![The Big Picture 3
superhighway is still a dirt road; more accurately, it is a set of isolated multilane high-
ways with cow paths for entrance and exit ramps. The introduction of broadband access
to residential customers by the telephone carriers and the cable operators is a step toward
eliminating those cow paths. However, digital subscriber line (DSL) and cable modems
are half-measures at best. Direct access to the fiber network by the end user [i.e., fiber
to the home (FTTH) or business user] is the ultimate way of removing the bottleneck
so that the network remains effective as demand for bandwidth grows. Although FTTH
was deployed many years ago in a few demonstration projects, it did not take hold for
several reasons, including cost and the absence of services of interest to the customers.
Today that has changed because of the proliferation of broadband Internet services. De-
ployment of glass is now moving from the network core through fiber access networks
to the end users. This will undoubtedly stimulate interest in new broadband services that
take advantage of high-speed access and in turn produce demand for more bandwidth.
At this writing, most of the world’s installed fiber capacity is underutilized – arguably
due to the last-mile bottleneck. That should change rapidly as progress in the removal of
the bottleneck results in a quantum jump in network traffic, making high-performance
optical networks indispensable.
Higher level issues such as deregulation, new ideas for improving the economics
of networking, and standardization of control and management techniques in multi-
vendor networks are also contributing to the growing effectiveness of optical networks.
Deregulation, which began in the United States in 1984 with the dismantling of ATT,
has brought with it a new level of competition, with long-haul carriers, local carriers,
Internet service providers (ISPs), and cable operators poaching on each other’s domains
and using optical fiber capacity to do so. Bandwidth trading has been introduced as
a way of improving the utilization of fibers and thereby optimizing profits. A carrier
with idle capacity sells it to another carrier with excess demand. This type of exchange
requires sophisticated control and management tools for network reconfiguration. More
generally, any large network requires complex control and management systems and
intelligent network elements for performance monitoring, network reconfiguration, and
fault recovery. The systems, protocols, and equipment for performing these functions in
traditional telephone and data networks were built over many years by the public carri-
ers and equipment manufacturers. The new optical networks require similar tools, and
this is especially important in multivendor environments. These are now making their
appearance in the form of a proposed control plane for optical networks and protocols
for systems management in these networks. As more sophisticated control and manage-
ment functions are incorporated into optical networks the network operators are in a
better position to offer high-quality service to their customers, improving the operator’s
revenue stream and customer loyalty.
Aboveall, thelessons of thepast show us that tomorrow’s networks must beflexibleand
versatile enough to adapt to a continuing barrage of new and as-yet-unknown services.
It is interesting to note that when optical networks were still in an embryonic form,
the typical uses envisaged for them were high-tech applications such as high-resolution
medical image archiving and remote supercomputer visualization – basically usages
generated by a minuscule, elite segment of the population. Today these applications](https://image.slidesharecdn.com/40992-250303025055-b39b7448/85/Multiwavelength-optical-networks-architectures-design-and-control-2nd-Edition-Thomas-E-Stern-48-320.jpg)
Multiwavelength optical networks architectures design and control 2nd Edition Thomas E. Stern
- 1. Visit https://ebookultra.com to download the full version and explore more ebooks Multiwavelength optical networks architectures design and control 2nd Edition Thomas E. Stern _____ Click the link below to download _____ https://ebookultra.com/download/multiwavelength- optical-networks-architectures-design-and-control-2nd- edition-thomas-e-stern/ Explore and download more ebooks at ebookultra.com
- 2. Here are some suggested products you might be interested in. Click the link to download Communication Networks Fundamental Concepts and Key Architectures 2nd Edition Alberto Leon Garcia And Indra Widjaja https://ebookultra.com/download/communication-networks-fundamental- concepts-and-key-architectures-2nd-edition-alberto-leon-garcia-and- indra-widjaja/ Networks Design and Management 2nd Edition Steven T. Karris https://ebookultra.com/download/networks-design-and-management-2nd- edition-steven-t-karris/ Communication Networks An Optimization Control and Stochastic Networks Perspective Srikant R. https://ebookultra.com/download/communication-networks-an- optimization-control-and-stochastic-networks-perspective-srikant-r/ Broadband Optical Access Networks 1st Edition Leonid G. Kazovsky https://ebookultra.com/download/broadband-optical-access-networks-1st- edition-leonid-g-kazovsky-2/
- 3. Broadband Optical Access Networks 1st Edition Leonid G. Kazovsky https://ebookultra.com/download/broadband-optical-access-networks-1st- edition-leonid-g-kazovsky/ Robust Control System Design Advanced State Space Techniques Automation and Control Engineering 2nd Edition Chia-Chi Tsui https://ebookultra.com/download/robust-control-system-design-advanced- state-space-techniques-automation-and-control-engineering-2nd-edition- chia-chi-tsui/ Tactical Wireless Communications and Networks Design Concepts and Challenges 2nd Edition George F. Elmasry https://ebookultra.com/download/tactical-wireless-communications-and- networks-design-concepts-and-challenges-2nd-edition-george-f-elmasry/ Batch Control Systems Design Application and Implementation 2nd Edition William M. Hawkins https://ebookultra.com/download/batch-control-systems-design- application-and-implementation-2nd-edition-william-m-hawkins/ Distillation Design and Control Using Aspen Simulation 2nd Edition William L. Luyben https://ebookultra.com/download/distillation-design-and-control-using- aspen-simulation-2nd-edition-william-l-luyben/
- 5. Multiwavelength optical networks architectures design and control 2nd Edition Thomas E. Stern Digital Instant Download Author(s): Thomas E. Stern, Georgios Ellinas, Krishna Bala ISBN(s): 9780521881395, 0521881390 Edition: 2 File Details: PDF, 11.40 MB Year: 2008 Language: english
- 7. This page intentionally left blank
- 8. Multiwavelength Optical Networks, Second Edition Updated and expanded, this second edition of the acclaimed Multiwavelength Optical Networks provides a detailed description of the structure and operation of modern optical networks. It also sets out the analytical tools for network performance evaluation and optimization for current and next generation networks, as well as the latest advances in enabling technologies. Backbone optical networks are evolving to mesh topologies utilizing intelligent net- work elements; a new optical control plane is taking shape based on GMPLS; and significant advances have occurred in Fiber to the Home/Premises (the “last mile”), metropolitan area networks, protection and restoration, and IP over WDM. Each of these is treated in depth, together with new research on all-optical packet-switched net- works, which combine the speed of optics with the versatility of packet switching. Also included are current trends and new applications on the commercial scene (wavelengths on demand, virtual private optical networks, and bandwidth trading). With its unique blend of coverage of modern enabling technologies, network archi- tectures, and analytical tools, the book is an invaluable resource for graduate and senior undergraduate students in electrical engineering, computer science, and applied physics, and for practitioners and researchers in the telecommunications industry. Thomas E. Stern is Professor Emeritus of Electrical Engineering at Columbia University, New York, and has served as department chair and technical director of Columbia’s Center for Telecommunications Research. A Fellow of the IEEE, he holds several patents in networking. He has also been a consultant to a number of companies, including IBM, Lucent, and Telcordia Technologies. Georgios Ellinas is an Assistant Professor in the Department of Electrical and Computer Engineering at the University of Cyprus, Nicosia. He has held prior positions as an Associate Professor at City College of New York, as a Senior Network Architect at Tellium Inc., and as a Senior Research Scientist at Bell Communications Research. He has authored numerous papers and holds several patents in the field of optical networking. Krishna Bala is currently the CEO of Xtellus, a company that manufactures fiber opti- cal switches. Krishna was the co-founder and CTO of Tellium (NASDAQ: TELM), a successful optical networking company. Prior to that he was a Senior Research Scien- tist at Bell Communications Research. He holds a Ph.D. in electrical engineering from Columbia University.
- 10. Multiwavelength Optical Networks, Second Edition Architectures, Design, and Control THOMAS E. STERN Columbia University GEORGIOS ELLINAS University of Cyprus, Nicosia KRISHNA BALA Xtellus
- 11. CAMBRIDGE UNIVERSITY PRESS Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo Cambridge University Press The Edinburgh Building, Cambridge CB2 8RU, UK First published in print format ISBN-13 978-0-521-88139-5 ISBN-13 978-0-511-48054-6 © Cambridge University Press 2009 2008 Information on this title: www.cambridge.org/9780521881395 This publication is in copyright. Subject to statutory exception and to the provision of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. Cambridge University Press has no responsibility for the persistence or accuracy of urls for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate. Published in the United States of America by Cambridge University Press, New York www.cambridge.org eBook (NetLibrary) hardback
- 12. To Monique, who has always been there for me. To our children and our grand- children (T.E.S.) To my loving mother, Mary, and sister, Dorita, and the memory of my beloved father, Nicos (G.E.) To my wife, Simrat, and our children, Tegh and Amrita (K.B.)
- 14. Contents Figures page xvii Tables xxix Preface to the Second Edition xxxi Acknowledgments xxxv 1 The Big Picture 1 1.1 Why Optical Networks? 1 1.2 Objectives of an Optical Network Architecture 4 1.3 Optics versus Electronics: The Case for Transparent Multiwavelength Networks 9 1.4 Optics and Electronics: The Case for Multilayered Networks 12 1.5 Network Hierarchies 16 1.6 A Little History 18 1.7 Overview and Road Map 22 2 The Layered Architecture and Its Resources 28 2.1 Layers and Sublayers 29 2.2 Network Links: Spectrum Partitioning 34 2.3 Optical Network Nodes: Routing, Switching, and Wavelength Conversion 39 2.3.1 Static Nodes 40 2.3.2 Dynamic Nodes 46 2.3.3 Wavelength Converters 63 2.4 Network Access Stations 67 2.4.1 Transmitting Side 70 2.4.2 Receiving Side 71 2.5 Overlay Processors 74 2.5.1 Regeneration 76 2.5.2 Wavelength Interchange 76 2.6 Logical Network Overlays 77 2.6.1 SONET Networks 79 2.6.2 ATM Networks 81
- 15. viii Contents 2.6.3 IP Networks 83 2.6.4 MPLS and Its Extensions 84 2.7 Summary 85 2.8 Problems 87 3 Network Connections 91 3.1 Connection Management and Control 96 3.1.1 Optical Connections 100 3.1.2 Logical Connections 100 3.2 Static Networks 102 3.2.1 Point-to-Point and Multipoint Connections 104 3.2.2 Packet Switching in the Optical Layer: The MAC Sublayer 111 3.2.3 Additional Comments on Broadcast-and-Select 121 3.3 Wavelength-Routed Networks 122 3.3.1 Routing and Channel Assignment 124 3.3.2 Routing and Channel Assignment Examples 128 3.4 Linear Lightwave Networks: Waveband Routing 133 3.4.1 Routing and Channel Assignment 135 3.4.2 Multipoint Subnets in LLNs 140 3.4.3 A Seven-Station Example 143 3.5 Logically-Routed Networks 151 3.5.1 Point-to-Point Logical Topologies 153 3.5.2 Multipoint Logical Topologies: Hypernets 156 3.6 Summary 162 3.7 Problems 163 4 Enabling Technology 165 4.1 Evolution of Transmission and Switching Technology 166 4.2 Overview of the Optical Connection 167 4.3 Optical Fibers 168 4.3.1 Principles of Guided-Wave Propagation 168 4.3.2 Optical Fiber Technology: Transmission Impairments 174 4.3.3 Solitons 187 4.3.4 Photonic Crystal Fibers 188 4.4 Amplifiers 190 4.4.1 Erbium-Doped Fiber Amplifiers 191 4.4.2 Raman Amplifiers 198 4.4.3 Semiconductor Optical Amplifiers 201 4.4.4 Amplification Trends in Metro Optical Networks: Amplets 204 4.5 Optical Transmitters 205 4.5.1 Lasers 205 4.5.2 Vertical Cavity Surface Emitting Lasers 211 4.5.3 Modulation Technology 212
- 16. Contents ix 4.6 Optical Receivers in Intensity-Modulated Direct-Detection Systems 217 4.6.1 Photodetectors 217 4.6.2 Front-End Amplifier: Signal-to-Noise Ratio 219 4.6.3 Digital Signal Detection: Noise, Interference, and Bit Error Rate 221 4.6.4 Analog Systems: Carrier-to-Noise Ratio 227 4.7 The End-to-End Transmission Channel 228 4.7.1 Modulation Formats 229 4.7.2 Forward Error Correction 231 4.7.3 Equalization 233 4.8 Coherent Optical Systems 234 4.9 Performance Impairments in a Network Environment 235 4.9.1 Cross-Talk 235 4.9.2 Signal Power Divergence 239 4.9.3 Chirp-Induced Penalty 240 4.9.4 Optical Filter Concatenation: Distortion-Induced Penalty 240 4.9.5 Polarization Mode Dispersion Impact on System Performance 241 4.10 Optical and Photonic Device Technology 241 4.10.1 Couplers and Switches 242 4.10.2 Reciprocity 255 4.10.3 Nonreciprocal Devices 257 4.10.4 Optical Filtering Technology 257 4.10.5 Multiwavelength Switch Technology 266 4.11 Wavelength Conversion and Signal Regeneration 274 4.11.1 All-Optical Wavelength Conversion 275 4.11.2 Opaque Wavelength Conversion and Signal Regeneration 278 4.12 Optical Switch Architectures 281 4.12.1 Space Switches 281 4.12.2 Wavelength-Selective Switches 288 4.13 Performance Evaluation: Methodology and Case Studies 297 4.13.1 Physical-Layer Simulation: Three-Step Approach 298 4.13.2 WDM Network Simulation Case Studies 301 4.14 Problems 311 5 Static Multipoint Networks 324 5.1 Shared Media: The Broadcast Star 324 5.2 Representative Multiplexing and Multiple-Access Schemes 327 5.2.1 Time-Wavelength-Division Multiplexing/Multiple Access 328 5.2.2 Subcarriers 336 5.2.3 Code Division Multiple Access 352
- 17. x Contents 5.3 Traffic Constraints in Shared-Channel Networks 367 5.3.1 Balanced Traffic 370 5.3.2 Unbalanced Traffic 370 5.4 Capacity Allocation for Dedicated Connections 371 5.4.1 Fixed-Frame Scheduling for Stream Traffic 371 5.4.2 Fixed-Frame Scheduling for Packet Traffic 383 5.5 Demand-Assigned Connections 389 5.5.1 Blocking Calculations in WDMA Networks 390 5.5.2 Blocking in Combined Time-Wavelength-Division Networks 395 5.6 Packet Switching in the Optical Layer 399 5.6.1 Uncontrolled Scheduling: Random Access 401 5.6.2 Scheduling with Loss 403 5.6.3 Lossless Scheduling: Reservations 405 5.6.4 Perfect Scheduling 407 5.6.5 Dynamic versus Fixed Capacity Allocation 408 5.7 The Passive Optical Network 409 5.7.1 ATM and Fixed-Frame PONs 412 5.7.2 Ethernet-Based PONs 414 5.7.3 WDM PONs 416 5.7.4 Optical-Wireless Access 420 5.7.5 Recent Trends 422 5.8 Summary 424 5.9 Problems 425 6 Wavelength/Waveband-Routed Networks 432 6.1 Introduction 432 6.2 Physical Topologies 434 6.3 Wavelength-Routed Networks: Static Routing and Channel Assignment 442 6.3.1 Flow Bounds: Matching the Physical and Logical Topologies 444 6.3.2 Nonblocking Stations 448 6.3.3 RCA as a Graph Coloring Problem 449 6.3.4 Rings 452 6.3.5 Ring Decomposition of General Mesh Networks 458 6.3.6 Multistar Wavelength-Routed Networks 462 6.3.7 RCA as an Optimization Problem 464 6.3.8 Heuristics for Static RCA 474 6.4 Wavelength-Routed Networks: Dynamic Routing and Channel Assignment 484 6.4.1 Some Basic Routing and Channel Assignment Algorithms 484
- 18. Contents xi 6.4.2 Case Study: Bidirectional Rings 491 6.4.3 Performance of Dynamic Routing Rules on Meshes 494 6.4.4 Case Study: An Interconnected Ring 495 6.4.5 Routing Multicast Connections in WRNs 497 6.5 Linear Lightwave Networks: Static Routing Rules 507 6.5.1 Routing of Optical Paths 509 6.5.2 Optical Connections: λ-Channel Assignment 516 6.5.3 Significance of Nonblocking Access Stations in LLNs 518 6.5.4 Local Access to LLNs 519 6.5.5 Routing Waveband and Channel Assignment on the Petersen Network 521 6.5.6 Channel Assignment 528 6.5.7 Multistar Linear Lightwave Networks 540 6.6 Linear Lightwave Networks: Dynamic Routing Rules 544 6.6.1 Point-to-Point Connections 544 6.6.2 Routing Multicast Connections in LLNs 558 6.7 Problems 568 7 Logically-Routed Networks 576 7.1 Introduction: Why Logically-Routed Networks? 576 7.1.1 Multitier Networks: Grooming 581 7.2 Point-to-Point Logical Topologies: Multihop Networks 585 7.2.1 ShuffleNets 587 7.2.2 Families of Dense Logical Topologies 589 7.3 Multihop Network Design 591 7.3.1 Logical-Layer Design 591 7.3.2 Physical-Layer Design 594 7.3.3 Traffic Grooming in Point-to-Point Logical Topologies 597 7.4 Multipoint Logical Topologies: Hypernets 607 7.4.1 Capacity of a Multipoint Subnet 611 7.4.2 Families of Dense Hypernets 613 7.4.3 Kautz Hypernets 615 7.4.4 Hypernet versus Multihop 628 7.4.5 Multicast Virtual Connections 631 7.5 Hypernet Design 632 7.5.1 Logical-Layer Design 632 7.5.2 Physical-Layer Design 634 7.5.3 Traffic Grooming in Multipoint Logical Topologies 637 7.5.4 Multistar Realizations 639 7.6 Summary 641 7.7 Problems 642
- 19. xii Contents 8 Survivability: Protection and Restoration 647 8.1 Objectives of Protection and Restoration 648 8.2 Current Fault Protection and Restoration Techniques in the Logical Layer 650 8.2.1 Point-to-Point Systems 650 8.2.2 SONET Self-Healing Rings 654 8.2.3 SONET Self-Healing Ring Interconnection Techniques 657 8.2.4 Architectures with Arbitrary Mesh Topologies 663 8.3 Optical-Layer Protection: Point-to-Point and Ring Architectures 669 8.3.1 Point-to-Point Systems 669 8.3.2 Self-Healing Optical Rings 672 8.4 Optical-Layer Protection: Mesh Architectures 677 8.4.1 Shared Optical Layer Line-Based Protection 679 8.4.2 Optical Path-Based Protection 692 8.4.3 Segment Protection 700 8.4.4 Survivability Techniques for Multicast Connections 702 8.5 Summary 703 8.6 Problems 706 9 Optical Control Plane 714 9.1 Introduction to the Optical Control Plane 716 9.1.1 Control-Plane Architecture 719 9.1.2 Control-Plane Interfaces 719 9.1.3 Control-Plane Functions 721 9.2 Overview of Multiprotocol Label Switching 722 9.2.1 Packet Transport through an MPLS Network 722 9.2.2 MPLS Protocol Stack 727 9.2.3 MPLS Applications 728 9.3 Overview of Generalized Multiprotocol Label Switching 729 9.3.1 Link Management in GMPLS 731 9.3.2 Routing in GMPLS 734 9.3.3 Signaling in GMPLS 742 9.4 Conclusions 751 10 Optical Packet-Switched Networks 756 10.1 Optical Packet-Switched Network Architectures 758 10.1.1 Unbuffered Networks 759 10.1.2 Deflection Routing 764 10.1.3 Performance Analysis of Deflection Routing 766 10.1.4 Buffering: Time Domain Contention Resolution 770 10.1.5 Buffering and Wavelength Conversion: Time/Wavelength Domain Contention Resolution 778
- 20. Contents xiii 10.1.6 Comparison of Contention Resolution Techniques for Asynchronous OPS Networks 782 10.1.7 Hybrid Electronic and Optical Buffering 784 10.2 OPS Enabling Technologies 787 10.2.1 Packet Synchronization 788 10.2.2 All-Optical 2R or 3R Regeneration 788 10.2.3 Optical Switching 788 10.2.4 Wavelength Conversion 789 10.2.5 Optical Header Processing 789 10.2.6 Optical Buffering 789 10.3 OPS Network Testbed Implementations 791 10.3.1 CORD Testbed 791 10.3.2 KEOPS Testbed 793 10.3.3 WASPNET Testbed 796 10.4 Optical Burst Switching 798 10.4.1 Just Enough Time Protocol 801 10.4.2 Just In Time Protocol 803 10.4.3 Contention Resolution in OBS Networks 806 10.5 Optical Label Switching 808 10.5.1 All-Optical Label Swapping 809 10.5.2 Contention Resolution Techniques 811 10.5.3 OLS Network Implementations 811 10.6 Conclusions 820 10.7 Problems 822 11 Current Trends in Multiwavelength Optical Networking 828 11.1 Business Drivers and Economics 828 11.1.1 Cost Issues for WDM Point-to-Point Systems 831 11.1.2 Cost Issues for WDM Rings 832 11.1.3 Cost Issues for WDM Cross-Connect Networks 833 11.1.4 Open versus Closed WDM Installations 835 11.2 Multiwavelength Optical Network Testbeds 838 11.2.1 Optical Networks Technology Consortium 838 11.2.2 All-Optical Network Consortium 839 11.2.3 European Multiwavelength Optical Network Trials 839 11.2.4 Multiwavelength Optical Network 840 11.2.5 National Transparent Optical Networks Consortium 840 11.2.6 The Importance of the Testbeds in Driving the Telecommunications Infrastructure 840 11.3 Metropolitan Area Networks 841 11.3.1 Metro Network Unique Characteristics 841 11.3.2 Defining the Metropolitan Networking Domain 842
- 21. xiv Contents 11.3.3 Metro Network Evolution 844 11.3.4 Metro Networking State of the Art 847 11.4 Long-Haul and Ultra Long-Haul Networks 854 11.4.1 Current Considerations in Wide Area Network Architectures 854 11.4.2 Some Recent Commercial Network Deployments 856 11.5 New Applications and Services 858 11.5.1 Wavelength-on-Demand 858 11.5.2 Virtual Private Optical Networks 858 11.5.3 Bandwidth Trading 859 11.6 Conclusions 861 A Graph Theory 869 A.1 Graphs 869 A.1.1 Cycle Double Covers 872 A.1.2 Eulerian Graphs 872 A.1.3 Planar Graphs 873 A.1.4 Matchings in Graphs 873 A.1.5 Graph Coloring 874 A.1.6 Digraphs 875 A.1.7 Moore Bounds 875 A.1.8 Max Flow–Min Cut 876 A.2 Hypergraphs 877 A.2.1 Undirected Hypergraphs 877 A.2.2 Directed Hypergraphs 878 B Fixed Scheduling Algorithm 879 B.1 Column/Row–Expansion Algorithm 880 B.2 Decomposition into Permutation Matrices 883 B.3 Column/Row–Compression Algorithm 883 C Markov Chains and Queues 884 C.1 Random Processes 884 C.2 Markov Processes 885 C.3 Queues 887 C.3.1 The M |M |1 Queue 888 C.3.2 The M |G |1 Queue 888 C.3.3 Little’s Formula 889
- 22. Contents xv D A Limiting-Cut Heuristic 890 D.1 The Multicommodity Flow Problem and Limiting Cuts 890 D.2 A Heuristic 891 D.2.1 Swap (X, Y ) 891 D.2.2 Limcut 892 E An Algorithm for Minimum-Interference Routing in Linear Lightwave Networks 893 E.1 The Image Network 893 E.2 The Min-Int Algorithm 894 E.3 Minimum Interference 895 F Synopsis of the SONET Standard 896 G A Looping Algorithm 900 Acronyms 903 Index 915
- 24. Figures 1.1 Multilayered network. page 6 1.2 Physical picture of the network. 8 1.3 Layered view of an optical network. 13 1.4 Alternative network approaches. 15 1.5 Hierarchical network. 17 1.6 Road map. 24 2.1 Layered view of optical network connections. 29 2.2 A typical connection. 32 2.3 Client server associations in an optical transport network. 33 2.4 Fiber resources. 34 2.5 Wavelength and waveband partitioning of the optical spectrum. 35 2.6 Network picture based on spectrum partitioning. 38 2.7 Tree physical topologies. 41 2.8 Directional coupler. 42 2.9 A 16 × 16 star coupler. 43 2.10 Static routing node. 44 2.11 Directed star. 45 2.12 Space switch connection matrices. 46 2.13 Unidirectional ring. 47 2.14 Crossbar switch. 48 2.15 Clos switch. 49 2.16 Recursion for Benes switch. 49 2.17 An 8 × 8 Benes switch. 50 2.18 Generalized optical switch. 53 2.19 δ–σ linear divider-combiner. 54 2.20 A node without loopback connections. 55 2.21 Three-stage realization of a waveband-space switch. 56 2.22 Multiwaveband directional coupler. 57 2.23 WADM–NAS combination. 58 2.24 Wavelength conversion as a linear operation. 64 2.25 Wavelength interchanger. 65 2.26 Wavelength-interchanging switch. 65 2.27 WIXC implementation. 66 2.28 Wavelength-routed network. 67
- 25. xviii Figures 2.29 Network access station. 68 2.30 Example of a logical connection between two NASs. 69 2.31 Optical transmitter. 70 2.32 Optical receivers. 72 2.33 Heterodyne receiver and spectra. 73 2.34 Overlay processor. 75 2.35 Logically routed network. 78 2.36 SONET DCS. 81 2.37 ATM cell format. 82 2.38 ATM switch connections. 83 2.39 TCP/IP and OSI. 84 2.40 Taxonomy of multiwavelength networks. 86 3.1 End systems: full connectivity. 92 3.2 Star physical topology. 92 3.3 Bidirectional ring physical topology. 93 3.4 The control plane in an optical network. 98 3.5 Connection management system. 99 3.6 Star coupler example. 105 3.7 Time-shared medium. 106 3.8 A TDM/TDMA schedule. 108 3.9 TDM/T-WDMA. 110 3.10 CSMA/CD. 114 3.11 CSMA/CD collision. 115 3.12 NAS equipped for packet switching. 116 3.13 Packet switching in the optical layer. 118 3.14 MAC protocol in the layered architecture. 120 3.15 Wavelength-routed star network. 123 3.16 Channel assignment example. 125 3.17 Nonblocking access link. 129 3.18 Bidirectional ring: single access fiber pair. 130 3.19 Bidirectional ring: two access fiber pairs. 132 3.20 A mesh network. 133 3.21 Inseparability. 136 3.22 Two violations of DSC. 137 3.23 Inadvertent violation of DSC. 137 3.24 Avoidance of DSC violations. 138 3.25 Color clash. 139 3.26 Seven stations on a mesh. 141 3.27 Tree embedded in mesh. 142 3.28 Multistar network. 144 3.29 Embedded star on a bidirectional ring. 146 3.30 Seven-node hypernet. 148 3.31 Assumed channel spacings. 150 3.32 A logical switching node in an optical network. 152
- 26. Figures xix 3.33 Eight-node ShuffleNet. 154 3.34 ShuffleNet embedding. 155 3.35 Details of ShuffleNet node. 157 3.36 Twenty-two node hypernet. 159 3.37 Hypernet embedding. 160 4.1 A point-to-point optical connection. 167 4.2 Refractive index profiles for fibers. 169 4.3 Snell’s law. 169 4.4 Ray propagation in a step-index fiber. 170 4.5 Ray propagation in a graded-index fiber. 171 4.6 Cylindrical coordinates. 173 4.7 Commercial fiber cables. 175 4.8 Attenuation as a function of wavelength. 176 4.9 Broadening of pulses due to dispersion. 178 4.10 Dispersion coefficients as a function of frequency. 179 4.11 Limitations due to nonlinear effects in multiwavelength systems. 186 4.12 Soliton. 188 4.13 Three types of microstructured fibers. 189 4.14 Basic erbium-doped fiber amplifier structures. 192 4.15 Energy levels in EDFA. 192 4.16 EDFA gain profile. 194 4.17 Illustration of noise figure. 196 4.18 Raman gain coefficient in bulk silica as a function of frequency shift. 198 4.19 Hybrid distributed-discrete amplification. 200 4.20 Signal and pump power in hybrid system. 200 4.21 Fabry–Perot laser. 206 4.22 Single-frequency lasers. 208 4.23 Laser array. 210 4.24 Typical VCSEL structure. 211 4.25 Pulse and accompanying chirp. 213 4.26 Mach–Zehnder interferometer. 215 4.27 Typical structure of an EA-DFB transmitter. 216 4.28 Absorption and chirp (linewidth enhancement factor) parameters versus reverse bias voltage for a typical EA-DFB transmitter. 216 4.29 Photodiode. 217 4.30 Transimpedance amplifier. 219 4.31 Binary receiver. 221 4.32 Typical waveforms in an IM/DD system. 222 4.33 Eye diagram. 223 4.34 Ideal detection. 225 4.35 BER as a function of Q. 226 4.36 Transmission channel processing operations. 228 4.37 Modulation formats. 229 4.38 FEC encoding/decoding functions. 232
- 27. xx Figures 4.39 Transversal decision-directed equalizer. 233 4.40 Heterodyne receiver. 234 4.41 Types of cross-talk. 237 4.42 Power penalty with homodyne cross-talk. 238 4.43 Controllable directional coupler. 245 4.44 Mach–Zehnder switch. 246 4.45 Two-stage Mach–Zehnder switch. 247 4.46 Y-branch switch. 248 4.47 Gate array switch. 249 4.48 Laser-activated bubble switch element. 250 4.49 2D mechanical switch using micromachined mirrors. 251 4.50 3D MEMS switch. 252 4.51 3D gimbaled mirror. 253 4.52 Liquid crystal holographic switch. 254 4.53 Two hologram N × N liquid crystal holographic switch. 255 4.54 Illustration of reciprocity. 256 4.55 Optical isolator. 258 4.56 Fabry–Perot filter and its spectral response. 260 4.57 MI filter. 262 4.58 MI filter array. 263 4.59 FBG used as a drop filter. 264 4.60 A Mach–Zehnder WADM. 264 4.61 Arrayed waveguide grating. 266 4.62 Acousto-optic tunable filter. 267 4.63 Liquid crystal MWS. 269 4.64 A MEMS-based WADM. 270 4.65 An MI filter-based WADM. 272 4.66 Wavelength-dilated switch. 273 4.67 Optoelectronic wavelength converter. 275 4.68 Performance of a difference frequency converter. 277 4.69 Opaque conversion and regeneration. 278 4.70 SA-based regenerator. 279 4.71 Nonlinear Mach–Zehnder regenerator. 279 4.72 Nonlinear optical loop mirror regenerator. 280 4.73 Optical crossbar switch. 282 4.74 Path-independent loss crossbar switch. 282 4.75 Circuit layout for 8 × 8 optical crossbar switch. 284 4.76 Router/selector. 285 4.77 Benes switch. 285 4.78 Orders of cross-talk. 286 4.79 Enhanced performance switch. 287 4.80 Space dilation. 288 4.81 OADMs in a network. 289
- 28. Figures xxi 4.82 Parallel and serial OADM architectures with capability for m wavelength add/drops. 290 4.83 Functional diagram of an OADM based on wavebands and wavelengths. 290 4.84 Typical B&S OADM architecture. Assumes (1+1) protection. 292 4.85 Typical 4 × 4 ROADM based on 4 × 1 wavelength selective switch and B&S-type of architecture. 292 4.86 Transparent OXC. 293 4.87 Opaque O-E-O OXC. 294 4.88 Opaque O-O-O OXC. 294 4.89 Hybrid waveband/wavelength switch. 297 4.90 Wavelength-domain simulation. 299 4.91 WADM chain. 302 4.92 WADM structure and simulation model. 302 4.93 Simulation results for the WADM chain. 303 4.94 Ring interconnect network architecture. Worst-case paths between A and B are indicated. 304 4.95 Histogram of all cross-talk terms accumulated at receiver B for the worst-case path of Figure 4.94. 306 4.96 Cross-talk-induced Q penalty in dB versus dominant cross-talk term power level. 307 4.97 Q-channel performance for the worst-case path of Figure 4.94 assuming OC-192 bit rate and EA-modulated transmitters. 308 4.98 A DWDM metro network deployment scenario. All rings represent typical SONET OC-12/48/192 designs. DWDM is deployed only between the superhub nodes (dark squares) in ring (solid) or possible mesh (dotted) configurations. 309 4.99 DWDM metro network case study based on the network deployment scenario presented in Figure 4.98. Nodes represent only superhub stations with typical distances (not shown to scale). 309 4.100 Simulation results for path A-F-D in Figure 4.99 comparing Q-channel performance with and without EDC. 311 5.1 Star networks. 325 5.2 A 3 × 3 example. 330 5.3 TDM/T-WDMA channel allocation schedules. 331 5.4 Illustrating channel reuse in an FT-TR system. 332 5.5 Illustrating optical spectral efficiency. 334 5.6 SCMA example. 337 5.7 Transmitting and receiving stations equipped for SCMA. 337 5.8 Subcarrier spectra. 340 5.9 Effect of OBI. 342 5.10 TDM/T-SCMA. 346 5.11 SCM/SCMA. 349
- 29. xxii Figures 5.12 SCM/WDMA/SCMA. 350 5.13 SCM/WDMA/SCMA example. 351 5.14 Block diagram of a direct-detection CDMA system. 354 5.15 Waveforms for a direct-detection CDMA system. 356 5.16 Orthogonal optical codes. 357 5.17 Parallel CDMA transceiver structure. 359 5.18 CDMA with all-optical processing. 359 5.19 Multidimensional codes. 361 5.20 FBG encoder for FFH-CDMA. 361 5.21 A 3D CDMA system. 363 5.22 Realization of coherent optical CDMA. 364 5.23 Shared-channel broadcast medium. 367 5.24 Normalized traffic matrices. 370 5.25 CASs for systems with a full complement of channels. 376 5.26 CASs for Examples 4, 5, 6, and 7. 378 5.27 Heterogeneous traffic scheduling. 379 5.28 Logical multicast CAS. 382 5.29 Single-server queue. 385 5.30 Throughput versus traffic intensity. 387 5.31 Markov chain model for demand-assigned traffic. 391 5.32 Comparison of Engset and Erlang models. 393 5.33 Normalized throughput versus traffic intensity. 394 5.34 Normalized throughput versus traffic intensity. 395 5.35 Matching time slots. 396 5.36 Framed system blocking probabilities. 397 5.37 Illustrating rearrangeability. 398 5.38 Slotted ALOHA. 403 5.39 Tell-and-go protocol. 404 5.40 Lossless scheduling. 406 5.41 Queues for perfect scheduling. 407 5.42 Passive optical network. 410 5.43 BPON frame. 412 5.44 Transmission scenario in a BPON system. 413 5.45 PON equipped for decentralized control. 416 5.46 LARNet. 417 5.47 RITE-Net. 418 5.48 WDM PON. 419 5.49 Integrated system for dual services. 421 5.50 Dual services testbed. 421 5.51 DWDM/TDM PON. 423 6.1 Number of vertices in known maximal graphs. 435 6.2 Thirty-eight-vertex graph. 435 6.3 Tessellations of the plane. 436 6.4 Undirected deBruijn and Kautz graphs. 437
- 30. Figures xxiii 6.5 Construction for d min. 438 6.6 Plot of d min as a function of N. 438 6.7 Internodal distances in random networks. 439 6.8 Recursive grid. 440 6.9 Hierarchical Petersen graph. 441 6.10 Limiting cuts for four networks. 447 6.11 Three-node network. 449 6.12 Illustrating RCA in a wavelength-routed network. 451 6.13 A four-fiber SPRING. 453 6.14 A two-fiber SPRING. 454 6.15 Bidirectional ring. 455 6.16 Five-node WDM ring. 458 6.17 Ring decomposition. 459 6.18 Bridged ring overlay. 462 6.19 A multistar network. 462 6.20 Layered view of RCA. 465 6.21 External traffic in flow conservation equations. 467 6.22 Wavelength savings by increasing fibers. 476 6.23 Mean values of Nλ versus α. 477 6.24 Minimum values of Nλ versus α. 478 6.25 Flow chart of the Monte Carlo algorithm. 481 6.26 Time trace of Monte Carlo algorithm. 483 6.27 An example of SPD routing. 488 6.28 Blocking on an 11-node WDM ring. 490 6.29 Gain in blocking; 11-node WDM ring, simulation. 491 6.30 Fairness ratio; 11-node WDM ring, simulation. 492 6.31 Fairness ratio improvement versus interchanger density; 11-node WDM ring with 32 wavelengths. 493 6.32 Simulation and asymptotic analysis; 195-node interconnected WDM rings. 496 6.33 Blocking improvement with wavelength interchange; 195-node interconnected WDM rings. 497 6.34 Fairness ratio improvement with wavelength interchange; 195-node interconnected WDM rings. 498 6.35 Fairness ratio improvement versus interchanger density; 195-node interconnected WDM ring, 32 wavelengths. 499 6.36 Multicast connection in a transparent network. 500 6.37 A P×P split-and-deliver switch. 501 6.38 A P×P multicast-capable optical cross-connect based on a split-and-deliver switch. 502 6.39 A P×P multicast-capable optical cross-connect based on splitter sharing. 503 6.40 Multicasting in a network with sparse splitting capabilities. 506
- 31. xxiv Figures 6.41 Petersen network. 509 6.42 Structure of a nonblocking access station for an LLN. 510 6.43 Optical paths. 512 6.44 Optical connection hypergraph. 518 6.45 Local access subnets on the Petersen network. 520 6.46 Embedded star on tree TA. 523 6.47 Waveband assignments: W = 5. 526 6.48 Connection interference graph. 529 6.49 Connection interference graph for Equation (6.60). 530 6.50 Optical connection hypergraph. 534 6.51 Fixed-frame scheduling for four LCs. 534 6.52 Directed hypernet GKH (2, 8, 4, 4). 542 6.53 Color clash. 547 6.54 Illustrating inseparability. 548 6.55 Illustrating Min-Int. 550 6.56 Random network. 551 6.57 Max Reuse versus Min Reuse channel allocation. 552 6.58 k-SP routing. 553 6.59 k-SP versus Min-Int routing. 554 6.60 Blocking in networks with multifiber links. 555 6.61 Blocking in networks with multiple wavebands. 556 6.62 Example of a multicast connection. 559 6.63 Example of a tree decomposition using MBFS-1. 562 6.64 Example of a tree decomposition using MBFS-4. 563 6.65 Illustrating routing on a tree. 564 6.66 Blocking probability for multicast connections. 568 7.1 Why logically-routed networks? 577 7.2 A schematic of a point-to-point LRN. 579 7.3 Two-tier architecture. 582 7.4 The architecture of a grooming node with optical bypass. 584 7.5 ShuffleNet: δ = 3, k = 2, N = 18. 587 7.6 Maximum throughput per node for ShuffleNet. 588 7.7 deBruijn and Kautz digraphs. 590 7.8 A traffic matrix and matched LCG. 593 7.9 ShuffleNet on Atlantis. 595 7.10 Benefit of traffic grooming. 598 7.11 A node in a SONET over WDM ring. 599 7.12 Advantage of grooming static traffic in SONET over WDM rings. 601 7.13 Construction of an auxiliary graph for grooming. 605 7.14 Layered view of a hypernet. 608 7.15 Hypernets. 609 7.16 Illustrating fan-out in hypergraphs. 613 7.17 Shuffle hypernet. 614 7.18 Orders of K H(2, D,r). 616
- 32. Figures xxv 7.19 Duality construction. 618 7.20 Directed hypergraph construction via duality. 619 7.21 Directed hypergraph construction via edge grouping. 620 7.22 Tripartite representation of GK H(2, 42, 3, 28). 622 7.23 Routing header. 624 7.24 Comparison of hypernets and multihop networks. 628 7.25 Multicast tree in DK H(1, 4, 2). 631 7.26 Multicast-capable logical-grooming switch. 638 8.1 Path versus line protection. 651 8.2 (1 + 1) SONET protection. 652 8.3 (1:1) SONET protection. 653 8.4 (1:N) SONET protection. 654 8.5 Single- and dual-access ring interconnection configurations. 658 8.6 Two-fiber UPSR-to-UPSR ring interconnection. 660 8.7 BLSR-to-BLSR ring interconnection. 661 8.8 BLSR-to-UPSR ring interconnection. 662 8.9 (1 + 1) Protection in the optical layer. 670 8.10 (1:1) Protection in the optical layer. 671 8.11 (1:N) Protection in the optical layer. 671 8.12 (1 + 1) Optical protection and (1:N) electronic protection for a WDM system. 672 8.13 Four-fiber WDM SPRING architecture. 673 8.14 Four-fiber WDM SPRING surviving a link failure. 674 8.15 Four-fiber WDM SPRING surviving a node failure. 675 8.16 Two-fiber WDM SPRING architecture. 676 8.17 A taxonomy of survivability schemes. 677 8.18 Rerouting around a failed link. 682 8.19 Directed cycles in a planar graph. 683 8.20 Directed cycles in a nonplanar graph: K5. 683 8.21 Face traversal for a planar national network. 684 8.22 Orientable CDC of the ARPANet. 686 8.23 Seven-node planar network with default protection switch settings. 687 8.24 Seven-node planar network after a link failure. 687 8.25 Failure recovery using the p-cycle approach. 688 8.26 Generalized loopback example. 691 8.27 (1+1) dedicated protection architecture. 692 8.28 (1:3) shared protection in a mesh network. 693 8.29 Spanning trees used in optical path protection. 697 8.30 Shared risk groups. 698 8.31 SRG classification. 699 8.32 SLSP protection scheme. 701 8.33 Examples of different types of islands centered on node 22. 702 8.34 Examples of segment and path-pair protection of multicast sessions. 703 8.35 Example illustrating the arc-disjoint and MC-CR algorithms. 704
- 33. xxvi Figures 9.1 A mesh optical network. 715 9.2 Example of an optical node architecture. 716 9.3 Provisioning a connection between two routers through an optical network. 717 9.4 Control plane architecture. 720 9.5 Control plane interfaces. 721 9.6 MPLS header format and MPLS packet format. 724 9.7 Two LSPs in an MPLS packet-switched network. 725 9.8 Label stacking. 727 9.9 MPLS protocol stack. 728 9.10 Link bundling illustration. 740 9.11 LSP hierarchy in GMPLS. 741 9.12 User-Network Interface. 743 9.13 Provisioning in GMPLS. 744 9.14 Path and Resv message flows in RSVP for resource reservation. 746 9.15 RSVP message format. 747 9.16 Protection signaling using GMPLS RSVP-TE. 751 10.1 Optical packet-switching node. 760 10.2 A generic OPS node architecture for an unslotted network. 761 10.3 A generic OPS node architecture for a slotted network. 761 10.4 An FDL-based synchronizer. 762 10.5 A generic packet format for a slotted network. 762 10.6 Optical packet contention. 764 10.7 Petersen network graph. 766 10.8 Paths from A to D. 766 10.9 Input buffered optical packet switch with WSXC. 771 10.10 Input buffered optical packet switch using multiple space switch planes. 772 10.11 FDL input buffer. 772 10.12 Example of head-of-the-line (HOL) blocking. 773 10.13 Feed-forward delay line architecture. 773 10.14 Feedback delay line architecture. 774 10.15 Dump-and-insert buffer architecture. 775 10.16 Typical packet sequence in DI buffers for a 4 × 4 optical switch. 776 10.17 Generic node architecture with TOWCs at the input lines. 779 10.18 Details of output buffers for TOWC switch. 780 10.19 Packet-loss probability versus number of FDLs with and without wavelength conversion. 780 10.20 Generic node architecture with TOWCs that are shared among input lines. 781 10.21 Node architectures for different contention resolution schemes: single-wavelength delay line, multiwavelength delay line, wavelength
- 34. Figures xxvii conversion, and wavelength conversion with multiwavelength buffering. 783 10.22 Switch architecture with electronic buffering and wavelength conversion. 784 10.23 All-optical buffering and switching architecture. 790 10.24 Physical implementation of the CRO device. 792 10.25 CORD testbed. 792 10.26 Packet format for the KEOPS project. 794 10.27 Proposed unicast node architecture for the KEOPS project. 794 10.28 Proposed multicast/broadcast node architecture for the KEOPS project. 795 10.29 SLOB architecture. Each stage is a photonic switch element (PSE). 796 10.30 WASPNET optical packet switch. 797 10.31 Optical packet switching and Optical burst switching. 799 10.32 OBS architecture concept. 800 10.33 Just enough time (JET) protocol. 802 10.34 Just in time (JIT) protocol. 804 10.35 Segmentation of a burst. 807 10.36 Packet-loss probability versus load for different contention resolution policies in OBS. 808 10.37 OLS network. 809 10.38 All-optical processor for OLS. 810 10.39 OLS subcarrier transmission system. 812 10.40 OLS network node. 813 10.41 Network node architecture for an OLS testbed demonstration. 814 10.42 FSK/IM orthogonal labeling scheme used in the STOLAS project. 814 10.43 Optical router and optical label swapper used in the STOLAS project. 815 10.44 Architecture of the edge-router in OPSnet. 816 10.45 Architecture of the core-router in OPSnet. 817 10.46 Core node configuration for label swapping and packet switching. 819 10.47 OCSS technique. 819 10.48 Experimental setup for multihop packet transmission and multirate payload. 820 10.49 K3,3 network. 822 11.1 Six Central Offices, including two hubs, with capacity exhaust. 831 11.2 Application of WDM point-to-point systems to alleviate capacity exhaust. 832 11.3 Six Central Offices, including two hubs, with capacity exhaust. 833 11.4 Economic case for WDM rings. 834 11.5 Node in Central Office: Electronic cross-connect. 835 11.6 Economic case for WDM optical cross-connect. 836 11.7 Open WDM network architecture: Opaque network. 837 11.8 Integrated closed WDM network architecture. 837 11.9 Current legacy SONET/SDH design in U.S. metropolitan regions. 842
- 35. xxviii Figures 11.10 Metro WDM interconnected-ring simulation case study. 844 11.11 Typical ILEC metro network in the 2004 time frame. 845 11.12 Typical view of a superhub in the ILEC metro network of Figure 11.11 (2004 time frame). 846 11.13 Typical view of a current superhub in an ILEC metro network. 847 11.14 Typical deployment of 10 GbE technology in the metro environment. 848 11.15 Typical metro network ring architectures. 850 11.16 Typical vendor generic WADM node architecture for the metro network application space of Figure 11.15. 851 11.17 Metro network case study. 853 11.18 Wavelength-brokering operational model. 860 11.19 Infrastructure swapping. 860 11.20 Multicarrier recovery. 861 A.1 A maximal independent set. 870 A.2 The complete graph K5. 870 A.3 The complete bipartite graph K3,3. 871 A.4 Orientable cycle double cover for K3,3. 872 A.5 Multigraphs: Non-Eulerian and Eulerian. 873 A.6 Maximum matching of a bipartite multigraph. 874 A.7 A diclique. 876 A.8 A cut. 877 B.1 Example of decomposition of an NQDS matrix. 881 B.2 Example of fixed-frame scheduling. 882 C.1 Two-state chain. 886 C.2 Birth–death process. 887 C.3 A queue. 888 E.1 Image network. 894 F.1 SONET STS-1 frame and overhead channels. 897 F.2 Creating an OC-N signal. 898 F.3 Structure of a concatenated SONET frame. 898 G.1 Looping, first step. 901 G.2 Final switch settings. 901
- 36. Tables 3.1 Tree routing on mesh 149 3.2 Seven-station comparisons 150 3.3 Wavelength assignments for ShuffleNet on a ring 156 5.1 SCM/WDMA/SCMA example 351 5.2 Multicast connections 382 6.1 Orders of some graphs 437 6.2 Comparative performance of three RCA heuristics 480 6.3 Routing table 517 6.4 Waveband routing in the Petersen graph: W = 3 524 6.5 Waveband routing in the Petersen graph: W = 5 527 6.6 Comparison of performance of various configurations of the Petersen network 538 6.7 Sizes of trees generated with the MBFS-d algorithm 567 7.1 ShuffleNet routing on Atlantis 596 7.2 Relative costs for different ring networks 602 7.3 Kautz hypergraphs 615 7.4 Performance of K H(2, D∗ ,r) 627 7.5 Performance comparison of DK H(d, D, s) to multihop networks 630 7.6 Hypernet trees on Atlantis 636
- 38. Preface to the Second Edition The first edition of this book was published when optical networks were just emerging from the laboratory, mostly in the form of government-sponsored testbeds. Since then there have been fundamental changes in many aspects of optical networking, driven by the move from the laboratory to commercial deployment and by the twists and turns of the world economy. The investment climate in which optical networks have developed has had two major swings as of this writing. During the technology bubble that began at the end of the 20th century, investment in research, product development, and network deployment increased enormously. The activities during this time of euphoria produced advances in the technology base that would not have been possible without the extraordinary momentum of that period. At the same time, commercial network deployment provided a reality check. Some ideas that were pursued in the late 1990s dropped by the wayside because they did not meet the test of commercial viability, and new ones came along to take their place. When the bubble burst after less than a decade of “irrational exuberance,” the pendulum swung the other way. Investors and executives who a short time earlier thought the sky was the limit now wondered if demand would ever materialize for all of the fiber capacity in the ground. At this writing a more reasoned approach has taken hold; that seemingly elusive demand has materialized and, hopefully, a more rational and sustainable growth period will ensue. This is the context for the second edition. It is designed to build on the foundations laid out in the first edition while reflecting the new developments of the past 9 years: a maturing underlying technology, new tools for network control, and a recognition of the latest directions of optical network deployment and research. These new directions include cost-effective metropolitan area network architectures tailored to the strengths of current optical transmission and switching equipment, passive optical networks to bring high-speed access to the end user, hybrid optical/electronic architectures supporting the merging of multiwavelength and Internet technologies, and networks of the future based on all-optical packet switching. As in the first edition, the emphasis of this book is on concepts and methodologies that will stand the test of time. The first three chapters provide a qualitative foundation for what follows, presenting an overview of optical networking (Chapter 1), the multiwave- length network architecture and its supporting components (Chapter 2), and a high-level view of the different network structures considered throughout the book (Chapter 3). A more detailed picture is provided in the remaining chapters, with a survey of enabling technology (Chapter 4) and in-depth studies of the three basic network structures: static
- 39. xxxii Preface to the Second Edition multipoint networks (Chapter 5), wavelength/waveband routed networks (Chapter 6), and optical/electronic (logically routed) networks (Chapter 7). The remaining chapters complete the networking picture: survivability (Chapter 8), network control (Chapter 9), optical packet switching (Chapter 10), and current trends (Chapter 11). The first three chapters are suitable for the reader who wishes to gain an understanding of multiwavelength networks without delving deeply into the analytical tools for net- work design and the physical underpinnings of the optical technology. These beginning chapters, together with Chapter 11, would be suitable for a short undergraduate course for electrical engineers and computer science majors. The first seven chapters provide a largely generic framework for understanding net- work architectures, performance, and design in an abstract setting. An exception is Chapter 4, which surveys enabling technology from theory to practice, thereby provid- ing the necessary background concerning the physical limitations and possibilities of the network technology. The material through Chapter 7, together with selected material from the remaining chapters (depending on the reader’s orientation), can form the basis of a comprehensive graduate course, introducing the student to the latest developments in the field and suggesting a host of different research directions. The networking developments since the publication of the first edition have served to reorient and expand our treatment in significant ways. r Recognizing the importance of current activity in the “last mile” (fiber to the home/premises), and in metropolitan area networks, we have added a new section on passive optical networks (PONs) in Chapter 5,1 and we have included new material in Chapters 4 and 11 to connect our generic networking approach to recent metro network developments. r Chapter 4 was substantially expanded and updated to provide a glimpse of the im- pressive new trends in photonic and electro-optic technology. Some of the new and/or expanded topics are photonic crystal fibers, Raman amplification, supercontinuum generation, amplification trends in metro networks, and forward error correction and equalization to improve transmission performance. There are also new and expanded sections on wavelength conversion and signal regeneration with emphasis on all- optical techniques, and a new section on microelectromechanical system (MEMS) devices. The treatment of optical switch architectures has been significantly enlarged with a focus on cost-effective architectures and opacity versus transparency. More emphasis is placed on the effects of signal impairments, including a new section on performance impairments in a network environment. Also, new case studies are in- cluded that illustrate methodologies for evaluating the performance of metropolitan area networks. r Chapter 8, on survivability, protection, and restoration, was extensively updated, con- sistent with the growing importance of optical layer fault management in current 1 It is interesting to note that the PON, epitomized by the broadcast star, was the first structure that demonstrated the possibilities of optical networking in the 1980s. However, it was largely ignored for large-scale network deployment until recently, when it has again come into its own as the vehicle of choice for extending optical networks to the end user.
- 40. Preface to the Second Edition xxxiii networks. It contains recent work on the subject, including shared line-based protection in mesh networks, path-based protection, ring-based protection, segment protection, the treatment of shared risk groups, and recovery of multicast connections. r Chapter 9, describing the control plane, was added to present the latest developments in optical network control. It describes the control plane architecture as it has developed through the recent activities of several standards organizations. The chapter offers a detailed discussion of Multiprotocol Label-Switching (MPLS) and Generalized MPLS (GMPLS) as it applies to optical networks. r Chapter 10, on optical packet-switched networks, was added to provide an introduction to this emergent field.2 It provides a window on a cutting-edge research area that has the potential to offer the next breakthroughs in optical networking. r Chapter 11, on current networking trends (replacing the original Chapter 9), is a completely updated description of the current networking environment. This includes a historical perspective describing the pioneering network testbeds, business drivers, and current trends in metro, long-haul, and ultra long-haul netwoks. Included also are some new applications that have emerged on the commercial scene, such as wavelengths on demand, virtual private optical networks, and bandwidth trading. r This edition places increased emphasis on the practical aspects of hybrid (i.e., elec- tronic/optical and wavelength/waveband) architectures. This includes the importance of grooming, which is required to pack electronically multiplexed channels efficiently into an optical wavelength channel, and to pack wavelength channels into wavebands. Also, the existence of transparent (purely optical) and opaque (electronic/optical) alter- natives to network design is stressed throughout. These practical aspects of networking have become important as optical networks have found their place in the real world. Exercises are provided for most of the chapters, and many of them suggest avenues for future study. The book is meant to offer several different alternatives for study depending on the interest of the reader, be it understanding the current state of the field; acquiring the analytical tools for network performance evaluation, optimization, and design; or performing research on next-generation networks. 2 Although the idea of using packet switching in optical networks is not new, it has attracted renewed interest as the technology for purely optical packet processing has developed over the past few years, and the advantages of merging Internet and WDM technology have become apparent.
- 42. Acknowledgments The first edition of this book had its origins in 1990, when we organized a small group within the Center for Telecommunications Research (CTR) at Columbia University, to investigate lightwave networks. Among the many colleagues, students, and friends who contributed in various ways to the first edition, there are several who have continued to interact with us in the preparation of the second edition. We specifically express our thanks to Eric Bouillet, Aklilu Hailemariam, Gang Liu, and G.-K. Chang. Special thanks go to Ioannis Roudas for useful discussions and comments. Mischa Schwartz, who was singled out as our guiding spirit in the first edition, is still an indefatigable contributor to communication networking and a continuing inspiration to us. We are especially indebted to Neophytos Antoniades for coauthoring Chapters 4 and 11 of the second edition. His understanding of the role of physical layer simulation and the evolution of optical networks in the metropolitan area domain provided invaluable additions to this edition. We also express our thanks to Phil Meyler at Cambridge University Press for his support and encouragement, and to Anna Littlewood at Cambridge and Barbara Walthall at Aptara for their help in putting everything together. Finally, Thomas Stern expresses his profound gratitude to his wife, Monique, for her everlasting support; Georgios Ellinas is deeply grateful to his mother, Mary, and sister, Dorita, for their unyielding support and understanding during this endeavor; and Krishna Bala is greatly indebted to his wife, Simrat, and children, Tegh and Amrita, for their patience and support.
- 44. Multiwavelength Optical Networks, Second Edition
- 46. 1 The Big Picture Since the beginning of the 21st century there has been a burgeoning demand for com- munications services. From the ubiquitous mobile phone, providing voice, images, mes- saging, and more, to the Internet and the World Wide Web, offering bandwidth-hungry applications such as interactive games, music, and video file sharing, the public’s ap- petite for information continues to grow at an ever-increasing pace. Underneath all of this, essentially unseen by the users, is the optical fiber-based global communications infrastructure – the foundation of the information superhighway. That infrastructure contains the multiwavelength optical networks that are the theme of this book. Our purpose is to present a general framework for understanding, analyzing, and designing these networks. It is applicable to current network architectures as they have evolved since the mid-1990s, but more importantly it is a planning and design tool for the future. Our approach is to use a generic methodology that will retain its relevance as networks, applications, and technology continue to evolve. 1.1 Why Optical Networks? Since the fabrication of the first low-loss optical fiber by Corning Glass in 1970, a vision of a ubiquitous and universal all-optical communication network has intrigued researchers, service providers, and the general public. Beginning in the last decades of the 20th century enormous quantities of optical fiber were deployed throughout the world. Initially, fiber was used in point-to-point transmission links as a direct substitute for copper, with the fibers terminating on electronic equipment. Glass fiber was and is the ideal medium because of its many superior properties: extraordinary bandwidth, low loss, low cost, light weight and compactness, strength and flexibility, immunity to noise and electromagnetic interference, security and privacy (it is difficult to tap them), and corrosion resistance. Although all of these qualities make the fiber a technological marvel, fibers do not become networks until they are interconnected in a properly structured architecture. For our purposes, an optical network is a telecommunications network with transmission links that are optical fibers, and with an architecture that is designed to exploit the unique features of fibers. (Most of the communication systems in use today, including many specialized networks such as cable TV and mobile phone systems, have optical fiber in them somewhere; however, this does not make them optical networks.) As we shall see, suitable architectures for high-performance lightwave
- 47. 2 Multiwavelength Optical Networks networks involve complex combinations of both optical and electronic devices. Thus, as used here, the term optical or lightwave network does not necessarily imply a purely optical network, but it does imply something more than a set of fibers interconnecting electronic switches. As optical and photonic technology has advanced, applications to point-to-point trans- mission have preceded advances in networking. For example, it was clear in the early years of optical fiber transmission that by introducing wavelength division multiplexing (WDM) on existing fibers the capacity of a fiber link could be increased manyfold at minimum cost. However, it was only since the early 2000s that the optical switching technology necessary to convert isolated fiber transmission links to optical networks matured sufficiently to permit the commercial deployment of these networks. In the mid-1990s, the optical network (as opposed to optical fiber transmission alone) was still a “blue sky” concept. New optical and photonic devices were being developed and incorporated into experimental networks. But full-fledged multiwavelength networks integrating optical transmission, switching, and user access were still in the research and development stage. At that time the technology push for networking was out in front, but demand for the seemingly unlimited capacity of these networks was essentially nonexistent. As this is being written, the promise of optical networking is finally being fulfilled. The demand pull for these networks has materialized. As low-cost broadband services are made available to the general public, demand for Internet-based applications continues to increase. Equipment manufacturers, communications carriers, and service providers have joined in moving optical networking from feasibility studies to commer- cial viability in both cost and performance. The focus in the networking community has now shifted to organization, control, manageability, survivability, standardization, and cost-effectiveness, a trend that reflects the maturing of the optical technology as well as the recognition that the optical network is the only way of supporting current and future demand. These networks have played a critical role in reducing communications costs, promoting competition among carriers and service providers, and thereby increasing the demand for new services. In addition to the technology push and demand pull, a number of other recent devel- opments are contributing to the expansion and effectiveness of optical networks. One is the accelerating removal of the bottleneck in the “last mile” – the distribution network that is the bridge between the high-speed fiber core network and the end users. Until the last decade of the 20th century this distribution network – composed of twisted pairs of copper wires connecting each residential subscriber to the local telephone Central Office – was specifically engineered to a limited bandwidth of 3000 Hz. As a result the user bit rates were restricted to a tiny trickle. This low-speed access link separated the various high-speed communications and computing devices located on the premises of the end users (e.g., PCs, TV displays, and music/image/video storage equipment in the home) from the high-capacity network serving this equipment. Considering that the processors in today’s PCs operate at speeds six orders of magnitude faster than a low-speed access link, and the optical fibers in the network have bandwidths nine orders of magnitude wider than the bandwidth of the access link, it is obvious that access was – and is – a severe problem. As long as the last-mile bottleneck is present, the information
- 48. The Big Picture 3 superhighway is still a dirt road; more accurately, it is a set of isolated multilane high- ways with cow paths for entrance and exit ramps. The introduction of broadband access to residential customers by the telephone carriers and the cable operators is a step toward eliminating those cow paths. However, digital subscriber line (DSL) and cable modems are half-measures at best. Direct access to the fiber network by the end user [i.e., fiber to the home (FTTH) or business user] is the ultimate way of removing the bottleneck so that the network remains effective as demand for bandwidth grows. Although FTTH was deployed many years ago in a few demonstration projects, it did not take hold for several reasons, including cost and the absence of services of interest to the customers. Today that has changed because of the proliferation of broadband Internet services. De- ployment of glass is now moving from the network core through fiber access networks to the end users. This will undoubtedly stimulate interest in new broadband services that take advantage of high-speed access and in turn produce demand for more bandwidth. At this writing, most of the world’s installed fiber capacity is underutilized – arguably due to the last-mile bottleneck. That should change rapidly as progress in the removal of the bottleneck results in a quantum jump in network traffic, making high-performance optical networks indispensable. Higher level issues such as deregulation, new ideas for improving the economics of networking, and standardization of control and management techniques in multi- vendor networks are also contributing to the growing effectiveness of optical networks. Deregulation, which began in the United States in 1984 with the dismantling of ATT, has brought with it a new level of competition, with long-haul carriers, local carriers, Internet service providers (ISPs), and cable operators poaching on each other’s domains and using optical fiber capacity to do so. Bandwidth trading has been introduced as a way of improving the utilization of fibers and thereby optimizing profits. A carrier with idle capacity sells it to another carrier with excess demand. This type of exchange requires sophisticated control and management tools for network reconfiguration. More generally, any large network requires complex control and management systems and intelligent network elements for performance monitoring, network reconfiguration, and fault recovery. The systems, protocols, and equipment for performing these functions in traditional telephone and data networks were built over many years by the public carri- ers and equipment manufacturers. The new optical networks require similar tools, and this is especially important in multivendor environments. These are now making their appearance in the form of a proposed control plane for optical networks and protocols for systems management in these networks. As more sophisticated control and manage- ment functions are incorporated into optical networks the network operators are in a better position to offer high-quality service to their customers, improving the operator’s revenue stream and customer loyalty. Aboveall, thelessons of thepast show us that tomorrow’s networks must beflexibleand versatile enough to adapt to a continuing barrage of new and as-yet-unknown services. It is interesting to note that when optical networks were still in an embryonic form, the typical uses envisaged for them were high-tech applications such as high-resolution medical image archiving and remote supercomputer visualization – basically usages generated by a minuscule, elite segment of the population. Today these applications
- 49. 4 Multiwavelength Optical Networks represent but a tiny part of the global network traffic, submerged in a torrent generated by the common man, who has only recently gained access to the enormous opportunities our worldwide communication system has to offer. The networks we conceive today must be “futureproof” so as to be ready for the next unforeseen developments. 1.2 Objectives of an Optical Network Architecture Today’s and tomorrow’s optical networks must provide the capacity, connectivity, and intelligence necessary to link together a global community of information providers and consumers. A well-designed network performs this function efficiently and reliably. To facilitate a systematic study of networks that achieve this goal, it is useful to formulate a generic model in the form of a multiwavelength network architecture (MWNA). As background for the MWNA we briefly review the current network structures and the services they support. Until the end of the second millennium, the world of networking consisted of two separate spheres: the traditional telephone networks mainly devoted to providing voice services (operated in a circuit-switched mode) and data networks (operated in a packet-switched mode) for communication between computers. Each type of network was specially engineered and optimized for its own type of service. Circuit switching was the preferred approach to voice transmission, because the voice signal was transmitted as a continuous stream, whereas packet switching was invented to carry data traffic because data signals were bursty in nature, making the circuit-switching approach very ineffi- cient. Because the voice networks operated by the public carriers contained virtually all of the world’s installed communication capacity, the early data networks were con- structed as overlays on these networks, running on lines leased from the public carriers – mainly ATT in the United States and the government administrations in Europe. The traffic flow in the early data networks was minute compared to voice traffic – essentially confined to businesses, universities, and research laboratories. For this rea- son the main players in data networking were originally government, research, business, and educational organizations and data-processing equipment manufacturers. As optical fiber became the dominant transmission medium, various standards for ex- ploitation of fiber were developed, including the synchronous optical network (SONET) standard in the United States and a similar synchronous digital hierarchy (SDH) stan- dard in Europe. The SONET/SDH transmission, multiplexing and switching equipment, adapted primarily to circuit-switched applications, was soon augmented by asynchronous transfer mode (ATM) switches and Internet Protocol (IP) routers (cell-switched and packet-switched, respectively) to handle a wide variety of data and multimedia services. By the late 1990s the traditional separation of voice and data networks changed signifi- cantly. In a very short time we moved from a voice-centric world to a data-centric world, and, more importantly, the techniques of carrying data (packet switching) were extended to an infinite variety of services having no resemblance to those in the traditional com- puter world. Internet/Web services, running the gamut from interactive computer games through telemedicine to peer-to-peer file sharing, now use IP for transmitting anything
- 50. The Big Picture 5 from computer data to video (Internet Protocol TV; IPTV) to old-fashioned voice (Voice over IP; VoIP).1 This brings us to the characteristics and requirements of the services supported by the optical networks discussed here. These are extremely diverse in terms of connectivity, bandwidth, performance, survivability, cost, and a host of other features. Consider the common Internet services offered to the general public (e.g., e-mail and search engines). They serve a vast globally distributed user community. In terms of connectivity, these types of services push networking to its ultimate limits; any end user wants rapid connectivity to anyone or anything in the network. However, in terms of performance, they are undemanding – they can tolerate errors, delays, and occasional downtimes due to congestion, programming bugs, and equipment failure. Total costs may be high, but they are spread over an enormous user base resulting in a very low cost per user. In contrast, consider a different type of application, the virtual private network (VPN). This is a subnet carved out of a larger network by a telecommunications carrier and put at the disposal of a single enterprise, which typically controls and manages it. Consequently it has a much smaller user group with more intense utilization per user, far fewer active connections, and tighter control of network performance, including security and reliability. Customer costs per user will be higher, but this is offset by higher performance and more responsiveness to the needs of the customer. Another example is telemedicine, which requires high-quality communication (e.g., high fidelity medical image encoding and transmission, and rapid response) and where cost is secondary. Different requirements apply to public safety services (e.g., police, fire, and disaster relief), which depend on a high degree of survivability, fault recovery, and availability2 in the face of equipment and line failures, natural disasters, or malicious attacks. Transmission quality is secondary. Similar requirements hold to a lesser degree for financial services (e.g., banks and brokerage houses). In public safety and financial service applications, cost is not the primordial issue. To ensure satisfactory service, large users of network services (e.g., enterprises op- erating VPNs) enter into service-level agreements (SLAs) with the service providers. For example, the SLA might specify a level of availability, network delay, packet loss, and other features. These represent promises from the provider to the user, and as such they must be backed up by suitable controls within the underlying network to achieve the performance stated in the SLA. These controls are enforced within a large network by identifying differentiated services, that is, traffic flows that are singled out to be pro- vided with a predictable quality of service (QoS) (e.g., limits on packet loss and delay). Traffic routed through a large network can be tagged to recognize its class of service (CoS), thereby facilitating the satisfaction of service requirements through mechanisms such as priority packet queueing, bandwidth allocation, and service recovery priority. 1 The increased interest in VoIP, because of its low cost and growing ubiquity is, to paraphrase Shakespeare, the most unkindest cut of all from the computer community to the traditional telephone carriers. 2 Availability is the percentage of time that a network is operational. For example, “five 9s” (99.999%) availability, which is a goal for public carriers, implies 5.25 minutes of downtime per year.
- 51. Another Random Scribd Document with Unrelated Content
- 52. name, and address it to the Dowager Countess? I see they allow you paper, pens, and ink. With all my heart, Van Noost, replied Smeaton. I am quite sure you would rather injure yourself than me. And he wrote down on a sheet of paper the words which had been required. When he had sanded the paper and was handing it to Van Noost, a sound of bolts being drawn was heard at the door. The statuary hurriedly concealed what he had received, and the next moment Richard Newark came in. He advanced towards the Earl with a frank bright look, and shook him warmly by the hand. Then, turning to Van Noost, he said, Ha! idol-maker! Are you here? Get you gone--get you gone to Emmeline, and stay with her till I come. The dear gouvernante has gone forth questing like a spaniel dog upon a pheasant, from a hint I gave her last night. Do not leave her for a minute; and, if the man refuses you admittance, pull his nose boldly, and walk in. He is an arrant coward; so you may venture safely. I will--I will, sir, replied Van Noost. He shall not stop me on such an errand. If there be two of them, continued Richard, knock down one. That will be enough for the other. Van Noost hurriedly took up his hat and left the room; and Richard Newark, taking Smeaton's hand in his, said, in a quieter tone than usual, Come, Eskdale, sit down and talk to me. I must try and keep my poor whirling brain steady for a minute or two, while you tell me all and everything with regard to your transactions with Lord Stair. There is your only chance of safety. If you can show that you were
- 53. driven into the insurrection against your own inclination by the conduct of others, as I know you were, a skilful lawyer tells me that you will certainly be pardoned. New listen to what I know, then fill up the gaps, give me some proofs, and I will follow the scent as keenly as my bloodhound, Bellmouth. You sent a letter long before the outbreak to Lord Stair. That letter never reached him. It was stopped by my father. You went over to Mount Place, led to believe that you would see nobody but one old fool; and you found twenty or thirty, young and old, assembled, on a hint from my father, to meet you and trap you into treason. The Exeter people sent down dragoons, who sought you at Mount Place, and thence tracked you to Keanton; for they had secret information from Ale Manor. But what could be your father's motive? asked Smeaton. Keanton, for the first; to get you out of the way of Emmeline, for the second, answered Richard. But never mind motives. Let us deal with facts. You afterwards, in the north, sent your servant with a letter to Lord Stair, on receiving intelligence that he was on before us at Wooler. Now, Eskdale, I doubt that letter ever having been seen by him. Nay, I am quite sure it was not. Higham assured me, said the young Earl, that it was put into his hand, that he opened it, read it, and returned it with contempt. What can make you think that he never saw it? Because Lord Stair was, on that very day and hour, more than seven hundred miles from Wooler as the crow flies, replied Richard. His regiment was there, true enough; but he was in Paris. A man cannot be in two places at once, noble friend. But come, do not pause and wonder. This is all I know. Fill up, fill up! Let me hear the whole; and I will try if my wits are not worth something, in spite of all folks may say against them. Smeaton did as he was bidden; and, sitting down at the table with his young companion, he gave him a clear and complete
- 54. narrative of everything that had occurred after his arrival at Ale Manor, and showed him the copies he had taken of his letters to Lord Stair. More than once Richard asked him to stop for a moment, and wrote down the heads of what he had heard; and then, looking at the letters, he said-- May I take these with me to copy? You shall have them to- morrow; for you may need them. Strange that a piece of paper should sometimes be the best armour for a man's neck! Take them, take them, replied Smeaton. They are but unauthenticated copies, and could not be given in evidence, if Lord Stair has not received them. Yet I can hardly believe that Higham would play me such a trick. Where did you hire him? asked Richard. He was recommended to me by the man in whose house I lodged, replied the young Earl; a good honest fellow, who had been a servant to the Earl of Oxford. Put about you by the Jacobites, replied Richard, with a laugh, to keep you steady in the cause, and commit you to it if you wavered. The man must be found and made to tell the truth. Hear you will have to seek him in the grave, said Smeaton; for he was sorely wounded at Preston, where he fought as boldly as a lion. Never mind, replied Richard. Some of these letters must have reached Lord Stair, I think; and, if I get at him, I will jump upon his back, and never take my spurs from his side till we have passed the winning-post. Good-bye, Eskdale, good-bye. Your trial will not come on for a month, they say; and you wont see me for a fortnight, perhaps; but I'll be working all the time. Tell Emmeline to mind well every step she takes; for the villain scoundrel, William Newark, alias Somerville, has made his peace with the court, pretends that he is
- 55. the most loyal subject of King George, has betrayed all that he knew of Kenmure's and Forster's secrets, and is watching with all his eyes to pounce upon Emmeline. He cannot rightly make out where she is; for I have puzzled him about it. But he thinks that if he could but get her into his hands, Ale Manor--which is hers, you know--would be his, and he would be a great man in his generation. Once more, good-bye, Eskdale; and, if you hear that I am drowned, shot, stabbed, or otherwise disposed of, do not forget me. Say to yourself- -'I was kind to the boy; and he loved me well.' Thus speaking, he hurried to the door, and halloo'd to the turnkey to let him out. CHAPTER XLII. I will not dwell upon the first interview between Emmeline and her husband; I will not dwell upon many that took place, for many did take place between the time of his arrival as a prisoner in London and the day of his trial. There are sanctities in the deep emotions of the heart, the violation of which nothing but a holy cause can justify. I have no right to eat the show-bread on the altar of their love. I have no right, be they real or be they ideal characters, to intrude into the secrets of their hearts, and place the thrilling nerves beneath a microscope for the public eye. Suffice it to say that they met often, daily, sometimes twice a-day, by the skilful management of her who had been the young Earl's nurse; and that no annoyance or inconvenience happened to the young Countess of Eskdale during nearly a month, although some circumstances of suspicion--a number of strange men hovering about the house, and
- 56. the appearance of others dogging them in their walk to the Tower-- caused some apprehension in the mind of the old housekeeper, and induced her to redouble her precautions. Emmeline had seen her cousin more than once. Kind, affectionate, self-devoted, he showed himself during their short and scanty interviews; but those interviews were not very many. Suddenly he disappeared, telling his fair cousin that he was about to visit Paris, but without mentioning the business on which he went; for, although he was very sanguine in all things, he loved her too well to give her hopes which might be disappointed, or to shackle her exertions in other directions by expectations from the uncertain projects he had in view. She knew that he went for the purposes of her husband's defence, and she thanked him with her whole heart; but this was all she knew, and, when he was gone, she felt anxious and eager for tidings which did not come. Thus passed the days of a long imprisonment; but several steps had been gained, notwithstanding. The extreme laxity of those who had charge of the prisoners had become apparent, and Smeaton had established a certain sort of friendship with his jailers; but the principal fact was that they showed themselves accessible to bribes; so that the probability of escape was reasonably added to the probability of acquittal or of pardon. Nevertheless, with hope for their guide, they flattered themselves that the delay in bringing the prisoners to trial arose from the intention of sparing them; but they experienced a bitter disappointment in the end, when Smeaton and the rest were impeached of high treason by the House of Commons, and their trial came on with unusual rapidity. As is well known, the greater part of the insurgent noblemen pleaded guilty. But Smeaton would not join in this plea. He acknowledged the whole share he had borne in the rebellion; he entered into minute details of all that had occurred; he showed, as well as he had the means of showing, that he was actually driven to join the insurgents; but he could bring no proof of the fact. Richard
- 57. was still absent, although he had promised to return in a fortnight, and nothing had been heard of him when the trial took place. Smeaton's mere unsupported word had little weight with the peers; but, while most of the others were, upon their own plea, condemned at once, a space of time was taken to consider and to allow for the collection of evidence before his trial. The lawyers laboured hard to induce him to withdraw his plea of not guilty, and cast himself upon the royal mercy; but, although his mind, till the insurrection had actually begun, had been in that doubtful and undecided state which is most painful to men of a determined and resolute character, yet, once having joined in it, either the prejudices of early education resumed their sway, or the enthusiasm of his companions infected his own mind, and he could not bring himself to believe that there was guilt in supporting by arms the sovereign whom all his family had served, and whose claim to the throne of England they had never on any occasion renounced. He did not feel himself guilty, and he would not plead guilty. It was a dishonouring word, a word that he would not have attached to any part of his conduct by his own act, and he resolutely adhered to his former plea. He gave no unnecessary trouble indeed; he admitted all the facts as they stood charged against him; but he contended that his acts were loyal and not treasonable, and it was only as an admission that he stated he had been willing to submit quietly to the existing state of things. To this, he added a detail of the transactions between himself and the Earl of Stair. His defence was frequently interrupted; for the English law often decrees that the evidence which would clearly exculpate any man from all moral blame shall not be received in his justification. But he persevered in his course, and the very men who condemned him felt for him, and hardly believed their own words when they pronounced him guilty. It is a strange thing, that law of treason, which affixes the most odious moral censure upon acts heroically mistaken and sometimes
- 58. sublimely just; which compels men, by rigid rules and the admission of false premises, to pronounce that to be guilt which they know to be virtue; which places the same stain upon the lowest and most selfish crimes, and upon the most elevated patriotic deeds. A great fault exists somewhere; it is true, order and respect for law must be maintained; the will of the majority must rule; it may be, even, that, for general security, men must be punished for bold attacks upon existing institutions; but let us not be called upon to denounce as guilt that which is mistake, or enthusiasm, or virtue. The dark scene was over; the verdict was given, the sentence pronounced, the blade of the axe turned towards the prisoner, and one more of the gallant and the true was carried back from the bar to the Tower, to await the fate of a traitor. In the anticipation of that moment, Smeaton had often felt how terrible it would be; he had doubted his own courage, his own fortitude; he had nerved his mind to resist all the impulses of his mortal nature, lest he should meanly and faint-heartedly supplicate for life, as others had done. He recollected that there were many endearing ties around him; that youth, and love, and hope, and high health, and all the bright amenities of being, attached him to the world in which he was; that it was full of delight and enjoyment to one so constituted mentally and bodily, and that the thought of parting with it in its hour of greatest excellence might well shake his resolution and undermine his firmness. But when each peer had pronounced his judgment, and when the frightful and barbarous sentence was passed, it was marvellous, even to his own mind, how calmly he bore himself, how firm and composed he felt. It seemed for the moment as if the tremulous, vibrating, anxious cord between hope and fear was snapped, and that his feet were firmly fixed upon the rock of fate. Take away hope, and there is no such thing as fear. During a short space of time all hope was over in his bosom. But, in the meanwhile, others were preparing hope for him, and to two separate scenes we must turn, where busy love was eagerly exerting
- 59. itself, in different ways and without concert, to avert the blow from his head. I know not which to depict first; for they both occurred on the same day, and very nearly at the same hour; but perhaps I had better choose the one which, from presenting few if any characters already brought under notice, may have the least interest for the reader. Into a gorgeous room of a palace, containing a number of distinguished persons--some marked out to the eye by the splendour of their apparel, some by their beauty or their grace--entered a middle-aged man, small in stature, insignificant in appearance, and with his somewhat large head rendered more ridiculously conspicuous by a huge Ramillies wig. He was dressed in tea- coloured velvet, with his sword by his side and his hat on, and the door by which he entered was thrown open for him by one of the high noblemen of the Court; while another, bearing a light in either hand, walked backwards into the room before him. He was a very mean-looking person; cold, unloveable in aspect, looking like a small dancing-master in a holiday suit; but yet he was a King. At one side of the room, supporting herself by the back of a chair, stood a tall and queenly woman of some sixty years of age. Her natural hair, as white as snow, appeared slightly from beneath the weeds of widowhood, and her striking and beautiful face--beautiful even in sorrow--was pale and worn with long and heavy sickness. The moment the king entered, she advanced towards him, with a step firm and dignified; but she sank upon her knees as she came near, and stretched out her hands towards him, holding what appeared to be a petition. Who are you, madam, who are you? asked the King, in French. I am the unhappy Countess of Eskdale, sire, replied the lady, in the same language. I do beseech you, hear me, and receive my petition for my poor son. Spare him, gracious monarch--spare him, and I pledge--
- 60. She was not permitted to finish the sentence. The cold-hearted King drew back at her first words, and, with a sort of frightened and repulsive look, turned towards a different door from that by which he had entered. But the lady caught him by the skirt of his coat, pleading with all the earnestness of maternal love for her son's life, while he rudely endeavoured to shake himself free, walking with a quick step towards the other side of the room, and literally dragging her after him as she still kept her hold, endeavouring to force the petition upon him. A gentleman with a cut upon his brow, who had entered with the monarch, now whispered in his ear in French: Be firm, sire! Be firm! Shall I remove her? The monarch made an eager motion of assent, and the other, casting his arms round Lady Eskdale, tore her away. The paper, which she held in her hand, dropped to the ground; and, instantly rising to her full height, as the monarch passed the door, she turned a look of dignified anger on him who had interposed to prevent the reception of her petition, and exclaimed aloud, in English-- Oh, William Newark, William Newark! Ever ready, like the viper, to sting the hand that has fostered you, and to aid in all that is hard and selfish! Poor lady! said the gentleman thus addressed, with a look of contemptuous pity, and he followed the King. But there was another who followed also; a grave-looking man of the middle age, with a calm and placid countenance and a blue ribbon across his breast. With a quick but easy step, he hurried on, and overtook King George just as he had crossed an ante-room and was about to enter a large drawing-room beyond--round which were grouped a great number of brilliant-looking people in a blaze of light. He ventured to stop the sovereign in his advance, saying something to him in a very low tone in the Latin language; for many of the first nobility of England, at
- 61. that period, did not speak French or German, and the first George's stock of English was not very copious. Who is he--who is he? asked the monarch, also speaking Latin, though not in its greatest purity. What does he want at this hour? He bears despatches from Lord Stair, sire, the nobleman answered who had spoken to him; and is charged to deliver them immediately into your Majesty's own hands. He is the young gentleman whom your Majesty declared to be more praiseworthy, on account of his speedy repentance and atonement, than others who had never joined the rebellion. He spoke still in a low tone; but the monarch replied, aloud, Admit him--admit him. He is a strange boy; but whatever comes from my Lord Stair is worthy of immediate attention. The despatches were to be delivered in private, sire, observed the other; but the bearer was detained for want of horses on the Dover road. Shall I-- So be it, so be it, replied the King. Close the doors again. Make everybody quit the room but you and Walpole, my lord; and then bring the young man in. The personage to whom he spoke proceeded to fulfil his commands, and William Newark, in obedience to those commands, quitted the room with a scowling brow, which was not brightened by the passing of Richard Newark in the very doorway. He did not venture to say anything, however, and the lad advanced with a small packet in his hand straight towards the King, without any other salutation than merely a low bow. Bend your knee, bend your knee, said the elderly nobleman, in a whisper, and the lad, after a moment's hesitation, did as he was directed.
- 62. I am glad to see you again, young gentleman, said King George. You have been to Paris, I suppose. And, at the same time, he took the packet and broke it open. It contained two sheets; but, before he proceeded to examine either of them, the monarch added a question. Do you know, he asked, why Lord Stair happened to address me personally instead of the Secretary? Because the matter was for your Majesty's own ear, replied Richard Newark, somewhat abruptly. We do not give an apple to one boy to hand it to another, for fear he should eat it himself. The King laughed good-humouredly, and proceeded to read the first sheet, which, beginning at the bottom of the first page, and ending at the top of the fourth page, did not seem to contain much matter. Whatever that matter was, it seemed to give the King great satisfaction. That is good; that is very good, he said. He is an invaluable man. We shall know how to honour him. All is safe in that quarter. He then turned to the other sheet, and his face instantly changed. Ha! he said, with a curling lip, and an irritable eye More about this Lord Eskdale! He joined the rebels wittingly, adhered to them till the last moment, was taken with arms in his hands, and he must die. I have signed the warrant. Then kill me first, sir, rejoined Richard Newark, bluffly, for I first helped to engage him in the rebellion; and, had it not been for his advice, I should never have quitted it. He went against his own will, as your Majesty will see if you read; and, if he dies, it will be as a bird that is caught in a trap because he was deceived by the baits set for him. Your Majesty cannot understand till you read, any more than I can see through that wall; for there is a great deal beyond your sight or mine, unless a door be opened for each of us to look through.
- 63. The King gazed at him for a moment in utter surprise, as if completely astounded by the lad's impudence; but gradually a sense of the justice of what he had heard seemed to overpower the slight sense of anger; and, without answering a syllable, he turned his eyes to the paper, and proceeded to read it to the very end. When he had done so, the expression of his countenance was again greatly changed; a hesitating and embarrassed look came upon his face. He put his finger under his large wig, rubbed his temple, and pulled up one of his stockings, which had somewhat slipped down the leg, and most likely tickled his shin; then, turning to another gentleman present, he said, Come with me, Mr. Walpole--come with me, my Lord. I will go to my cabinet for a moment. Thus saying, he took two steps towards the door by which he had entered, but then turned a sharp glance upon Richard Newark, who was standing by with a vacant air, looking down at the hilt of his sword. It was the same sword which Smeaton had given to him. The monarch's look was certainly not very placable at first; but something seemed to touch the risible organs in his brain or heart-- wherever they may lie--and we all know that in those organs a great deal of the milk of human kindness is secreted. He laughed, low but gaily, and said-- Get away, sir, get away. Lord Stair has trusted his letters to a somewhat indiscreet messenger. The best in the world could not have done better, your Majesty, replied Richard Newark, boldly; for he has delivered them safely into the best hands in the realm. If he meant it, nothing could have been more dexterous than his reply. It was a compliment, slightly veiled under a rudeness. But I very much doubt whether he did mean it. However, King George smiled most graciously, saying: Go, sir, go. We shall not forget you.
- 64. Richard Newark bowed and retired, while the King again took a step or two towards the door. Before he passed out of the room, however, the King turned to a gentleman with a florid countenance, saying: We shall not meet that woman again, I hope; for I have not quite made up my mind. Keep that man, Sir William Newark, from me. I do not like him as I did. So saying, and suffering Mr. Walpole and one of his attendants to precede him, he followed slowly and thoughtfully out of the room. The adjoining chamber was by this time vacant; the unhappy Lady Eskdale had quitted it the moment after she had received so violent a rebuff, and the courtiers who had been present when she sought to force her petition upon the King, concluding that he had passed on into the drawing-room, had thronged thither by another way. But a full hour elapsed before the monarch joined his guests. Now let us turn to the other scene which I have mentioned, in which strong affection was busily engaged for Smeaton's deliverance, but in a different manner. Let us break into the middle of it, however; for what is to follow will explain what is passed. No, no, dearest lady, said old Mrs. Culpepper, in a low but eager tone. It must not be. The boat is prepared, the ship ready to sail the moment his foot is on board. You must go with him, and all will be safe. Then who is to stay and personate him in the prison? asked Emmeline. Indeed it must be as I have said. Although you have bribed the people to shut their eyes, yet I do not believe they dare venture to let three people pass out when only two have passed in. In this I will have my way, indeed. I fear nothing. I do not believe there is any man so cruel as to punish a wife for saving her husband's life, I will wrap myself in his roquelaure, and sit brooding
- 65. over the fire. My heart may beat; but no one will see it. My eyes may overflow; but I will cover them with my hands. The first plan was the best--far the best, and it is my bounden duty, as well as my earnest wish, to risk anything to myself for his sake. Oh, Heaven what happiness will it be hereafter, even if they should shut me in a prison and never let me see his face again, to think that I have saved him! It is the same plan still, dear lady, replied Mrs. Culpepper, with her usual calm and quiet manner; but you must not, cannot execute it in the way you propose. Consider your height, the difference between your tiny figure and his. They would be blind indeed to mistake you, and we cannot expect them to be so blind as that. I am shorter than he is, but still I am very tall, and the difference will not easily be seen. They will not mark very exactly, especially if he put his handkerchief to his face and seem to weep. My clothes will nearly fit him too; and-- And will you--will you stay in his place? asked Emmeline, gazing in her face, with a look of wonder and gratitude. What will you say when they find you there? You have no such excuse as I have. I will say, lady, replied the woman, earnestly, that he drank the milk from this breast as an infant; that he was to me as a child, when God had taken my own; that he was my nursling, my beloved, my only one, when I had lost all else, on earth who loved me, or whom. I could love. Then, if they choose to shorten my days or make me pass them in a prison, it is but little they can take away and little they can inflict. It must be so indeed, Lady; and now we are only losing time. They will not let us pass in or out after eleven. It is now past nine, and it will take some time to disguise him as we wish. Haste then, to get on your hood. I am quite ready. With this sacque above my other clothes, and a large French capote, everything is ready to hide his face and figure.
- 66. Emmeline looked down thoughtfully; but she said nothing, for her heart was too full to speak, and in a few minutes they set out upon their adventure, followed by two men servants, whom the old housekeeper had already prepared for the task in hand. The moment they were gone, however, one of Sir John Newark's men, who had lived at Ale for several years, and who had been accustomed to act as one of his spies upon all that took place in the house, crept silently out and pursued them with a stealthy step down the little street. He saw them cross Tower Hill, and obtain admission at the gates; and then, turning to the right, he approached a house in a neighbouring street, hurrying his walk as much as he could without converting it into a run. At the moment he reached the door, one of the ordinary hackney coaches of the day drew up, and a gentleman in somewhat brilliant attire descended with a slow step. The man waited till he had paid the fare, and then plucked him by the sleeve, whispering something in his ear. The gloomy and discontented face of the other instantly cleared up, and he exclaimed, with a mocking laugh--Ha, ha! Then they have put themselves in the trap. I will away to the Tower. You stay and watch at the gates. But no--better let them be caught in the very act, just when they fancy themselves secure. It will be more meritorious to bring him back after he has actually escaped than to prevent him from doing so. You are sure, quite sure? It would never do to take an old raven instead of a young hawk. I am quite sure, replied the man; for I overheard it all, as I listened at the hole I have made in the wall. This morning, I could not make out which of the two it is who is to play his part; but just now I heard, and I am quite certain. The old woman was his nurse, it seems, and is ready to sacrifice her life for him. Well, well, go to the gates and watch, rejoined William Newark. Give instant information if they come forth. I will go and get a messenger. There is one lives hard by.
- 67. The servant did as the other bade him; but he had not remained many minutes near the gates of the Tower when some quick steps approached, and he turned round towards the new corners. Ha, ha, old Truepenny! said Richard Newark, taking the man's arm in a firm grasp; what are you on the watch for here? Nothing, Master Richard, answered the man. I am only just taking the air. You won't let your intentions take the air, at all events, retorted Richard Newark. I know you, serviceable knave! This is the fellow, he continued, turning to the two young men who accompanied him, this is the fellow who informed of the smuggled tea. Then I will baste him to a stock fish, cried one of the youths, brandishing his cudgel. No, no, interposed Richard, with a laugh. Wait till you get him back at Ale, and then tar and feather him. Hasten off, Argus, or we will leave you no eyes to see out of. The man had no hesitation in obeying; and, as soon as the young gentleman had relaxed his grasp, ran across the open space as fast as his legs would carry him. Richard Newark then turned towards the gates again; but, taking three steps in advance, paused, and, after a moment's thought, with his hand pressed upon his brow, quietly glided away to a little distance, followed by the two lads. CHAPTER XLIII.
- 68. At the hour of half past ten, two persons issued forth from the room in the Tower in which the young Earl of Eskdale had been long confined. Both were dressed in female apparel; both were apparently much affected, and it appeared very natural that they should be so, as the following morning was appointed for the bloody spectacle of an execution on Tower Hill. The limbs of the younger and shorter lady trembled so much that they could hardly bear her up; but the other, though apparently weeping and holding a handkerchief to her eyes, seemed much more firm, and contrived to support the wavering steps of her companion as they passed out into the passage. The jailer who opened the door to give them exit from the room looked in and saw a tall figure wrapped in a red cloak laced with gold, seated by the fire, with the head leaning on the hand. All is right, he cried, speaking to another man at the top of the stairs hard by. Pass them out! Hastening onward through the passages and courts of the Tower, as fast as the agitation of the fair girl would permit, they came without obstruction to the outer gate, where the two men servants were waiting in the little gate-house. The turnkey who accompanied them seemed to be a kind-hearted man for one in such an office; and, while the wicket was being opened, he said--Don't take it so much to heart, lady. Perhaps he may be pardoned after all. One of the tall warders who stood near gave him a grim contemptuous look, and uttered a short cruel laugh; but the two visitors, without reply, passed unopposed through the wicket, and stood upon Tower Hill. The men servants followed, and the gate was closed. Still keeping profoundly silent, they all walked on with great speed, not towards the little street in which Emmeline had lived, but
- 69. towards the end of another street. When they were half way across the open space, the latter of the two bent down, saying in a whisper--Bear up, bear up, dear Emmeline. We are well nigh safe now. But hardly were the words uttered when two or three men came quickly across, and one of them taught hold of the apparently elder woman's arm, exclaiming, with a mocking laugh--You are a tall lady, upon my soul, to walk upon Tower Hill of a night! Gadzooks, we must see more of your ladyship! Another man--who subsequently turned out to be a messenger sent in pursuit--at the same moment seized the young Earl (for I need hardly say it was he) with a hard strong grasp, exclaiming-- Henry, Earl of Eskdale, I charge you, in the King's name, to make no resistance. With a faint despairing cry Emmeline sank to the ground, while they dragged Smeaton away from her side. The two servants, running up, demanded--Who are you who dare to stop these ladies? and angry words began to pass; but Smeaton interposed, saying--It is in vain, it is in vain. Look to your lady, my good men. Convey her home safely. God bless you, my Emmeline! What is the matter, what is the matter here? cried Richard Newark, suddenly appearing with two or three more, while the man who had first seized upon Smeaton left him in the hands of the messenger, and raised Emmeline from the ground. Ah, Master Dick! he exclaimed, have you a finger in this pretty pie? Better put yourself out of harm's way, young man, as fast as possible. How dare you touch that lady, scoundrel? demanded Richard, in a voice furious with passion, as he recognised the person of William Newark. Take that for your pains! And, holding the scabbard of his
- 70. sword with his left hand, he struck his cousin a furious blow with the right. William Newark started forward and drew his sword; Richard's was not long in the sheath; but the servants interposed, and parted them for the time, though not till words had been spoken--some in loud anger, some in the low tones of intense hate--which bore their fruit soon after. The last four of those words were uttered in a whisper. At seven, and alone, said Richard, in his cousin's ear. The other nodded his head, and turned sullenly away, while Richard aided to raise the unhappy girl, whose last hope had been extinguished by her husband's recapture, and carried her, still insensible, to her dwelling. In the mean time, the messenger and two of his men conducted their prisoner back to the gates of the Tower with feelings in the bosom of Smeaton too dark, too painful, for description. To his own fate his mind had been long made up, and the extinction of a brief hope of escape added little to the load he had to bear; but the thought of what might befall Emmeline in consequence of her effort to save him, and of the certain consequences to the devoted woman who had placed her liberty and even her life in peril for him, was too heavy to be borne with anything like calmness. Arrived at the gates of the Tower, they found the wicket, to their surprise, open, and a good deal of confusion under the archway of the gate-house. Some twelve or fourteen men were collected; a buzz of tongues was going on; and some loud and angry words were being spoken. The lieutenant-governor himself, in a silk dressing- gown, was present, with a man beside him, holding a lantern; and just as the messenger passed the wicket, still holding the prisoner fast by the arm, they heard that officer exclaim--
- 71. Shut the gate, shut the gate! Every one keep silence! If you can be discreet, no harm may come of this. If not, some of your necks may pay for it.--Ha! who have we here? An escaped prisoner, Mr. Lieutenant, answered the messenger, who was willing to take all possible credit to himself. I am sharp enough; and I got information of this fine plot. The lieutenant-governor stared at him coldly, with no great appearance of satisfaction in his countenance. Pray, Mr. Messenger, he said, after a moment's thought, had you any warrant for what you have done? The man looked aghast at the question, but replied, in a somewhat insolent tone-- I needed no warrant to apprehend a convicted traitor whom you have suffered one way or another to slip out of the Tower. The lieutenant still gazed at him with a frowning brow and teeth tight shut, and then said-- You may have to prove, Mr. Messenger, that you possess such a justification of your conduct. I tell you, you have not. Then, turning to one of the warders, he said, in a sharp tone-- Shut the wicket, I say, and lock it. Let no one pass in or out till I return. Keep that man safe too, he continued, pointing to the messenger, and be perfectly silent with him. Let no one exchange a word with him, as you value the King's favour.--My Lord of Eskdale, will you do me the honour of accompanying me back to your chamber? I wish to speak a few words with you.--Let go his arm, sir, this instant!
- 72. The messenger instantly relaxed his grasp; and Smeaton, not less astonished than his captor, followed the lieutenant in silence back to the room where he had been confined. They found the door open; but within stood the turnkey, looking gloomy enough, with his arms crossed upon his chest, and old Mrs. Culpepper, with the young Lord's roquelaure now cast off, seated in her usual attire before the fire. The moment she heard steps, however, she started up, and, gazing at Smeaton, clasped her hands together in silence, with a look of unutterable anguish. Remove her to my lodging, said the lieutenant, speaking to the turnkey, and keep her there under your guard till I come. The young Earl, however, started forward, and took her by the hand. Thanks, excellent woman! he exclaimed, a thousand thanks! I pray God, as one of my last prayers, that he may defend you and my Emmeline, and shield you from all the ill consequences of this night. Before she could reply--for her voice was choked with sobs--she was removed from the room; and the lieutenant, carefully closing the door, said, with a faint and rueful smile-- That dress does not become you, my Lord. Let me beg you to throw it off, for I hardly know whether I am speaking to the Earl of Eskdale or an old woman. That is easily done, replied Smeaton, casting off the loose garment called a sacque, which was, for three-quarters of a century, a favourite habiliment of the ladies of France and England. Now, sir, I am your prisoner again. I beseech you to leave me, for the last few hours of my life, to the thoughts which befit the occasion; and, if it be possible, to conceal the events which have taken place, so as to shield that excellent creature and all others from the consequences.
- 73. Welcome to our website – the ideal destination for book lovers and knowledge seekers. With a mission to inspire endlessly, we offer a vast collection of books, ranging from classic literary works to specialized publications, self-development books, and children's literature. Each book is a new journey of discovery, expanding knowledge and enriching the soul of the reade Our website is not just a platform for buying books, but a bridge connecting readers to the timeless values of culture and wisdom. With an elegant, user-friendly interface and an intelligent search system, we are committed to providing a quick and convenient shopping experience. Additionally, our special promotions and home delivery services ensure that you save time and fully enjoy the joy of reading. Let us accompany you on the journey of exploring knowledge and personal growth! ebookultra.com