End-to-end adaptive congestion control in TCP/IP networks

Establishing adaptive control as an alternative framework to design and analyze Internet congestion controllers, End-to-End Adaptive Congestion Control in TCP/IP Networks employs a rigorously mathematical approach coupled with a lucid writing style to provide extensive background and introductory material on dynamic systems stability and neural network approximation; alongside future internet requests for congestion control architectures. Designed to operate under extreme heterogeneous, dynamic, and time-varying network conditions, the developed controllers must also handle network modeling structural uncertainties and uncontrolled traffic flows acting as external perturbations. The book also presents a parallel examination of specific adaptive congestion control, NNRC , using adaptive control and approximation theory, as well as extensions toward cooperation of NNRC with application QoS control.

Features:

Uses adaptive control techniques for congestion control in packet switching networks Employs a rigorously mathematical approach with lucid writing style Presents simulation experiments illustrating significant operational aspects of the method; including scalability, dynamic behavior, wireless networks, and fairness Applies to networked applications in the music industry, computers, image trading, and virtual groups by techniques such as peer-to-peer, file sharing, and internet telephony Contains working examples to highlight and clarify key attributes of the congestion control algorithms presented

Drawing on the recent research efforts of the authors, the book offers numerous tables and figures to increase clarity and summarize the algorithms that implement various NNRC building blocks. Extensive simulations and comparison tests analyze its behavior and measure its performance through monitoring vital network quality metrics. Divided into three parts, the book offers a review of computer networks and congestion control, presents an adaptive congestion control framework as an alternative to optimization methods, and provides appendices related to dynamic systems through universal neural network approximators.

Author Notes

Christos N. Houmkozlis is currently in the Department of Electrical and Computer Engineering at Aristotle University of Thessaloniki. His research interests include nonlinear systems, robust adaptive control, modeling and control of communications networks, control over heterogeneous networks, resource management, and pricing in networks.

George A. Rovithakis is Associate Professor in the Department of Electrical and Computer Engineering at Aristotle University of Thessaloniki. His research interests include nonlinear robust adaptive control, neural networks for identification, control of uncertain systems, and control issues arising in computer networks.

List of Figures	p. xiii
List of Tables	p. xix
Preface	p. xxi
1 Introduction	p. 1
1.1 Overview	p. 1
1.2 Future Internet	p. 2
1.3 Internet Congestion Control	p. 4
1.4 Adaptive Congestion Control	p. 8
I Background on Computer Networks and Congestion Control	p. 13
2 Controlled System: The Packet-Switched Network	p. 15
2.1 Overview	p. 15
2.2 Network Connectivity	p. 17
2.2.1 Links and Nodes	p. 17
2.2.2 Sub-Networks	p. 17
2.2.3 Network Classification	p. 19
2.2.4 LAN Topologies	p. 21
2.3 Network Communication	p. 24
2.3.1 Packet Switching	p. 24
2.3.2 Protocols and Layering	p. 26
2.3.3 Internet Architecture	p. 28
2.3.4 Transfer Control Protocol (TCP)	p. 32
2.3.5 User Datagram Protocol (UDP)	p. 37
2.3.6 Internet Protocol (IP)	p. 38
2.4 Performance Characteristics	p. 40
2.4.1 Queue Size	p. 40
2.4.2 Throughput	p. 41
2.4.3 Link Utilization	p. 41
2.4.4 Packet Loss Rate	p. 41
2.4.5 Round Trip Time	p. 41
2.4.6 Fairness	p. 42
2.5 Applications	p. 43
2.5.1 E-Mail	p. 44
2.5.2 World Wide Web	p. 44
2.5.3 Remote Access	p. 45
2.5.4 File Transfer	p. 45
2.5.5 Streaming Media	p. 46
2.5.6 Internet Telephony (VOIP)	p. 46
2.6 Concluding Comments	p. 47
3 Congestion Issues and TCP	p. 49
3.1 Overview	p. 49
3.2 Core Issues in Congestion Control	p. 50
3.3 TCP: Flow Control and Congestion Control	p. 51
3.3.1 Slow Start	p. 52
3.3.2 Congestion Avoidance	p. 53
3.3.3 Fast Retransmit and Fast Recovery	p. 55
3.4 TCP Problems	p. 57
3.5 Managing Congestion	p. 59
3.5.1 TCP Friendliness	p. 59
3.5.2 Classification of Congestion Control Protocols	p. 60
3.5.2.1 Window-Based vs. Rate-Based	p. 60
3.5.2.2 Unicast vs. Multicast	p. 61
3.5.2.3 End-to-End vs. Router-Based	p. 62
3.6 Concluding Comments	p. 63
4 Measuring Network Congestion	p. 65
4.1 Overview	p. 65
4.2 Drop Tail	p. 66
4.3 Congestion Early Warning	p. 67
4.3.1 Packet Drop Schemes	p. 68
4.3.2 Packet Marking Schemes	p. 72
4.4 Concluding Comments	p. 77
5 Source-Based Congestion Control Mechanisms	p. 79
5.1 Overview	p. 79
5.2 Traditional TCP	p. 80
5.3 TCP Modifications for Networks with Large Bandwidth Delay Products	p. 81
5.3.1 Scalable TCP (STCP)	p. 82
5.3.2 HighSpeed TCP (HSTCP)	p. 82
5.3.3 BIC	p. 84
5.3.4 CUBIC	p. 85
5.4 Delay-Based Congestion Control	p. 86
5.4.1 TCP Vegas	p. 87
5.4.2 FAST TCP	p. 88
5.5 Congestion Control for Wireless Networks	p. 89
5.5.1 TCP Westwood	p. 90
5.5.2 TCP Veno	p. 91
5.6 Congestion Control for Multimedia Applications	p. 92
5.6.1 Rate Adaptation Protocol (RAP)	p. 92
5.6.2 TFRC	p. 94
5.7 Concluding Comments	p. 95
6 Fluid Flow Model Congestion Control	p. 97
6.1 Overview	p. 97
6.2 The Fluid Flow Model	p. 98
6.3 Network Representation	p. 99
6.4 Congestion Control as a Resource Allocation Problem	p. 101
6.4.1 Dual Approach	p. 103
6.4.2 Primal Approach	p. 104
6.4.3 Utility Function Selection	p. 104
6.5 Open Issues	p. 106
6.5.1 Stability and Convergence	p. 106
6.5.2 Implementation Constraints	p. 107
6.5.3 Robustness	p. 107
6.5.4 Fairness	p. 108
6.6 Concluding Comments	p. 109
II Adaptive Congestion Control Framework	p. 111
7 NNRC: An Adaptive Congestion Control Framework	p. 113
7.1 Overview	p. 113
7.2 Packet Switching Network System	p. 114
7.3 Problem Statement	p. 117
7.4 Throughput Improvement	p. 118
7.5 NNRC Framework Description	p. 120
7.5.1 Future Path Congestion Level Estimator	p. 121
7.5.2 Feasible Desired Round Trip Time Estimator	p. 122
7.5.3 Rate Control	p. 122
7.5.4 Throughput Control	p. 123
7.6 Concluding Comments	p. 123
8 NNRC: Rate Control Design	p. 125
8.1 Overview	p. 125
8.2 Feasible Desired Round Trip Time Estimator Design	p. 125
8.2.1 Proof of Lemma 8.1	p. 129
8.2.2 Proof of Lemma 8.2	p. 130
8.3 Rate Control Design	p. 132
8.3.1 Guaranteeing Boundness of Transmission Rate	p. 137
8.3.2 Reducing Rate in Congestion	p. 138
8.4 Illustrative Example	p. 140
8.4.1 Implementation Details	p. 141
8.4.2 Network Topology	p. 142
8.4.3 Normal Scenario	p. 142
8.4.4 Congestion Avoidance Scenario	p. 143
8.5 Concluding Comments	p. 148
9 NNRC: Throughput and Fairness Guarantees	p. 151
9.1 Overview	p. 151
9.2 Necessity for Throughput Control	p. 151
9.3 Problem Definition	p. 153
9.4 Throughput Control Design	p. 154
9.4.1 Guaranteeing Specific Bounds on the Number of Channels	p. 156
9.4.2 Reducing Channels in Congestion	p. 157
9.5 Illustrative Example	p. 158
9.5.1 Implementation Details	p. 159
9.5.2 Normal Scenario	p. 160
9.5.3 Congestion Avoidance Scenario	p. 163
9.5.4 Throughput Improvement	p. 163
9.6 Concluding Comments	p. 169
10 NNRC: Performance Evaluation	p. 171
10.1 Overview	p. 171
10.2 Network Topology	p. 172
10.3 Scalability	p. 174
10.3.1 Effect of Maximum Queue Length	p. 174
10.3.2 Effect of Propagation Delays	p. 176
10.3.3 Effect of Bandwidth	p. 178
10.4 Dynamic Response of NNRC and FAST TCP	p. 180
10.4.1 Bursty Traffic	p. 182
10.4.2 Re-Routing	p. 183
10.4.3 Non-Constant Number of Sources	p. 186
10.5 NNRC and FAST TCP Interfairness	p. 188
10.6 Synopsis of Results	p. 199
10.7 Concluding Comments	p. 200
11 User QoS Adaptive Control	p. 203
11.1 Overview	p. 203
11.2 Application Adaptation Architecture	p. 204
11.2.1 QoS Mapping	p. 204
11.2.2 Application QoS Control Design	p. 205
11.3 NNRC Source Enhanced with Application Adaptation	p. 207
11.4 Illustrative Example	p. 208
11.4.1 Application Adaptation Implementation Details	p. 209
11.4.2 Simulation Study	p. 210
11.5 Concluding Comments	p. 210
III Appendices	p. 215
A Dynamic Systems and Stability	p. 217
A.1 Vectors and Matrices	p. 217
A.1.1 Positive Definite Matrices	p. 219
A.2 Signals	p. 221
A.3 Functions	p. 223
A.3.1 Continuity	p. 223
A.3.2 Differentiation	p. 224
A.3.3 Convergence	p. 225
A.3.4 Function Properties	p. 226
A.4 Dynamic Systems	p. 227
A.4.1 Stability Definitions	p. 229
A.4.2 Boundedness Definitions	p. 231
A.4.3 Stability Tools	p. 232
B Neural Networks for Function Approximation	p. 247
B.1 General	p. 247
B.2 Neural Networks Architectures	p. 249
B.2.1 Multilayer Perceptron (MLP)	p. 250
B.2.2 Radial Basis Function Networks (RBF)	p. 252
B.2.3 High-Order Neural Networks (HONN)	p. 253
B.3 Off-Line Training	p. 255
B.3.1 Algorithms	p. 257
B.3.1.1 Gradient Algorithms	p. 257
B.3.1.2 Least Squares	p. 261
B.3.1.3 Backpropagation	p. 262
B.4 On-Line Training	p. 263
B.4.1 Filtering Schemes	p. 263
B.4.1.1 Filtered Error	p. 264
B.4.1.2 Filtered Regressor	p. 265
B.4.2 Lyapunov-Based Training	p. 266
B.4.2.1 LIP Case	p. 266
B.4.2.2 NLIP Case	p. 266
B.4.3 Steepest Descent Training	p. 267
B.4.4 Recursive Least Squares Training	p. 268
B.4.5 Robust On-Line Training	p. 269
Bibliography	p. 273
Index	p. 301

Available:*

On Order

Summary

Summary

Author Notes

Table of Contents