| Preface | p. xiii |
| Acknowledgments | p. xxi |
| About the authors | p. xxiii |
| QOS Requirements and Service Level Agreements | p. 1 |
| Introduction | p. 1 |
| SLA Metrics | p. 4 |
| Network Delay | p. 4 |
| Propagation Delay | p. 5 |
| Switching Delay | p. 6 |
| Scheduling Delay | p. 6 |
| Serialization Delay | p. 6 |
| Delay-jitter | p. 8 |
| Packet Loss | p. 9 |
| Bandwidth and Throughput | p. 12 |
| Layer 2 Overheads | p. 13 |
| VPN Hose and Pipe Models | p. 16 |
| Per Flow Sequence Preservation | p. 18 |
| Availability | p. 20 |
| Network Availability | p. 20 |
| Service Availability | p. 21 |
| Quality of Experience | p. 22 |
| Voice | p. 23 |
| Video | p. 24 |
| On-line Gaming | p. 24 |
| Application SLA Requirements | p. 24 |
| Voice over IP | p. 26 |
| VoIP: Impact of Delay | p. 29 |
| VoIP: Impact of Delay-jitter | p. 31 |
| VoIP: Impact of Loss | p. 33 |
| VoIP: Impact of Throughput | p. 36 |
| VoIP: Impact of Packet Re-ordering | p. 37 |
| Video | p. 38 |
| Video Streaming | p. 38 |
| Video Conferencing | p. 57 |
| Data Applications | p. 58 |
| Throughput Focussed TCP Applications | p. 59 |
| Interactive Data Applications | p. 70 |
| On-line Gaming | p. 74 |
| Marketed SLAs versus Engineered SLAs | p. 16 |
| End-to-End SLAs vs Segmented SLAs | p. 77 |
| Inter-provider SLAs | p. 77 |
| Intserv and Diffserv SLAs | p. 78 |
| References | p. 79 |
| Introduction to QOS Mechanics and Architectures | p. 87 |
| What is Quality of Service? | p. 87 |
| Quality of Service vs Class of Service or Type of Service? | p. 88 |
| Best-effort Service | p. 89 |
| The Timeframes that Matter for QOS | p. 90 |
| Why IP QOS? | p. 91 |
| The QOS Toolset | p. 91 |
| Data Plane QOS Mechanisms | p. 94 |
| Classification | p. 94 |
| Implicit Classification | p. 95 |
| Complex Classification | p. 95 |
| Deep Packet Inspection/Stateful Inspection | p. 96 |
| Simple Classification | p. 96 |
| Marking | p. 99 |
| Policing and Metering | p. 100 |
| RFC 2697: Single Rate Three Color Marker | p. 102 |
| RFC 2698: Two Rate Three Color Marker | p. 106 |
| Color-aware Policers | p. 108 |
| Metering | p. 111 |
| Queuing, Scheduling, Shaping, and Dropping | p. 112 |
| Queuing and Scheduling | p. 112 |
| Dropping | p. 128 |
| Shaping | p. 137 |
| Link Fragmentation and Interleaving | p. 140 |
| IP QOS Architectures | p. 141 |
| A Short History of IP Quality of Service | p. 141 |
| Type of Service/IP Precedence | p. 142 |
| IP Precedence | p. 144 |
| Type of Service | p. 145 |
| IPv6 Traffic Class Octet | p. 147 |
| Integrated Services Architecture | p. 147 |
| Differentiated Services Architecture | p. 147 |
| DS Field | p. 150 |
| Per-Hop Behaviors | p. 154 |
| Per-Domain Behaviors | p. 159 |
| Explicit Congestion Notification | p. 160 |
| Diffserv Tunneling Models | p. 165 |
| IPv6 QOS Architectures | p. 170 |
| MPLS QOS Architectures | p. 171 |
| MPLS and Intserv/RSVP | p. 172 |
| MPLS and Diffserv | p. 173 |
| IP Multicast and QOS | p. 181 |
| Typical Router QOS Implementations in Practice | p. 183 |
| Layer 2 QOS | p. 189 |
| ATM | p. 190 |
| Mapping Diffserv to ATM QOS | p. 193 |
| Frame-relay | p. 194 |
| Ethernet | p. 196 |
| Complementary Technologies | p. 197 |
| Where QOS cannot make a difference | p. 198 |
| References | p. 199 |
| Precedence, TOS, and DSCP Conversion | p. 204 |
| Notation | p. 204 |
| Conversion | p. 205 |
| Deploying Diffserv | p. 209 |
| Introduction | p. 209 |
| Deploying Diffserv at the Network Edge | p. 211 |
| Why is the Edge Key for Tight SLA Services? | p. 211 |
| Edge Diffserv Case Study | p. 212 |
| SLA Specification | p. 212 |
| Diffserv Meta-Language | p. 218 |
| High-speed Edge Design | p. 218 |
| Design Variations | p. 225 |
| Edge SLA Summary | p. 241 |
| How Many Classes are Enough? | p. 241 |
| What Marking Scheme to Use? | p. 244 |
| VoIP - How Much is Enough at the Edge? | p. 245 |
| Deploying Diffserv in the Network Backbone | p. 249 |
| Is Diffserv Needed in the Backbone? | p. 249 |
| Core Case Study | p. 253 |
| Core Classes of Service and SLA Specification | p. 253 |
| "Prioritized" Diffserv Core Model | p. 254 |
| Detailed Core Design | p. 256 |
| Design Variations | p. 261 |
| Core-marking Scheme | p. 263 |
| Tuning (W)RED | p. 268 |
| Tuning the Exponential Weighting Constant | p. 269 |
| Tuning Minth and Maxth | p. 270 |
| Mark Probability Denominator | p. 271 |
| In- and Out-of-contract | p. 271 |
| References | p. 272 |
| Capacity Admission Control | p. 275 |
| Introduction | p. 275 |
| When is Admission Control Needed? | p. 277 |
| A Taxonomy for Admission Control | p. 282 |
| What Information is Needed for Admission Control? | p. 285 |
| Parameterized or Measurements-based Algorithms | p. 286 |
| Parameterized Algorithms | p. 286 |
| Measurement-based Algorithms | p. 288 |
| Topology-unaware Off-path CAC | p. 290 |
| Topology-aware Off-path CAC: "Bandwidth Manager" | p. 292 |
| Example Bandwidth Manager Method of Operation: Next Generation Network Voice CAC | p. 294 |
| The Integrated Services Architecture/RSVP | p. 303 |
| RSVP | p. 304 |
| RSVP Example Reservation Setup | p. 307 |
| Application Signaling Interaction | p. 314 |
| Intserv over Diffserv | p. 316 |
| RSVP Aggregation | p. 320 |
| RSVP Traffic Engineering | p. 325 |
| NSIS | p. 326 |
| End-system Measurement-based Admission Control | p. 328 |
| Summary | p. 329 |
| References | p. 330 |
| SLA and Network Monitoring | p. 335 |
| Introduction | p. 335 |
| Passive Network Monitoring | p. 336 |
| How Often to Poll? | p. 337 |
| Per-link Statistics | p. 337 |
| Monitoring Classification | p. 338 |
| Monitoring Policing | p. 339 |
| Monitoring Queuing and Dropping | p. 342 |
| System Monitoring | p. 346 |
| Core Traffic Matrix | p. 347 |
| Active Network Monitoring | p. 348 |
| Test Stream Parameters | p. 349 |
| Packet Size | p. 350 |
| Sampling Strategy | p. 351 |
| Test Rate | p. 354 |
| Test Duration and Frequency | p. 355 |
| Protocols, Ports, and Applications | p. 357 |
| Active Measurement Metrics | p. 358 |
| Delay | p. 358 |
| Delay-jitter | p. 360 |
| Packet Loss | p. 362 |
| Bandwidth and Throughput | p. 363 |
| Re-ordering | p. 363 |
| Availability | p. 363 |
| Quality of Experience | p. 364 |
| Deployment Considerations | p. 364 |
| External versus Embedded Agents | p. 364 |
| Active Monitoring Topologies | p. 365 |
| Measuring Equal Cost Multiple Paths | p. 369 |
| Clock Synchronization | p. 370 |
| References | p. 371 |
| Core Capacity Planning and Traffic Engineering | p. 375 |
| Core Network Capacity Planning | p. 375 |
| Capacity Planning Methodology | p. 376 |
| Collecting the Traffic Demand Matrices | p. 377 |
| Determine Appropriate Over-provisioning Factors | p. 382 |
| Simulation and Analysis | p. 388 |
| IP Traffic Engineering | p. 389 |
| The Problem | p. 390 |
| IGP Metric-based Traffic Engineering | p. 394 |
| MPLS Traffic Engineering | p. 397 |
| MPLS TE Example Tunnel Establishment | p. 397 |
| Diffserv-aware MPLS Traffic Engineering | p. 404 |
| MPLS TE Deployment Models and Considerations | p. 408 |
| Setting Tunnel Bandwidth | p. 412 |
| References | p. 414 |
| Index | p. 419 |
| Table of Contents provided by Ingram. All Rights Reserved. |
