Understanding Infrastructure Edge Computing

UNDERSTANDING INFRASTRUCTURE EDGE COMPUTING A comprehensive review of the key emerging technologies that will directly impact areas of computer technology over the next five yearsInfrastructure edge computing is the model of data center and network infrastructure deployment which distributes a large number of physically small data centers around an area to deliver better performance and to enable new economical applications. It is vital for those operating at business or technical levels to be positioned to capitalize on the changes that will occur as a result of infrastructure edge computing.This book provides a thorough understanding of the growth of internet infrastructure from its inception to the emergence of infrastructure edge computing. Author Alex Marcham, an acknowledged leader in the field who coined the term ‘infrastructure edge computing,’ presents an accessible, accurate, and expansive view of the next generation of internet infrastructure. The book features illustrative examples of 5G mobile cellular networks, city-scale AI systems, self-driving cars, drones, industrial robots, and more—technologies that increase efficiency, save time and money, and improve safety. Covering state-of-the-art topics, this timely and authoritative book:Presents a clear and accurate survey of the key emerging technologies that will impact data centers, 5G networks, artificial intelligence and cyber-physical systems, and other areas of computer technology Explores how and why Internet infrastructure has evolved to where it stands today and where it needs to be in the near future Covers a wide range of topics including distributed application workload operation, infrastructure and application security, and related technologies such as multi-access edge computing (MEC) and fog computing Provides numerous use cases and examples of real-world applications which depend upon underlying edge infrastructure Written for Information Technology practitioners, computer technology practitioners, and students, Understanding Infrastructure Edge Computing is essential reading for those looking to benefit from the coming changes in computer technology.

Preface xv About the Author xvii Acknowledgements xix 1 Introduction 1 2 What Is Edge Computing? 3 2.1 Overview 3 2.2 Defining the Terminology 3 2.3 Where Is the Edge? 4 2.3.1 A Tale of Many Edges 5 2.3.2 Infrastructure Edge 6 2.3.3 Device Edge 6 2.4 A Brief History 8 2.4.1 Third Act of the Internet 8 2.4.2 Network Regionalisation 10 2.4.3 CDNs and Early Examples 10 2.5 Why Edge Computing? 12 2.5.1 Latency 12 2.5.2 Data Gravity 13 2.5.3 Data Velocity 13 2.5.4 Transport Cost 14 2.5.5 Locality 14 2.6 Basic Edge Computing Operation 15 2.7 Summary 18 References 18 3 Introduction to Network Technology 21 3.1 Overview 21 3.2 Structure of the Internet 21 3.2.1 1970s 22 3.2.2 1990s 22 3.2.3 2010s 23 3.2.4 2020s 23 3.2.5 Change over Time 23 3.3 The OSI Model 24 3.3.1 Layer 1 25 3.3.2 Layer 2 25 3.3.3 Layer 3 26 3.3.4 Layer 4 26 3.3.5 Layers 5, 6, and 7 27 3.4 Ethernet 28 3.5 IPv4 and IPv6 29 3.6 Routing and Switching 29 3.6.1 Routing 30 3.6.2 Routing Protocols 31 3.6.3 Routing Process 34 3.7 LAN, MAN, and WAN 41 3.8 Interconnection and Exchange 42 3.9 Fronthaul, Backhaul, and Midhaul 44 3.10 Last Mile or Access Networks 45 3.11 Network Transport and Transit 46 3.12 Serve Transit Fail (STF) Metric 48 3.13 Summary 51 References 52 4 Introduction to Data Centre Technology 53 4.1 Overview 53 4.2 Physical Size and Design 53 4.3 Cooling and Power Efficiency 54 4.4 Airflow Design 56 4.5 Power Distribution 57 4.6 Redundancy and Resiliency 58 4.7 Environmental Control 61 4.8 Data Centre Network Design 61 4.9 Information Technology (IT) Equipment Capacity 65 4.10 Data Centre Operation 66 4.10.1 Notification 67 4.10.2 Security 67 4.10.3 Equipment Deployment 67 4.10.4 Service Offerings 68 4.10.5 Managed Colocation 68 4.11 Data Centre Deployment 69 4.11.1 Deployment Costing 69 4.11.2 Brownfield and Greenfield Sites 69 4.11.3 Other Factors 70 4.12 Summary 70 References 70 5 Infrastructure Edge Computing Networks 71 5.1 Overview 71 5.2 Network Connectivity and Coverage Area 71 5.3 Network Topology 72 5.3.1 Full Mesh 74 5.3.2 Partial Mesh 74 5.3.3 Hub and Spoke 75 5.3.4 Ring 76 5.3.5 Tree 76 5.3.6 Optimal Topology 76 5.3.7 Inter-area Connectivity 77 5.4 Transmission Medium 78 5.4.1 Fibre 78 5.4.2 Copper 78 5.4.3 Wireless 79 5.5 Scaling and Tiered Network Architecture 80 5.6 Other Considerations 81 5.7 Summary 82 6 Infrastructure Edge Data Centres 83 6.1 Overview 83 6.2 Physical Size and Design 83 6.2.1 Defining an Infrastructure Edge Data Centre 84 6.2.2 Size Categories 84 6.3 Heating and Cooling 102 6.4 Airflow Design 105 6.4.1 Traditional Designs 107 6.4.2 Non-traditional Designs 109 6.5 Power Distribution 113 6.6 Redundancy and Resiliency 114 6.6.1 Electrical Power Delivery and Generation 116 6.6.2 Network Connectivity 118 6.6.3 Cooling Systems 120 6.6.4 Market Design 122 6.6.5 Redundancy Certification 124 6.6.6 Software Service Resiliency 125 6.6.7 Physical Redundancy 126 6.6.8 System Resiliency Example 127 6.7 Environmental Control 128 6.8 Data Centre Network Design 131 6.9 Information Technology (IT) Equipment Capacity 134 6.9.1 Operational Headroom 135 6.10 Data Centre Operation 135 6.10.1 Site Automation 136 6.10.2 Single or Multi-tenant 142 6.10.3 Neutral Host 144 6.10.4 Network Operations Centre (NOC) 145 6.11 Brownfield and Greenfield Sites 147 6.12 Summary 151 7 Interconnection and Edge Exchange 153 7.1 Overview 153 7.2 Access or Last Mile Network Interconnection 153 7.3 Backhaul and Midhaul Network Interconnection 158 7.4 Internet Exchange 160 7.5 Edge Exchange 164 7.6 Interconnection Network Technology 167 7.6.1 5G Networks 168 7.6.2 4G Networks 169 7.6.3 Cable Networks 170 7.6.4 Fibre Networks 172 7.6.5 Other Networks 173 7.6.6 Meet Me Room (MMR) 173 7.6.7 Cross Connection 174 7.6.8 Virtual Cross Connection 176 7.6.9 Interconnection as a Resource 179 7.7 Peering 180 7.8 Cloud On-ramps 181 7.9 Beneficial Impact 183 7.9.1 Latency 183 7.9.2 Data Transport Cost 184 7.9.3 Platform Benefit 185 7.10 Alternatives to Interconnection 186 7.11 Business Arrangements 187 7.12 Summary 188 8 Infrastructure Edge Computing Deployment 189 8.1 Overview 189 8.2 Physical Facilities 189 8.3 Site Locations 191 8.3.1 kW per kM2 192 8.3.2 Customer Facility Selection 193 8.3.3 Site Characteristics 194 8.4 Coverage Areas 195 8.5 Points of Interest 197 8.6 Codes and Regulations 198 8.7 Summary 200 9 Computing Systems at the Infrastructure Edge 203 9.1 Overview 203 9.2 What Is Suitable? 203 9.3 Equipment Hardening 204 9.4 Rack Densification 205 9.4.1 Heterogenous Servers 207 9.4.2 Processor Densification 208 9.4.3 Supporting Equipment 210 9.5 Parallel Accelerators 211 9.5.1 Field Programmable Gate Arrays (FPGAs) 213 9.5.2 Tensor Processing Units (TPUs) 213 9.5.3 Graphics Processing Units (GPUs) 214 9.5.4 Smart Network Interface Cards (NICs) 215 9.5.5 Cryptographic Accelerators 216 9.5.6 Other Accelerators 217 9.5.7 FPGA, TPU, or GPU? 217 9.6 Ideal Infrastructure 218 9.6.1 Network Compute Utilisation 218 9.7 Adapting Legacy Infrastructure 221 9.8 Summary 221 References 222 10 Multi-tier Device, Data Centre, and Network Resources 223 10.1 Overview 223 10.2 Multi-tier Resources 223 10.3 Multi-tier Applications 226 10.4 Core to Edge Applications 228 10.5 Edge to Core Applications 230 10.6 Infrastructure Edge and Device Edge Interoperation 231 10.7 Summary 234 11 Distributed Application Workload Operation 235 11.1 Overview 235 11.2 Microservices 235 11.3 Redundancy and Resiliency 236 11.4 Multi-site Operation 237 11.5 Workload Orchestration 238 11.5.1 Processing Requirements 240 11.5.2 Data Storage Requirements 240 11.5.3 Network Performance Requirements 241 11.5.4 Application Workload Cost Profile 241 11.5.5 Redundancy and Resiliency Requirements 242 11.5.6 Resource Marketplaces 243 11.5.7 Workload Requirement Declaration 243 11.6 Infrastructure Visibility 244 11.7 Summary 245 12 Infrastructure and Application Security 247 12.1 Overview 247 12.2 Threat Modelling 247 12.3 Physical Security 249 12.4 Logical Security 250 12.5 Common Security Issues 251 12.5.1 Staff 251 12.5.2 Visitors 252 12.5.3 Network Attacks 252 12.6 Application Security 253 12.7 Security Policy 254 12.8 Summary 255 13 Related Technologies 257 13.1 Overview 257 13.2 Multi-access Edge Computing (MEC) 257 13.3 Internet of Things (IoT) and Industrial Internet of Things (IIoT) 258 13.4 Fog and Mist Computing 259 13.5 Summary 260 Reference 260 14 Use Case Example: 5G 261 14.1 Overview 261 14.2 What Is 5G? 261 14.2.1 5G New Radio (NR) 262 14.2.2 5G Core Network (CN) 263 14.3 5G at the Infrastructure Edge 264 14.3.1 Benefits 264 14.3.2 Architecture 264 14.3.3 Considerations 265 14.4 Summary 266 15 Use Case Example: Distributed AI 267 15.1 Overview 267 15.2 What Is AI? 268 15.2.1 Machine Learning (ML) 268 15.2.2 Deep Learning (DL) 269 15.3 AI at the Infrastructure Edge 270 15.3.1 Benefits 270 15.3.2 Architecture 271 15.3.3 Considerations 272 15.4 Summary 273 16 Use Case Example: Cyber-physical Systems 275 16.1 Overview 275 16.2 What Are Cyber-physical Systems? 275 16.2.1 Autonomous Vehicles 276 16.2.2 Drones 278 16.2.3 Robotics 280 16.2.4 Other Use Cases 280 16.3 Cyber-physical Systems at the Infrastructure Edge 280 16.3.1 Benefits 280 16.3.2 Architecture 281 16.3.3 Considerations 282 16.4 Summary 282 Reference 283 17 Use Case Example: Public or Private Cloud 285 17.1 Overview 285 17.2 What Is Cloud Computing? 286 17.2.1 Public Clouds 286 17.2.2 Private Clouds 287 17.2.3 Hybrid Clouds 287 17.2.4 Edge Cloud 288 17.3 Cloud Computing at the Infrastructure Edge 288 17.3.1 Benefits 288 17.3.2 Architecture 289 17.3.3 Considerations 290 17.4 Summary 290 18 Other Infrastructure Edge Computing Use Cases 291 18.1 Overview 291 18.2 Near Premises Services 291 18.3 Video Surveillance 293 18.4 SD-WAN 294 18.5 Security Services 295 18.6 Video Conferencing 296 18.7 Content Delivery 297 18.8 Other Use Cases 298 18.9 Summary 299 19 End to End: An Infrastructure Edge Project Example 301 19.1 Overview 301 19.2 Defining Requirements 301 19.2.1 Deciding on a Use Case 302 19.2.2 Determining Deployment Locations 304 19.2.3 Identifying Required Equipment 306 19.2.4 Choosing an Infrastructure Edge Computing Network Operator 307 19.2.5 Regional or National Data Centres 307 19.3 Success Criteria 307 19.4 Comparing Costs 308 19.5 Alternative Options 309 19.6 Initial Deployment 310 19.7 Ongoing Operation 311 19.7.1 SLA Breaches 312 19.8 Project Conclusion 312 19.9 Summary 314 20 The Future of Infrastructure Edge Computing 315 20.1 Overview 315 20.2 Today and Tomorrow 315 20.3 The Next Five Years 316 20.4 The Next 10 Years 316 20.5 Summary 316 21 Conclusion 317 Appendix A: Acronyms and Abbreviations 319 Index 323