Książki

Design your networks to successfully manage their growing complexity

Network professionals have often been told that today’s modern control planes would simplify their networks. The opposite has happened: Technologies like SDN and NFV, although immensely valuable, are exacerbating complexity instead of solving it. Navigating Network Complexity is the first comprehensive guide to managing this complexity in both deployment and day-to-day operations.

Russ White and Jeff Tantsura introduce modern complexity theory from the standpoint of the working network engineer, helping you apply it to the practical problems you face every day. Avoiding complex mathematical models, they show how to characterize network complexity, so you can understand it and control it.

The authors examine specific techniques and technologies associated with network control planes, including SDNs, fast reroute, segment routing, service chaining, and cloud computing. They reveal how each of these affects network design and complexity and help you anticipate causes of failure in highly complex systems.

Next, they turn to modern control planes, examining the fundamental operating principles of SDNs, such as OpenFlow and I2RS, network and other service function virtualization, content distribution networks, Layer 2 fabrics, and service chaining solutions. You’ll learn how each of these might both resolve and increase complexity in network design and operations and what you can do about it.

Coverage includes:

• Defining complexity, understanding its components, and measuring it
• Mastering a straightforward “state, speed, and surface” model for analyzing complexity
• Controlling complexity in design, deployment, operations, protocols, and programmable networks
• Understanding how complex network systems begin to fail and how to prevent failure
• Recognizing complexity tradeoffs in service virtualization and service chaining
• Managing new challenges of complexity in virtualized and cloud environments
• Learning why constructs such as hierarchical design, aggregation, and protocol layering work and when they work best
• Choosing the right models to contain complexity as your network changes

From start to finish, Navigating Network Complexity helps you assess the true impact of new network technologies, so they can capture more value with fewer problems.

Navigating Network Complexity: Next-generation routing with SDN, service virtualization, and service chaining

Introduction . . . . . . . . . . . . xvii

    How This Book Is Organized . . . . . . . xvii

Chapter 1: Defining Complexity . . . . . . . . . 1

    What Is Complexity? . . . . . . . . . 3

        Anything I Don’t Understand . . . . . . 3

        Anything with a Lot of Parts . . . . . . . 5

        Anything for Which There Is More State Than Required to

        Achieve a Goal . . . . . . . . . 9

        Unintended Consequences . . . . . . 11

    Why So Much Complexity? . . . . . . . . 13

        Future Extensions versus New Protocols . . . . 16

        Unexpected Errors . . . . . . . . 17

    Why Not Build Infinitely Complex Systems? . . . . . 18

        Quick, Cheap, and High Quality: Choose Two . . . 20

        Consistency, Availability, and Partition Tolerance:

        Choose Two . . . . . . . . . 21

    Journey into the Center of Complexity . . . . . . 22

Chapter 2: Components of Complexity . . . . . . . 25

    Network Convergence . . . . . . . . . 26

        Path Vector: A BGP Example . . . . . . 26

        Distance Vector: An EIGRP Example . . . . . 28

        Link State: OSPF and IS-IS Convergence . . . . 30

    State . . . . . . . . . . . . 31

        Amount of Information . . . . . . . 32

        An Example of State Failure in the Real World . . . 33

        Final Thoughts on State . . . . . . . 34

    Speed . . . . . . . . . . . . 35

        The Network That Never Converges . . . . . 35

        The Flapping Link . . . . . . . . 37

        Final Thoughts on Speed . . . . . . . 38

    Surface . . . . . . . . . . . . 38

    The Hourglass Model . . . . . . . . . 41

    Optimization . . . . . . . . . . . 44

    A Final Thought . . . . . . . . . . 46

Chapter 3: Measuring Network Complexity . . . . . . 47

    Some Measures of Network Complexity . . . . . 49

        Network Complexity Index . . . . . . 49

        Modeling Design Complexity . . . . . . 51

        NetComplex . . . . . . . . . 53

    Organized Complexity . . . . . . . . . 55

    Is This a Waste of Time? . . . . . . . . . 58

        A Final Thought . . . . . . . . . . 58

Chapter 4: Operational Complexity . . . . . . . . 61

    Exploring the Problem Space . . . . . . . . 61

        The Cost of Human Interaction with the System . . . 61

        Policy Dispersion versus Optimal Traffic Handling . . 66

    Solving the Management Complexity Problem . . . . 69

        Automation as a Solution to Management Complexity . 69

        Modularity as a Solution to Management Complexity . 72

        Protocol Complexity versus Management Complexity . 74

        A Final Thought . . . . . . . . . . 77

Chapter 5: Design Complexity . . . . . . . . . 79

    Control Plane State versus Stretch . . . . . . . 81

        Aggregation versus Stretch . . . . . . . 83

        Traffic Engineering . . . . . . . . 85

        State versus Stretch: Some Final Thoughts . . . . 87

    Topology versus Speed of Convergence . . . . . . 88

        Ring Topology Convergence . . . . . . 88

        Redundancy versus Resilience . . . . . . 90

        Topology versus Speed of Convergence:

        Some Final Thoughts . . . . . . . 93

    Fast Convergence versus Complexity . . . . . . 94

        Improving Convergence with Intelligent Timers: Talk Faster . . . . . . . . . 96

        Removing Timers from Convergence: Precompute . . 99

        Working around Topology: Tunneling to the Loop-Free

        Alternate . . . . . . . . . 101

        Some Final Thoughts on Fast Convergence . . . 104

    Virtualization versus Design Complexity . . . . . 106

        Functional Separation . . . . . . . 108

        Forwarding Plane Complexity . . . . . . 109

        Control Plane Complexity . . . . . . 109

        Shared Fate Risk Groups . . . . . . . 111

    A Final Thought . . . . . . . . . . 111

Chapter 6: Managing Design Complexity . . . . . . 113

    Modularity . . . . . . . . . . . 113

        Uniformity . . . . . . . . . 114

        Interchangeable Modules . . . . . . . 120

        How Modularity Attacks the Complexity Problem . . 121

    Information Hiding . . . . . . . . . 122

        Aggregation . . . . . . . . . 122

        Failure Domains and Information Hiding . . . . 126

        Final Thoughts on Information Hiding . . . . 128

    Models . . . . . . . . . . . 129

        Waterfall . . . . . . . . . . 129

        Places in the Network . . . . . . . 131

        Hierarchical . . . . . . . . . 132

        UML . . . . . . . . . . . 134

    A Final Thought . . . . . . . . . . 136

Chapter 7: Protocol Complexity . . . . . . . . 137

    Flexibility versus Complexity: OSPF versus IS-IS . . . . 138

    Layering versus Protocol Complexity . . . . . . 141

        The Seven-Layer Model . . . . . . . 143

        The Four-Layer Model . . . . . . . 146

        The Iterative Model . . . . . . . . 147

        Protocol Stacks and Design . . . . . . 148

    Protocol Complexity versus Design Complexity . . . . 149

        Microloops and Fast Reroute . . . . . . 149

        EIGRP and the Design Conundrum . . . . . 158

    Final Thoughts on Protocol Complexity . . . . . 162

Chapter 8: How Complex Systems Fail . . . . . . . 163

    Feedback Loops . . . . . . . . . . 164

        Positive Feedback Loops in Network Engineering . . 169

        Speed, State, and Surface: Stability in the

        Network Control Plane . . . . . . . 174

    Shared Fate . . . . . . . . . . . 177

        Virtual Circuits . . . . . . . . . 177

        TCP Synchronization as a Shared Fate Problem . . 179

    A Final Thought . . . . . . . . . . 181

        Thoughts on Root Cause Analysis . . . . . 181

        Engineering Skills and Failure Management . . . 182

Chapter 9: Programmable Networks . . . . . . . 185

    Drivers and Definition . . . . . . . . . 186

        Business Drivers . . . . . . . . 186

        The Ebb and Flow of Centralization . . . . . 188

        Defining Network Programmability . . . . . 191

    Use Cases for Programmable Networks . . . . . 193

        Bandwidth Calendaring . . . . . . . 193

        Software-Defined Perimeter . . . . . . 196

    Programmable Network Interfaces . . . . . . 200

    The Programmable Network Landscape . . . . . 201

        OpenFlow . . . . . . . . . . 202

        YANG . . . . . . . . . . 204

        Path Computation Element Protocol . . . . . 207

        Interface to the Routing System . . . . . 210

    A Final Thought . . . . . . . . . . 212

Chapter 10: Programmable Network Complexity . . . . . 215

    The Subsidiarity Principle . . . . . . . . 216

    Policy Management . . . . . . . . . 217

        Policy Dispersion . . . . . . . . 220

        Policy Consistency . . . . . . . . 222

        Policy Complexity . . . . . . . . 223

    Surface and the Programmable Network . . . . . 224

    Impact on Failure Domains . . . . . . . . 226

        Wide Area Failure Domains . . . . . . 227

        Data Center Failure Domains . . . . . . 228

        Application to Control Plane Failure Domain . . . 229

        Controller to Controller Failure Domain . . . . 229

        Final Thoughts on Failure Domains . . . . . 229

    A Final Thought . . . . . . . . . . 230

Chapter 11: Service Virtualization and Service Chaining . . . . 233

    Network Function Virtualization . . . . . . . 234

        NFV: A Use Case . . . . . . . . 236

    Service Chaining . . . . . . . . . . 242

        Service Function Chaining . . . . . . 243

        Segment Routing . . . . . . . . 245

    A Final Thought . . . . . . . . . . 248

Chapter 12: Virtualization and Complexity . . . . . . 249

    Policy Dispersion and Network Virtualization . . . . 250

        State and Service Chaining . . . . . . 253

        State and Optimization . . . . . . . 254

        Surface and Policy Interaction . . . . . . 255

        Surface and Policy Proxies . . . . . . 255

    Other Design Considerations . . . . . . . 256

        Coupling and Failure Domains . . . . . . 257

        Troubleshooting . . . . . . . . 260

    The Orchestration Effect . . . . . . . . 262

    Managing Complexity . . . . . . . . . 264

    A Final Thought . . . . . . . . . . 266

Chapter 13: Complexity and the Cloud . . . . . . . 267

    Where Does the Complexity Live? . . . . . . 268

        Cloud Centric . . . . . . . . . 269

        Vendor Centric . . . . . . . . . 270

        Network Centric . . . . . . . . 271

        Is There a “Right Way?” . . . . . . . 271

    Centralize What? . . . . . . . . . . 272

    Cloudy Complications . . . . . . . . . 273

        Security . . . . . . . . . . 273

        Data Portability . . . . . . . . 276

    A Final Thought . . . . . . . . . . 277

Chapter 14: A Simple Ending . . . . . . . . . 279

    Defining Complexity . . . . . . . . . 279

        Difficult to Understand . . . . . . . 280

        Unintended Consequences . . . . . . 280

        Large Numbers of Interacting Parts . . . . . 280

        What Makes Something “Too Complex”? . . . 281

    Complexity Is a Tradeoff . . . . . . . . 282

    Modeling Complexity . . . . . . . . . 284

    Managing Complexity in the Real World . . . . . 286

        Don’t Ignore Complexity . . . . . . . 286

        Find a Model to Contain the Complexity . . . . 287

    A Final Thought . . . . . . . . 288

TOC, 9780133989359, 10/28/2015