106,99 €
Mehr erfahren.
- Herausgeber: John Wiley & Sons
- Kategorie: Wissenschaft und neue Technologien
- Serie: The ComSoc Guides to Communications Technologies
- Sprache: Englisch
Discover the latest advances in computer architecture and software at the dawn of the 5G/6G era
In Edge Computing Acceleration: From 5G to 6G and Beyond, distinguished researchers Dr. Patrick Hung, Hongwei Kan, and Greg Knopf deliver a comprehensive overview of personal computer architecture and software design usage in the upcoming 5G decade. The authors begin by introducing key components and exploring different hardware acceleration architectures. They move on to discuss 5G data security and data integrity and offer a survey of network virtualization technologies, including accelerated virtualization technologies.
The book analyzes 5G/6G system performance, investigating key design considerations and trade-offs and introducing high-level synthesis flow. It concludes with chapters exploring design verification and validation flow, illustrations of 5G applications based on artificial intelligence and other emerging technologies and offering highlights of emerging 6G research and roadmaps.
Readers will enjoy the combination of accessible descriptions of new technologies presented side-by-side as a step-by-step guide to designing effective 5G systems. The book also includes:
- A thorough introduction to key 5G/6G components, including new wireless communication protocols, edge and fog computing, acceleration technologies, IoE architectures, software-designed networks, network function virtualization, and data security
- Explorations of various hardware acceleration architectures, like FPGA and GPU acceleration architectures
- Practical discussions of 5G/6G data security, data integrity, and a survey of network virtualization technologies
- In-depth treatments of 5G/6G system performance, key design considerations, high-level synthesis flow, design verification, and validation flow
Perfect for undergraduate and graduate students in programs related to communications technology, engineering, and computer science, Edge Computing Acceleration: From 5G to 6G and Beyond is a must-have resource for engineers, programmers, system architects, technical managers, communications business executives, telco operators, and government regulators who regularly interact with cutting-edge communications equipment.
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 464
Veröffentlichungsjahr: 2024
Ähnliche
Table of Contents
Cover
Table of Contents
Title Page
Copyright
About the Authors
Foreword (Professor Ray Cheung)
Foreword (Raghu Nambiar)
Preface
Acknowledgment (Patrick Hung)
Acknowledgment (Greg Knopf)
Part I: Introduction
1 Introduction
1.1 Introducing 5G and Internet of Everything
1.2 Edge Computing Architecture
1.3 Custom Computing
1.4 Deployment Considerations
References
2 Overview of 5G and 6G
2.1 5G Timeline
2.2 5G Spectrum
2.3 Characteristics of 5G
2.4 5G New Radio
2.5 Data Plane and Control Plane Separation
2.6 5G Applications
2.7 Smooth Transition to 6G
2.8 6G Expected Timeline, Spectrum, and Characteristics
2.9 6G Potential Applications
2.10 Edge, Fog, and Cloud Computing in Relation to 5G and 6G
References
Part II: Theory
3 High-Level Synthesis (HLS)
3.1 Why Use High-Level Synthesis?
3.2 Common HLS Languages and Platforms
3.3 Limitations and Challenges of HLS
3.4 Using HLS in 5G Edge Computing
References
4 Coding Design
4.1 Overview
4.2 Error Correction Codes (ECCs)
4.3 Security Codes
4.4 Emerging 5G Security Design Acceleration
References
Part III: Architecture
5 Hardware Architecture
5.1 Development Timeline
5.2 Operating Spectrum
5.3 Core Requirements
5.4 New Radio Access Technology
5.5 Network Architecture
5.6 Performance Improvement
References
Note
6 Software Architecture
6.1 End-to-End Example of 5G System
6.2 Network Slicing Architecture, Software-Defined Network, and Network Function Virtualization
6.3 Software Acceleration
References
Part IV: Applications
7 Killer Applications
7.1 Metaverse and Its Trends
7.2 Technologies Behind Metaverse
7.3 Applications of Metaverse
7.4 Accelerating Killer Apps
References
8 From Concept to Production
8.1 System Design Process
8.2 Some Examples
8.3 Standards Compliance
8.4 Other Design Metrics
8.5 Summary
References
Part V: Future Roadmap
9 The Road Ahead
9.1 Spatial Computing and Networking
9.2 Supporting 5G/6G Spatial Computing and Networking
9.3 Migrating to 6G
9.4 Enabling Technologies for 5G and Beyond
9.5 Some Final Thoughts
References
Index
End User License Agreement
List of Tables
Chapter 1
Table 1.1 Round-trip time (RTT) from Tokyo to overseas cities.
Table 1.2 Comparison of 5G private network deployment architectures.
Table 1.3 Comparison of three 5G operation models.
Table 1.4 Some CAPEX estimation for 5G deployment.
Chapter 5
Table 5.1 NR operating bands in FR1.
Table 5.2 NR operating bands in FR2.
Table 5.3 Comparison of per-bit energy dissipation in data transmission via ...
Chapter 6
Table 6.1 Timeline of NFV release.
Chapter 8
Table 8.1 Comparison between two hardware design options.
List of Illustrations
Chapter 1
Figure 1.1 Comparison between 4G and 5G features and performance metrics.
Figure 1.2 Latency and bandwidth requirements of some typical 4G and 5G appl...
Figure 1.3 3GPP 5G network architecture.
Figure 1.4 Ubiquitous Internet of Everything (IoE) devices.
Figure 1.5 Edge versus cloud computing architecture.
Figure 1.6 Edge versus cloud network latencies.
Figure 1.7 Edge computing is driving many 5G applications.
Figure 1.8 5G computing and communications networks.
Figure 1.9 Key challenges to fully realize 5G potentials.
Figure 1.10 Many different ways to attack 5G network.
Figure 1.11 Employment of custom computing in mobile edge computing and fog ...
Figure 1.12 An NVIDIA Tesla T4 PCIe card and a Xilinx Alveo U200 PCIe card. ...
Figure 1.13 Xilinx Alveo U25 SmartNIC acceleration card encompasses network,...
Figure 1.14 5G fixed wireless access (FWA) application.
Figure 1.15 Small cells versus macrocells.
Figure 1.16 Deployment architectures A, B, C, and D.
Figure 1.17 Different 5G operation models.
Figure 1.18 CAPEX requirements for various coverage models.
Figure 1.19 Optimizing coverage using shared small cell layer.
Chapter 2
Figure 2.1 The latest 5G timeline from ITU and 3GPP, which uses a system of ...
Figure 2.2 5G standard requirements.
Figure 2.3 Multi-subcarrier signals arrangement in comparison to a single ca...
Figure 2.4 Example MIMO configurations: 2 × 2 and 4 × 4.
Figure 2.5 Antenna beam targeting particular user equipment.
Figure 2.6 Multiuser MIMO configuration.
Figure 2.7 Key 5G applications.
Figure 2.8 AMD Virtex Ultrascale (VCU108) FPGA Evaluation Board. Advanced Mi...
Figure 2.9 Potential 6G applications.
Figure 2.10 Simplified computing infrastructure showing the edge, fog, and c...
Chapter 3
Figure 3.1 Overview of HLS showing (a) input, (b) key tasks, and (c) output....
Chapter 4
Figure 4.1 Simplified model of an ECC system.
Figure 4.2 Encoder structure of turbo codes.
Figure 4.3 Parity-check matrix and Tanner graph of a (10,5) code.
Figure 4.4 VN update diagram.
Figure 4.5 CN update diagram.
Figure 4.6 Capacity of polarized sub-channels with block length of 1024 [11]...
Figure 4.7 A simplified public key infrastructure (PKI).
Figure 4.8 Symmetric key cryptography.
Figure 4.9 Asymmetric key cryptography.
Figure 4.10 Blockchain for 5G [22]/with permission of Elsevier.
Figure 4.11 Feistel structure in block ciphers for (a) SIMON and (b) SPECK [...
Figure 4.12 Encryption process of PRESENT block cipher [31]/MDPI/CCBY 4.0/Pu...
Figure 4.13 Encryption process of GIFT block cipher [31]/MDPI/CCBY 4.0/Publi...
Figure 4.14 Multicast network illustrating network coding [36].
Figure 4.15 Homomorphic encryption illustration.
Figure 4.16 Illustration of zero-knowledge proof.
Chapter 5
Figure 5.1 The 5G timeline: ITU-R and 3GPP releases from 5G network and 3GPP...
Figure 5.2 5G standard requirements.
Figure 5.3 Multi-subcarrier signals arrangement in comparison to a single ca...
Figure 5.4 Example of MIMO configurations: 2 × 2 and 4 × 4.
Figure 5.5 Antenna beam targeting particular user equipment.
Figure 5.6 Multiuser MIMO configuration.
Figure 5.7 A generic 5G network.
Figure 5.8 NG-RAN architecture.
Figure 5.9 5G system software-based architecture [9]/https://www.etsi.org/de...
Figure 5.10 Comparison of data communication flows among four different Nvid...
Figure 5.11 The key components within the Inter-Kernel Links (IKL) architect...
Figure 5.12 The distributed FPGA-based acceleration platform, jointly develo...
Figure 5.13 The Smart_xPU architectural design follows a novel design method...
Figure 5.14 The on-chip and off-chip data transfers between two GPU nodes ar...
Figure 5.15 The data path of traditional Ethernet and direct addressing.
Figure 5.16 The xPU architecture with tightly coupled communication and comp...
Figure 5.17 Comparison of the OpenCL programming paradigm and the traditiona...
Figure 5.18 The Smart_xPU software architecture overview.
Figure 5.19 The Smart_xPU prototype architecture.
Chapter 6
Figure 6.1 Example of an end-to-end 5G system.
Figure 6.2 Network slicing architecture.
Figure 6.3 SDN architecture.
Figure 6.4 DPDK-assisted user space approach.
Figure 6.5 Simplified virtual network function with and without DPDK. (a) No...
Figure 6.6 Arrangement showing DPDK with SR-IOV.
Figure 6.7 Test results in north–south direction [11]/with permission of Int...
Figure 6.8 Test results in east–west direction [11]/with permission of Intel...
Figure 6.9 Illustration of data of conventional setup (using TCP/IP) and RDM...
Figure 6.10 CXL protocols.
Figure 6.11 Example use of DPU [23].
Chapter 7
Figure 7.1 Technologies supporting metaverse.
Chapter 8
Figure 8.1 Key activities ensuring final products meet MRD and other specifi...
Figure 8.2 The power management feature in the downstream inference unit can...
Figure 8.3 The driver implementation is the performance bottleneck in this f...
Chapter 9
Figure 9.1 Concept of PIM.
Figure 9.2 A disaggregated SoC comprising multiple chiplets [19].
Figure 9.3 HBM structure showing memory chips arranged in a 3D-stacked forma...
Guide
Cover
Table of Contents
Title Page
Copyright
About the Authors
Foreword (Professor Ray Cheung)
Foreword (Raghu Nambiar)
Preface
Acknowledgment (Patrick Hung)
Acknowledgment (Greg Knopf)
Begin Reading
Index
End User License Agreement
Pages
ii
iii
iv
xi
xiii
xiv
xv
xvii
xviii
xix
xx
xxi
xxii
1
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
241
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
IEEE Press445 Hoes LanePiscataway, NJ 08854
IEEE Press Editorial BoardSarah Spurgeon, Editor-in-Chief
Moeness Amin
Jón Atli Benediktsson
Adam Drobot
James Duncan
Ekram Hossain
Brian Johnson
Hai Li
James Lyke
Joydeep Mitra
Desineni Subbaram Naidu
Tony Q. S. Quek
Behzad Razavi
Thomas Robertazzi
Diomidis Spinellis
Edge Computing Acceleration
From 5G to 6G and Beyond
Patrick Hung, Hongwei Kan, and Greg Knopf
Copyright © 2025 by The Institute of Electrical and Electronics Engineers, Inc. All rights reserved.
Published by John Wiley & Sons, Inc., Hoboken, New Jersey.Published simultaneously in Canada.
No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission.
Trademarks: Wiley and the Wiley logo are trademarks or registered trademarks of John Wiley & Sons, Inc. and/or its affiliates in the United States and other countries and may not be used without written permission. All other trademarks are the property of their respective owners. John Wiley & Sons, Inc. is not associated with any product or vendor mentioned in this book.
Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.
For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002.
Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com.
Library of Congress Cataloging-in-Publication Data Applied for:
Hardback ISBN: 9781119813842
Cover Design: WileyCover Image: © DrPixel/Getty Images
About the Authors
Patrick Hung received his PhD in electrical engineering from Stanford University and served as a consulting assistant professor there from 2003 to 2009. He has been recognized as a CBI overseas scholar in the United Kingdom and a Taishan scholar in China. Patrick co-founded several high-tech companies and is the co-director of the CALAS Laboratory at City University of Hong Kong. He also serves as the vice chair of the IEEE Hong Kong Computer Society.
Hongwei Kan is the General Manager of the Emerging Technologies Research Laboratory at Tsinghua Unigroup. His research interests include xPU and chiplet design, heterogeneous computing platforms, emerging computer architectures, and artificial intelligence. Before joining Tsinghua Unigroup, Hongwei was the Chief Architect at Inspur Group, overseeing high-performance computing and storage platforms. Professor Kan is an IEEE Senior Member.
Greg Knopf is the Senior Director of Server Customer Engineering at AMD, where he oversees the development of AMD EPYC-based server platforms from concept to launch. His research interests include software/hardware co-development, cloud-native technologies, at-scale debug, and hardware acceleration technologies in artificial intelligence and other datacenter applications. Before joining AMD, Greg was the Senior Director of Engineering at Intel, leading multiple worldwide design teams.
Foreword (Professor Ray Cheung)
As a university professor in computer engineering, teaching system-on-chip, and a good friend to the authors, I am pleased to introduce you to Edge Computing Acceleration: From 5G to 6G and Beyond. Authored by Patrick, Hongwei, and Greg, this book is a comprehensive guide to the rapidly evolving world of edge computing and its intersection with 5G and 6G technologies.
We are at a pivotal point in the technological landscape where the convergence of edge computing and advanced communication networks is set to redefine our digital experiences. This book is a testament to the authors’ deep understanding and expertise in high-performance computing, semiconductor chip design, digital security, and communication networks from academic and industry perspectives.
The book is structured into nine insightful chapters, each devoted to a critical aspect of edge computing acceleration. It begins with an introduction to the fundamentals of edge and custom computing, leading us through the complexities of 5G and 6G networks. The authors expertly discuss high-level synthesis (HLS) as a tool for accelerating hardware design and delve into the essential coding design principles for ensuring security in edge computing systems.
This book will serve as an invaluable resource and handbook in my senior-year undergraduate and postgraduate classes. It is designed to equip students, researchers, and practitioners with the knowledge and skills necessary to navigate and contribute to the field of edge computing acceleration. The blend of theoretical concepts and practical examples fosters a deep understanding of the subject matter and encourages innovation in this exciting domain.
I have had the distinct pleasure of co-teaching a university course on system-on-chip design with Patrick, one of the authors of this book. His deep understanding of the subject matter and his ability to convey complex concepts in an accessible manner have made him a favorite among students. Patrick is at the forefront of academic study, constantly pushing boundaries and exploring new areas in the field. His extensive experience in academia and his direct engagement with students have given him a unique perspective on the needs and challenges faced by learners in this rapidly evolving field.
As we embark on this journey of exploration and learning together, I am confident that Edge Computing Acceleration: From 5G to 6G and Beyond will guide you through the complexities and opportunities of this transformative era in computing and communication. I wholeheartedly recommend this book to anyone interested in the future of edge computing and the evolution of 5G and 6G technologies.
Associate Provost, City University of Hong KongChair, IEEE Hong Kong Section
Professor Ray Cheung
Foreword (Raghu Nambiar)
The industry landscape is evolving at an unprecedented pace, driven by rapid technological progress and a shift toward smarter and more efficient edge devices. The advent of 5G and the anticipation of 6G are fundamentally reshaping the digital infrastructure landscape. Edge Computing Acceleration: From 5G to 6G and Beyond, authored by industry experts Patrick Hung, Hongwei Kan, and Greg Knopf, offers a profound exploration of the technologies shaping our interconnected tomorrow.
The subjects explored in this book hold personal significance for me. In my current role as the lead for software and solution engineering at AMD, and in my previous position as the CTO of the server business at Cisco overseeing emerging technology solutions, I have engaged closely with these topics. This book aligns seamlessly with my market insights and aspirations. I am convinced it offers valuable insights for both knowledge seekers and practitioners.
In my career journey, I have personally witnessed the transformative power of collaboration in driving technological progress. This book is a shining example of such collaboration, as its authors bring together a wealth of knowledge from both academic and industry perspectives. Over the years, I have had the privilege of closely collaborating with Greg Knopf, an industry leader who has played a crucial role in the successful deployment of millions of servers in data centers across the globe. I wholeheartedly extend my congratulations and appreciation to him and the co-authors for their outstanding teamwork, seamlessly blending academic insights with practical industry perspectives.
As you delve into this enlightening guide, I urge you to envision the boundless potential of these technologies. The capability to process data at the edge, in data centers, and throughout the network unlocks unparalleled opportunities across diverse industries. Truly, we are living in an exhilarating era in computing and telecommunications. My wish is that this book not only serves as a comprehensive roadmap but also sparks inspiration to innovate and push the limits of what is achievable.
Corporate Vice President, AMDSan Francisco Bay Area
Raghu Nambiar
Preface
We are standing at the cusp of a technological revolution, where the confluence of 5G, 6G, and edge computing is reshaping the landscape of digital communication and computation. It is a pivotal moment that demands a comprehensive understanding of these technologies and their potential to accelerate the performance and capabilities of edge computing. In this context, we have endeavored to present Edge Computing Acceleration: From 5G to 6G and Beyond, a book that aims to bridge the gap between theoretical concepts and practical applications in this rapidly evolving domain.
The genesis of this book lies in our diverse expertise across high-performance computing, semiconductor chip design, digital security, and communication networks. Patrick Hung, an expert in computer architecture and SOC design; Hongwei Kan, an architect of heterogeneous acceleration platforms; and Greg Knopf, a specialist in CPU and HPC platforms, have collaboratively crafted a comprehensive resource that addresses the multifaceted aspects of edge computing acceleration in the era of 5G and 6G.
This book is structured into nine chapters, each focusing on a critical aspect of edge computing acceleration. The journey begins with an introduction to the fundamentals of edge and custom computing and its significance in the context of 5G and 6G networks. We then delve into the technical intricacies of 5G and 6G, exploring their architecture, capabilities, and the role they play in enhancing edge computing. High-level synthesis (HLS) is discussed as a pivotal tool for accelerating hardware design, followed by an examination of coding design principles for ensuring security and integrity in edge computing systems.
The hardware and software architectures form the backbone of efficient edge computing systems, and we dedicate two chapters to dissect these components and their interplay. As the potential of edge computing is realized through its applications, we explore the “Killer Applications” that are set to transform industries. The transition from concept to production is a critical phase, and we provide insights and guidelines for navigating this process successfully. Lastly, the “The Road Ahead” chapter offers a forward-looking perspective on the future developments in edge computing and the evolving landscape of 5G and 6G technologies.
Intended for use in a semester-long graduate-level course, this book is designed to equip students, researchers, and practitioners with the knowledge and skills necessary to navigate and contribute to the field of edge computing acceleration. Through a blend of theoretical concepts and practical examples, we aim to foster a deep understanding of the subject matter and inspire innovation in this exciting domain.
As we embark on this journey together, we hope that Edge Computing Acceleration: From 5G to 6G and Beyond serves as a valuable resource, guiding you through the complexities and opportunities of this transformative era in computing and communication.
March 2024
Patrick Hung, Hongwei Kan, and Greg Knopf
Acknowledgment (Patrick Hung)
I am profoundly grateful to everyone who has contributed to the creation of this book. This work represents a collaborative endeavor to delve into the cutting-edge of edge acceleration technology and its evolution through the eras of 5G, 6G, and beyond.
In this book, we tackle the challenges of the 5G era, where an array of artificial intelligence (AI) and robotic applications necessitate hardware acceleration through accelerators such as graphic processing units (GPUs), field programmable gate arrays (FPGAs), and data processing units (DPUs). Regrettably, the proprietary nature of most commercial solutions makes it challenging to grasp the bigger picture. This project was conceived to systematically educate students and engineers in edge acceleration technology, filling the knowledge gap and fostering a deeper understanding of this crucial field.
I extend my heartfelt appreciation to my co-authors, Greg Knopf and Hongwei Kan, for their invaluable contributions, insights, and collaborative spirit. Greg’s expertise in CPUs and servers, coupled with Hongwei’s proficiency in high-performance computing, has been instrumental in this project. Working alongside such knowledgeable and dedicated individuals has been both an honor and a privilege.
Special acknowledgment goes to Professor Qin Huang from Bei Hang University, who provided most of the materials for Chapter 4. Professor Huang is dedicated to advanced coding research and has the remarkable ability to simplify complex mathematical concepts, significantly enhancing the quality and accessibility of this book.
I am deeply indebted to Kam Sum and Abdurrashid Ibrahim Sanka, two dear friends who dedicated countless hours to reviewing every chapter. Without their assistance, this book would not have been possible. We are also grateful to our editor and anonymous reviewers for their constructive criticism and guidance, which have greatly improved the quality and clarity of our work.
We owe a debt of gratitude to our colleagues and peers in the academic and industry communities for their invaluable insights, feedback, and contributions, which have enriched the content and depth of this book. I would like to express my profound gratitude to Professor Michael Flynn, whose support and guidance were instrumental in my research on computer architecture. Special thanks also go to our research teams and assistants, whose dedication, hard work, and attention to detail have been instrumental in bringing this project to fruition.
I would like to express my heartfelt thanks to my wife, Jane, and my daughter, Madeline, for their constant support and encouragement. Their belief in me has been a constant source of inspiration.
Our appreciation extends to the various organizations and institutions that provided the necessary resources, infrastructure, and financial support for our research. Their support has been essential in enabling us to explore and push the boundaries of edge computing technology.
Lastly, we acknowledge the readers and the broader community interested in edge computing and its future developments. We hope that this book serves as a valuable resource and inspires further innovation and exploration in the field.
Thank you all for being a part of this journey.
March 2024
Sincerely,
Patrick Hung
Acknowledgment (Greg Knopf)
I would like to express my immense gratitude to everyone who contributed to the creation of this book and to the many influential people in my life who provided support and inspiration throughout my academic and professional journey.
First, I would like to thank my co-authors, Dr. Patrick Hung and Hongwei Kan. I am deeply grateful to Patrick for initiating the project, graciously inviting me to participate, and steadfastly leading it through to completion as we juggled our various responsibilities in life and work. Patrick has been a great friend and mentor since we first collaborated on server platform development and memory technology in the mid-2000s. We have had nearly two decades of deep and exciting conversations about industry trends and emerging technologies, often stretching our dinners in the San Francisco Bay Area or Hong Kong until restaurants are practically pushing us out the door. I am equally grateful to our co-author Hongwei Kan, who brought extremely valuable insights and contributions to the project from his extensive background in high-performance computing.
I am tremendously grateful to my wife, Adrienne, for her constant support and encouragement along the way. She helped me find time and space to focus and pursue this endeavor with my colleagues. I am also immensely grateful for our lovely daughter Heidi, my never ending source of inspiration.
I would also like to thank my lifelong friends Allen White, John Dering, David Dieruf, William Morrow III, David Wolford, Chad Hartley, and Pat Evans for always motivating and encouraging each other in our various pursuits.
At AMD, I have had the honor to work with renowned engineering leaders who are actively shaping the computing industry today, and I feel very fortunate to have the opportunity to learn from them. Dr. Lisa Su, AMD CEO, has been a spectacular leader and constant source of inspiration. It has also been an immense pleasure and learning experience to work closely with Mark Papermaster, AMD CTO, as we have grown the EPYC CPU product line and business.
I would also like to thank Jay Kirkland for his many years of leadership, collaboration, and friendship. Jay’s support and partnership over many years helped to create numerous opportunities and relationships that played an important role in the path to developing this book.
Finally, I want to acknowledge and thank all of the readers of this book and the broader community. We are living in an extremely exciting time in the world of computing, and we hope that this book helps to fuel new innovations and developments in the industry.
Thank you,D. Greg Knopf
Boulder, ColoradoMay 2024
Part IIntroduction
1Introduction
The fifth-generation (5G) and the upcoming sixth-generation (6G) wireless networks are poised to be the most significant and transformative technology in the coming decade, disrupting numerous industries ranging from energy, agriculture, manufacturing to transportation, retail, healthcare, entertainment, and financial services. Surprisingly and unbeknownst to most outsiders, one of the key technology enablers is the new computer architecture and software design model, in addition to the wireless communications technology [1, 2].
The main objective of this book is to help university students and professional engineers understand the 5G/6G edge computing architecture and to describe step by step how to unleash the full 5G/6G potential using custom edge computing technologies.
This book is divided into five parts:
Part 1
(Introduction):
–
Chapters 1
and
2
introduce the concept of edge and custom computing and provide an overview of 5G/6G technologies.
Part 2
(Theory):
–
Chapters 3
and
4
discuss how to use high-level synthesis (HLS) and coding theory to realize secure edge acceleration with high performance.
Part 3
(Architecture):
–
Chapters 5
and
6
elaborate the 5G/6G hardware and software acceleration architecture.
Part 4
(Applications):
–
Chapters 7
and
8
describe a few 5G/6G killer applications with acceleration technology including some practical development strategies.
Part 5
(Future Roadmap):
–
Chapter 9
discusses the road ahead in the coming decade.
1.1 Introducing 5G and Internet of Everything
With 20 times higher bandwidth and 10 times reduction in network latency, fifth-generation (5G) technologies enable many new applications, such as remote surgery and autonomous driving. These incredible applications require both high bandwidth and extremely low latency. To realize these useful applications, we need high-performance communication and computing platforms demanding different kinds of acceleration technologies, known as heterogeneous acceleration architecture (Figure 1.1).
Although the fundamental wireless technology – wireless spectral efficiency – has only increased three times from 4G/LTE to 5G, the area traffic capacity is 100 times larger. This is made possible mainly by building many more 5G cell towers, each covering a smaller area and supporting a wider spectrum. A typical 4G/LTE cell tower may have a range of 10 km, but a typical 5G tower may have a range of 500 m or less.
Figure 1.1 Comparison between 4G and 5G features and performance metrics.
The 3rd-Generation Partnership Project (3GPP) is responsible for driving the 5G telecommunications standards, which are detailed in the 3GPP Specifications Release 15 (NR Phase 1) and Release 16 (NR Phase 2) documents. The 3GPP 5G network architecture defines the network entities based on their functions and nature (control and data planes) [3–6].
There are three technical specification groups (TSGs) within the 3GPP:
TSG SA
(services and systems aspects) focuses on the overall architectures and services.
TSG CT
(core network and terminals) focuses on the core network (CN) architecture and terminal interfaces.
TSG RAN
(radio access networks) focuses on the radio transmission and its technical requirements.
The TSG SA defines three different sets of requirements for new 5G usages:
Enhanced Mobile Broadband (eMBB):
a new requirement that defines higher data rates, traffic and connection densities, and user mobility.
Massive Machine-Type Communications (mMTC):
a new requirement that supports very high traffic densities of devices.
Ultra-Reliable Low Latency Communications (URLLC):
a new requirement that provides very low latency and very high communications service availability (
Figure 1.2
).
Figure 1.2 compares the bandwidth and latency requirements of some 4G and 5G applications. The typical 4G network latency is 10–20 ms, while the minimum 5G network latency can go as low as 1 ms. The maximum 4G network bandwidth is around 100 Mbps, but the 5G network bandwidth can go up to 1 Gbps. Some new mobile applications, such as mobile telepresence, may only require higher network bandwidth. On the other hand, most disruptive applications, such as autonomous driving and smart energy grid, may demand very short network latencies.
Although human beings may not perceive any difference between 5 and 50 ms, this difference is so important in many industrial robotic systems. Indeed, an interconnection of machine systems – called the Internet of Everything (IoE) – can take full advantage of the three new 5G features (eMBB, mMTC, and URLLC) [7, 8] (Figure 1.3).
In 2012, Cisco Systems extended the concept of the Internet of Things (IoT) and coined the term “the Internet of Everything (IoE),” representing a networked connection of people, processes, data, and things. It goes beyond simple machine-to-machine (M2M) communications, forming a network of networks connecting all data, technologies, processes, and people [9–12].
Figure 1.2 Latency and bandwidth requirements of some typical 4G and 5G applications.
Figure 1.3 3GPP 5G network architecture.
Figure 1.4 Ubiquitous Internet of Everything (IoE) devices.
Today many IoE devices are adopting a tethered communication scheme – using Wi-Fi or Bluetooth to first connect to a smartphone or an access point – to communicate with other IoE devices. Unfortunately, tethered communication is slow, unreliable, and insecure, limiting the performance of these applications [13, 14].
The 5G/6G wireless network enables IoE devices to adopt an untethered communication scheme, thus allowing each device to have a secure channel connecting to a radio access network (RAN) with guaranteed network latency, bandwidth, and security (Figure 1.4).
According to Juniper Research, IoE devices reached 46 billion units by 2021. Figure 1.4 depicts a framework of different connected devices and things. The top IoE applications include smart factories, smart buildings, smart grids, connected gaming and entertainment, smart vehicles, remote healthcare, remote education, and connected marketing and advertisement.
These ubiquitous IoE devices have some essential characteristics:
Resource Limitations:
Many IoE devices are limited by size, weight, and power (SWaP) and do not have sufficient resources to process or store information locally. Accordingly, the information is offloaded to a remote server for storage and processing.
Network Latency:
Extremely low network latency is needed for time-critical applications. Industrial control systems can only tolerate delays of the order of milliseconds. Autonomous vehicles and virtual reality applications also have similar latency requirements. The network latency can be excessive if the servers are in a remote cloud.
Network Bandwidth:
A large IoE network can generate a huge amount of real-time data. For example, 120,000 CCTV devices transmitting 1080p video to a remote cloud may generate 1 Tbps traffic. If the servers are in a remote cloud, the 1 Tbps network data can cause serious traffic congestion in the Internet backbone.
Security and Reliability:
Many critical applications, such as energy grids and smart factories, must be ultra-secure and reliable. Due to the multi-hop nature of the Internet backbone, it is probably impossible to guarantee ultra-security and reliability with a remote cloud server.
Reconfigurability:
After deployment, an IoE system oftentimes requires future security patches and performance updates. While it is easier to update a generic software architecture, it is difficult to update a high-performance hardware architecture. This is an important reason why custom computing is desirable for 5G computing architecture.
1.2 Edge Computing Architecture
Although 5G can provide large bandwidths and low latencies, if the other end of the communication is far away, these parameters cannot be guaranteed. In particular, the communication latency is limited by the speed of light. For example, a distance of 4000 km – between Los Angeles and New York – has at least 20 ms delay on fiber optics [15–17].
In reality, the actual latencies, including all electronic and electro-optic delays, are much longer. Table 1.1 shows the typical round-trip latencies from Tokyo to some target cities.
Table 1.1 Round-trip time (RTT) from Tokyo to overseas cities.
Source: Adapted from http://Wondernetwork.com.
Target city
Round-trip time from Tokyo (ms)
Cape Town
360
New York
176
London
218
Los Angeles
128
Paris
236
San Paulo
275
Shanghai
67
Sydney
114
It is important to note that a saving of 20 ms – migrating from a 4G network to a 5G network – is not meaningful if the application involves a network round-trip time of 360 ms. We must keep this in mind when we implement a remote application, such as remote surgery or Internet gaming. The remote server cannot be too far away, or the underlying algorithm must be able to hide the latency from the users.
As much as possible, the IoE servers should be placed close to the wireless networks to optimize the network security, reliability, bandwidth, and latency. This requirement gives rise to the emerging edge computing architectures, where servers are located at the edge of the Internet near the wireless devices. Therefore, edge computing is a distributed computing paradigm or a model where the processing and storage elements are placed close to the sensing device (data source) for improved latency, throughput, security, and several other benefits. The traditional method sends the sensor data to the cloud for processing and storage. As shown in Figure 1.5, some edge computing nodes do not require the cloud because they are independent. The edge computing server is placed together (in the same housing) with the sensing device for many independent edge computing systems (Figure 1.5).
Figure 1.5 Edge versus cloud computing architecture.
1.2.1 Edge Versus Cloud Computing
Cloud computing, which has been a dominant enabler in enterprise IT delivery in the past decades, offers many key advantages [18]:
An enterprise can avoid the huge capital expenditure (CAPEX) of building a datacenter.
Cloud computing can lower operating expenses (OPEX) by exploiting economies of scale.
Cloud computing can minimize electricity costs by strategically locating servers in regions with the lowest electricity and air conditioning costs.
While cloud computing is a good value proposition for many enterprise applications, it is not a viable option for many time-critical IoE applications. Edge computing is a new paradigm in which substantial computing and storage resources are placed in close proximity to mobile devices or sensors [19–21] (Figure 1.6).
There have been many research activities and industry developments related to edge computing in the past decade:
In early 2013, Nokia and IBM jointly introduced an edge computing platform called the Radio Applications Cloud Server (RACS).
In 2014, a mobile edge computing standardization effort began under the auspices of the European Telecommunications Standards Institute (ETSI).
Figure 1.6 Edge versus cloud network latencies.
In 2015, multiple mobile technology organizations – led by Vodafone, Intel, and Huawei – launched the Open Edge Computing Initiative (OEC).
In late 2015, Cisco, Microsoft, Intel, Dell, and ARM, in partnership with Princeton University, created the OpenFog Consortium.
The origin of edge computing dates back to the 1990s, when Akamai introduced content delivery networks (CDNs) to accelerate web performance. A CDN uses nodes at the edge close to users to pre-fetch and cache web content. These edge nodes can also perform some content customizations, such as adding location-aware advertisements. The emerging edge computing architectures generalize and extend the CDN concept by leveraging cloud computing infrastructure within a 5G network.
1.2.2 Edge Design Options
Broadly speaking, there are two different architectural options for edge computing:
Mobile Edge Computing (MEC):
MEC consists of commercial off-the-shelf (COTS) computers located near the 5G RAN to reduce latency and improve context awareness. It is a standalone architecture that can function without a remote cloud.
Fog Computing (FC):
FC is a decentralized computing infrastructure with fog computing nodes (FCNs) placed between the end devices and a cloud. The FCNs are based on heterogeneous network elements including routers, switches, and gateways. Unlike MEC, FC is tightly linked to a cloud and cannot function without a remote cloud.
Both MEC and FC are computing and storage networks, which are built on top of 5G communications networks. Indeed, we are migrating from a simple 4G/LTE communications network to an integrated computing and communications network (Figure 1.7).
Interestingly, the 5G communications network has two similar layers:
Control Plane
(CP), also known as user plane function (UPF), consists of generic computers running a network operating system and controlling the network domain. The CP controls and manages the user equipment’s access to the network. It performs the access and mobility function (AMF), session management function (SMF), authentication server functions (AUSF), unified data management (UDM), and policy control functions.
Figure 1.7 Edge computing is driving many 5G applications.
Figure 1.8 5G computing and communications networks.
Data Plane
(DP) consists of the configurable network devices within the network domain. The data plane interfaces with the radio access network (RAN) and the data network (DN), and then performs the UPF on the received user data. Therefore, the data plane performs tasks such as packet inspection and processing, routing, and ensuring the quality of service (QoS) (
Figure 1.8
).
In an integrated 5G communications and computing network, MEC and CP can likely share the same servers, while FC and DP can share the same network elements. Nevertheless, each implementation is different, and the mobile network operator (MNO) would determine the network architecture. Table 1.2 compares the two edge computing design options.
1.2.3 Key Benefits of Edge Computing
Edge computing provides the following key benefits:
Responsive Services:
Some mobile applications are very sensitive to the long latency in cloud computing. Applications such as autonomous driving, virtual reality (VR), augmented reality (AR), robotics, and energy grids can work better in the edge computing environment.
Network Scalability:
Edge computing can reduce massive ingress traffic to the cloud, thus improving network scalability. It processes massive raw data locally and can cut up to 99.9% of cloud traffic.
Table 1.2 Comparison of 5G private network deployment architectures.
Deployment architecture
RAN
DP
CP
MP
Perf. grade
Flexibility grade
Security grade
Investment efforts
Management simplicity
Resource efficiency
DA-A
On-prem
On-prem
On-prem
On-prem
Best
Best
Best
Worst
Worst
Worst
DA-B
On-prem
On-prem
On-prem
Remote
Best
Best
Medium
Worst
Medium
Medium
DA-C
On-prem
On-prem
Remote
Remote
Medium
Medium
Medium
Medium
Best
Best
DA-D
On-prem
Remote
Remote
Remote
Worst
Worst
Worst
Best
Best
Best
The table compares four different 5G private network deployment architectures (DA-A, DA-B, DA-C, and DA-D). DA-A is the most expensive and most flexible architecture, and DA-D is the cheapest and the least flexible architecture. DA-B and DA-C are between these two extremes.
Masking Cloud Outages:
Edge servers are independent and can function even without a remote cloud. When cloud outages are short, mobile devices may not even be aware of the outages.
Distributed Security:
Instead of storing all data in a centralized server, critical data are distributed in multiple edge computing servers. This creates a robust distributed security architecture.
1.3 Custom Computing
The 5G technologies have presented many market opportunities and challenges at the same time. Here are the seven most critical challenges (Figure 1.9).
We have discussed about the edge computing architecture in the last several sections. Although the standard edge computing architecture, using software-defined networking (SDN) and network function virtualization (NFV) [22, 23], can be cost-effective, scalable, flexible, and reliable, it may not be able to meet the stringent latency and bandwidth requirements. Thus, custom edge computing architecture can be used to dramatically speed up the 5G performance while maintaining the reconfigurability [24, 25].
1.3.1 Introduction to Custom Computing
Custom computing systems are special computer systems designed for specific applications, such as signal processing or database operations, and can be reconfigurable. These systems are desirable when the general-purpose system has undesirable features such as high hardware resource consumption, high latency, low throughput, and high-power consumption. Custom computing systems need to meet some SWaP requirements, usually employ special acceleration hardware elements, including:
Figure 1.9 Key challenges to fully realize 5G potentials.
Field-Programmable Gate Arrays
(FPGA) chips, such as Xilinx Virtex-7.
Graphic Processing Unit
(GPU) chips, such as NVIDIA Titan V GPU.
System-on-Chip
(SOC), such as Google Tensor Processing Unit (TPU).
Custom computing is gaining popularity among all major cloud providers, thanks to the emergence of machine learning and big data applications. For example,
Amazon EC2 F1 Instances and Alibaba ECS F3 Instances use FPGA-based PCI Express cards running on generic servers to enable the delivery of custom FPGA hardware acceleration.
Amazon EC2 P2/P3/G3/G4 Instances and Alibaba EGS Instances use GPU-based PCI Express cards running on generic servers to enable the delivery of custom GPU acceleration.
By running the applications in hardware instead of software, custom computing can be used to improve their computational performance and reduce power consumption and costs. Unlike software algorithms, hardware implementations in FPGA and GPU are less susceptible to cyberattacks. As a result, custom computing architectures are ideal for implementing security elements, such as network firewalls and intrusion detection systems (IDSs).
1.3.2 5G/6G Security Concerns
Among the seven major challenges in the 5G era, security considerations are arguably the most critical concerns, receiving the most media attention in the last few years. Specifically, the 5G stakeholders are worried about the potential cyberattacks from foreign agents and malicious hackers [26–33].
Thanks to many emerging applications, the 5G networks are susceptible to many new threats. The common attack types include:
Denial of Service (DoS):
Attackers can temporarily disable the network services.
Network Message Modification:
Attackers can temporarily take over the network.
Data Corruption:
Attackers can erase important network data.
Eavesdropping:
Attackers can intercept confidential data from a 5G network.
Fraud:
Attackers can use network services at the expense of another subscriber.
In general, these cyberattacks may be originated from:
Fake base stations.
Rogue IoT devices, rogue base stations, and rogue network elements.
Roaming networks.
The Internet (
Figure 1.10
).
Figure 1.10 Many different ways to attack 5G network.
A DoS attack on a 5G/6G network is not only an inconvenient event but can also wreak havoc in an autonomous driving vehicle or a smart manufacturing factory. Accordingly, it is important to build a reliable 5G/6G network that can:
Prevent cyberattacks.
Detect cyberattacks quickly.
Limit detrimental consequences of cyberattacks.
To prevent cyberattacks, it is useful to use custom computing technology to implement 5G high-performance security firewalls:
GTP firewalls
defend the network against GPRS tunneling protocol (GTP) based attacks. These cyberattacks may be initiated from RAN or roaming networks.
SGi/Gi firewalls
defend the network against TCP/IP attacks from the Internet. These attacks may exploit vulnerabilities in SMTP, NFS, FTP, Telnet, NTP, and HTTP.
Diameter firewalls
defend the network against diameter protocol-based attacks. Diameter protocol contains subscriber information, similar to the traditional SS7 protocol.
To detect cyberattacks, most enterprise networks today may already have IDSs. For example, Splunk IDS uses log data and history to identify network anomalies. However, Splunk IDS typically does not process the log data in real time. Accordingly, it is useful to use custom computing technology to implement 5G real-time IDSs.
To limit the detrimental consequences of a successful cyberattack, it is important not to use a centralized authentication and encryption server. Instead, it is useful to implement a distributed authentication and encryption architectures. These architectures can take advantage of custom computing technologies to implement complicated security algorithms.
Figure 1.11 Employment of custom computing in mobile edge computing and fog computing. In both cases, custom computing can reduce network latency and power dissipation and enhance network bandwidth and digital security.
1.3.3 Custom Edge Computing Cards
Custom edge computing systems can be employed in both mobile edge computing (control plane) and fog computing (user/data plane) (Figure 1.11).
In custom mobile edge computing architectures, one or more acceleration cards are added to COTS servers. These acceleration cards run the original software tasks in hardware, improving the performance by two or three orders of magnitude. At the same time, the power consumption can also be dramatically reduced (Figure 1.12).
In custom fog computing networks, special SOCs and FPGAs are added to the proprietary network computing elements. Oftentimes, the fog computing networks consist of proprietary embedded systems running non-standard operating systems. An emerging architecture is called the “Smart Network Interface” (or SmartNIC). Unlike a standard network interface card, a SmartNIC card may be FPGA based or SOC based and can be reconfigurable. A key feature of SmartNIC is minimizing the network latency (Figure 1.13).
Figure 1.12 An NVIDIA Tesla T4 PCIe card and a Xilinx Alveo U200 PCIe card. In a 5G MEC or CP network, these acceleration cards can be used to increase performance and reduce overall power consumption.
Figure 1.13 Xilinx Alveo U25 SmartNIC acceleration card encompasses network, storage, and compute functions, providing a comprehensive SmartNIC platform.
1.4 Deployment Considerations
According to MTN Consulting, the revenues of MNOs have remained stagnant in the last few years with a 1.6% decline in 2019 even before the onset of the COVID-19 pandemic. On the other hand, building a 5G base station was three times more expensive than building a 4G base station, costing somewhere between US$30K and US$70K in 2019. To make the situation even worse, a typical 5G base station consumes twice or more the power of a 4G base station [34–36].
Construction and operation costs are the most critical barriers to the deployment of 5G network infrastructures. The conundrum is how we can unleash the full 5G potential without breaking the bank or taking excessive risks.
Figure 1.14 5G fixed wireless access (FWA) application.
What is the best way to deploy 5G? It depends on:
Existing infrastructure.
Use cases.
Business models.
For example, some 5G equipment providers, including Ericsson and Huawei, actively promote the 5G fixed wireless access (FWA) market. Here, the household uses a 5G wireless router to provide Wi-Fi and LAN used within the house (Figure 1.14).
In this use case, there are two interesting points:
First, the quality and reliability of the FWA network will depend on both the 5G wireless network and the fiber optic networks. Much of the 5G capital expenses are used to build the fiber optic infrastructure.
Second, FWA is a good value proposition for many rural and sub-urban areas. However, FWA may not work well in some urban areas where fiber-to-the-premise (FTTP) is already commonplace. In urban areas, it makes sense to focus on other mobile applications.
In this section, we examine various 5G network deployment options and how they can affect the 5G network architecture.
1.4.1 5G/6G Cell Architecture
While the cell radius of a 4G network cell tower – called macrocell – varies between 2 and 40 km, a small cell may have a radius of less than 500 m [37–39]. A smaller 5G network cell can support higher traffic densities by taking advantage of:
Higher frequency spectrums, which have shorter transmission ranges.
More effective spectrum reuse across different cells.
Figure 1.15 Small cells versus macrocells.
According to McKinsey’s research, sites with traffic density above 0.5 petabytes per square kilometer per year will require a cell radius of less than 200 m. The traffic densities of many major cities are projected to reach 1–2 petabytes per square kilometer per year.
To leverage the existing infrastructure, most MNOs adopt a two-tier approach by [40, 41]:
Upgrading the macrocells to support massive multi-input multi-output (mMIMO) protocols, with multiple antenna and beamforming technologies.
Building small cells, as needed, within each macrocell (
Figure 1.15
).
These two approaches are complementary and somewhat independent. Several research papers discuss optimizing network resources. In reality, an MNO may decide a certain deployment option based on business considerations only.
There are different types of small cells:
Microcells:
with a radius of less than 2 km – are usually used to provide good wireless coverage within an in-building area, such as a train station or a shopping mall.
Picocells:
with a radius of less than 200 m – are usually used to cover a small area for a temporary event, such as a sporting event.
Femtocells:
with a radius of less than 10 m – are usually installed at home or in office.
Femtocells are often installed and maintained by end users, whereas microcells and picocells may be installed and maintained by either end users or MNOs. In all three cases, a 5G NR small cell must comply with 3GPP TS 38 series NG-RAN gNB specifications.
Here are some characteristics of a small cell:
A small cell is connected to its 5G core network (5GC) by an NG interface, and the small cells are interconnected to each other by an Xn interface.
Unlike a macrocell, a small cell may not support mMIMO with more than 16TX and 16RX.
A small cell may support only 5G NR standards or both 4G/LTE and 5G NR standards.
1.4.2 5G/6G Private Network
In the previous section, we have discussed macrocells and small cells. Specifically, end users may install, maintain, and operate small cells. In this section, we will examine this specific market segment, called 5G private network.
There are a growing number of private 5G networks – aka non-public 5G networks – as we migrate from 4G/LTE to 5G. A private 5G network uses the same technology as public 5G networks but it allows an enterprise to manage its own 5G network to serve a limited geographic area with optimized services using dedicated equipment [42, 43].
A key component of a private 5G network is to provide:
Lower Latency:
In smart manufacturing, a private 5G network could be tailored for application delivery prioritization and to ensure predictable latency in all network traffic.
Higher Security:
A private 5G network could provide cellular-grade security, which can be used to keep sensitive data encrypted, authenticated, protected, and local to the premises.
Wider Coverage:
Many remote areas are not sufficiently covered by public cellular infrastructure. A private 5G network could be installed to provide wide coverage in agricultural fields, container ports, warehouses, or deep inside buildings.
Private 5G network use cases include:
A smart factory can take advantage of a private 5G network to support services, such as predictive diagnostics, workforce management, machine automation, and factory safety management.
A complex warehouse that fulfills pick-and-pack customer orders using robots and control software. A private 5G network is useful to help meet strict performance requirements needed to control countless fast-moving robots from a single base station.
A large mine can use a private 5G network to support a complex heterogeneous system, comprising in-pit CCTV monitoring, intelligent earth sensing, telemetry, and many other monitoring services. This can help improve the reliability and safety of the mine.
Figure 1.16 Deployment architectures A, B, C, and D.
There are many design alternatives in a private 5G network. Generally speaking, there are two main design aspects: deployment architecture and operation model [44–46]. Each design aspect has a few design options.
5G network architecture encompasses four main blocks:
Management Plane
(MP), aka management and orchestration layer, consists of network software managing a network domain.
Control Plane
(CP) consists of generic computers running a network operating system and controlling the network domain.
Data Plane
(DP), also known as the User Plane Function (UPF), consists of the configurable network devices within the network domain.
Radio Access Network (RAN) consists of legacy and 5G NR (New Radio) interfaces supporting mMIMO and beamforming technologies (
Figure 1.16
).
In the most flexible private 5G network architecture (Deployment Architecture A), the above four blocks are all on-premise. While this deployment architecture is very secure and provides the best performance, it is also the most expensive. On the other extreme (Deployment Architecture D), only the RAN block is on-premise. The security level is the worst but the investment is the minimum (Table 1.2).
Figure 1.17 Different 5G operation models.
A private 5G network may adopt a certain operation model, independent from the deployment architecture discussed previously.
First, the network infrastructure resources may be owned by a single user and used exclusively for a particular vertical application or are shared by multiple users and used for multiple vertical applications.
