerted efforts by the research community. Chapter 20 gives a broad overview of traffic characterization and modeling issues, with focus on achieving a comprehensive and refined understanding of network traffic spanning both long and short time scales. Chapter 21 describes a set of research problems and themes categorized into workload characterization, performance analysis, and traffic control. Some problems are more aptly described as research programs whereas other are more focused in their scope and nature.
As co-editors, we greatly appreciate the generous efforts of all the contributors to this volume. Because of their cooperation, flexibility, and willingness in helping us achieve a measure of coherence and balanced representation, this project has been a productive and timely occasion, and a delightful experience for us. We are confident that despite the rapidly changing conditions that have become a trademark of modern communication networks, this book contains insights and lessons that are less transient and will withstand the test of time. We hope the book will be of service as a comprehensive, in-depth, and up-to-date reference on self-similar network traffic for the larger networking and communication research communities. Our work would have been much more difficult and time consuming without the help of Wiley and its professional staff, especially, Andrew Smith who participated in the initial idea of the book and Rosalyn Farkas who provided critical editing support. We would like to extend our appreciation and thanks.
KIHONG PARK
WALTER WILLINGER
Purdue University
AT&T Labs
May 2000