Virtualizing 5g And Beyond 5g Mobile Networks Larry J Horner

Virtualizing 5g And Beyond 5g Mobile Networks
Larry J Horner download
https://ebookbell.com/product/virtualizing-5g-and-
beyond-5g-mobile-networks-larry-j-horner-50168732
Explore and download more ebooks at ebookbell.com

Here are some recommended products that we believe you will be
interested in. You can click the link to download.
Virtualizing Hadoop How To Install Deploy And Optimize Hadoop In A
Virtualized Architecture Vmware Press Technology George Trujillo
Charles Kim Steve Jones Rommel Garcia Justin Mu
https://ebookbell.com/product/virtualizing-hadoop-how-to-install-
deploy-and-optimize-hadoop-in-a-virtualized-architecture-vmware-press-
technology-george-trujillo-charles-kim-steve-jones-rommel-garcia-
justin-mu-55290680
Virtualizing Microsoft Tier 1 Applications With Vmware Vsphere 4
Charles A Windom
https://ebookbell.com/product/virtualizing-microsoft-
tier-1-applications-with-vmware-vsphere-4-charles-a-windom-4313866
Virtualizing Sql Server With Vmware Doing It Right Michael Corey
https://ebookbell.com/product/virtualizing-sql-server-with-vmware-
doing-it-right-michael-corey-5412114
Virtualizing An Infrastructure With System P And Linux Ibm Redbooks
https://ebookbell.com/product/virtualizing-an-infrastructure-with-
system-p-and-linux-ibm-redbooks-1397606

Virtualizing Desktops Apps With Windows Server 2012 R2 Inside Out
Svidergol
https://ebookbell.com/product/virtualizing-desktops-apps-with-windows-
server-2012-r2-inside-out-svidergol-6760786
Visualizing Social Science Research Johannes Wheeldon And Mauri
Ahlberg
https://ebookbell.com/product/visualizing-social-science-research-
johannes-wheeldon-and-mauri-ahlberg-45803170
Visualizing The Street New Practices Of Documenting Navigating And
Imagining The City Cities And Cultures Pedram Dibazar
https://ebookbell.com/product/visualizing-the-street-new-practices-of-
documenting-navigating-and-imagining-the-city-cities-and-cultures-
pedram-dibazar-46412278
Visualizing Taste Ai Hisano
https://ebookbell.com/product/visualizing-taste-ai-hisano-46768140
Visualizing Complexity Modular Information Design Handbook Darjan Hil
https://ebookbell.com/product/visualizing-complexity-modular-
information-design-handbook-darjan-hil-47187754

Include bar code
ISBN 13: 978-1-63081-930-9
B O S T O N I L O N D O N
A R T E C H H O U S E
www.artechhouse.com
3165 C
7702 C
The fifth generation (5G) mobile network brings significant new capacity and
opportunity to network operators while also creating new challenges and additional
pressure to build and operate networks differently. Transformation to 5G mobile
networks creates the opportunity to virtualize significant portions of the radio access
network (RAN) and network core, allowing operators to better compete with over-the-
top and hyperscaler offerings.This book covers key concepts of virtualization that will
solve problems of operational and support considerations, development and lifecycle
management, and vendor and team dynamics when deploying virtualized mobile
networks. Geared toward mobile network engineers and telecom professionals, the
book demonstrates the benefits of network virtualization, enabling operators to better
address the ever-increasing traffic load on their network while maintaining costs and
bringing increased agility to both their operations and business offerings.
Larry J. Horner is a principal engineer and senior solution architect at Intel. He is a
Life Senior Member of the IEEE, member of the board of the local Communication
Society, North America Region 5 ComSoc representative, and general cochair for the
International IEEE NFV SDN conference.
Kurt Tutschku is a professor for telecommunication systems at the Blekinge Institute
of Technology (BTH). He is leading BTH’s team on secure and distributed systems
(SDS). He served as one of the general cochairs of the IEEE Conference on NFV-SDN
from 2017 to 2022.
Andrea Fumagalli is a professor of electrical and computer engineering at the
University of Texas at Dallas (UTD). His research interests include aspects of wireless,
optical, Internet of Things (IoT), and cloud networks, and related protocol design and
performance evaluation.
ShunmugaPriya Ramanathan is pursuing her doctoral degree in 5G network function
virtualization at the University of Texas at Dallas (UTD). Her research focuses on the
performance evaluation of various open-source reliability schemas for the virtualized
5G RAN.
MOBILE COMMUNICATIONS
VIRTUALIZING
5G
AND
BEYOND
5G
MOBILE
NETWORKS
Larry J. Horner
Kurt Tutschku
Andrea Fumagalli
ShunmugaPriya Ramanathan
Horner
•
Tutschku
Fumagalli
•
Ramanathan
VIRTUALIZING 5G AND
BEYOND 5G MOBILE
NETWORKS

Virtualizing 5G and
Beyond 5G Mobile Networks

For a complete listing of titles in the
Artech House Mobile Communications Library,
turn to the back of this book.

Virtualizing 5G and
Beyond 5G Mobile Networks
Larry J. Horner
Kurt Tutschku
Andrea Fumagalli
ShunmugaPriya Ramanathan

Library of Congress Cataloging-in-Publication Data
A catalog record for this book is available from the U.S. Library of Congress.
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library.
Cover design by Andy Meaden Creative
ISBN 13: 978-1-63081-930-9
Cover image courtesy of Adobe Stock/Blue Planet Studio
© 2023 ARTECH HOUSE
685 Canton Street
Norwood, MA 02062
All rights reserved. Printed and bound in the United States of America. No part of this book
may be reproduced or utilized in any form or by any means, electronic or mechanical, including
photocopying, recording, or by any information storage and retrieval system, without permission
in writing from the publisher.
All terms mentioned in this book that are known to be trademarks or service marks have been
appropriately capitalized. Artech House cannot attest to the accuracy of this information. Use of
a term in this book should not be regarded as affecting the validity of any trademark or service
mark.
10 9 8 7 6 5 4 3 2 1

To Maggie, Tiago, and Eli, the future is what you make it, all the best
—LJH
To Beate, Lorenz, and Mathis, your love carries me
—KTT
To Daura, Lorella, and Tommaso, the lights of my life
—AF
To Nanda, Naren, and Shreeya, the loves of my life
—SR

vii
Contents
Acknowledgments xix
Part I
Fundamentals of Virtualization in Communication
Service Provider Networks 1
1 Virtualizing of the 5G Radio Access and Core Network 3
1.1 Introduction to Virtualizing the Mobile Network 3
1.1.1 The Beginning of Network Function Virtualization 3
1.2 Expanding on the First Vision of Virtualization 6
1.3 Breaking Down the Fundamentals Driving Virtualization 7
1.4 Applying This Discussion to the Mobile Radio Network 8
1.5 Transforming the Mobile Network One G at a Time 9
1.6 Evolving Small Steps on the Gs 12
1.7 Which Network Is This Exactly? 13
1.8 Acronyms and Domain-Specific Terms Abound 14
1.9 Telecom Providers Go by Many Names 14

viii Virtualizing 5G and Beyond 5G Mobile Networks
1.10 Addressing the Various Audiences 15
1.11 To Those New to This Industry 16
1.12 Structure of the Remaining Chapters 16
1.12.1 The Fundamentals: Chapters 1–5 16
1.12.2 Engineering of Virtualized 5G and B5G Systems:
Chapters 6–11 18
1.12.3 Future Developments: Chapters 12–14 20
1.12.4 Acronyms and Terms 20
References 20
2 Benefits of NFV for 5G and B5G Networks and
Standards Bodies 23
2.1 Why Use NFV for Networks? 23
2.1.1 Transformation of a Large Legacy Business Is Difficult 23
2.2 The Existing NEP Ecosystem of Vendors 24
2.3 Changing Business Models Midstream 25
2.4 Independent Software Vendors as NEPs 26
2.5 Green-Field Entrants into the CSP Business 26
2.6 Transformation from Hardware-Centric to Software-
Centric Networks 27
2.6.1 Data Traffic Dominates the Network 27
2.6.2 There Is a Fixed Cost to Moving Bits 28
2.6.3 A Tale of Two Models 29
2.7 Applying the Cloud Model to the Telco 30
2.8 Paths Taken to Evolve the Telco Network 32
2.8.1 3G Data Begins to Be the Primary Content in the
Network 32
2.8.2 Interfaces Connecting Endpoints in the Network 32
2.9 The Ever-Evolving Introduction of Technology into
the Network 33
2.9.1 Making the Network Global 33
2.9.2 This Global Network Comes at a High Cost 34
2.9.3 Relating This Back to the 5G Network 35

Contents ix
2.10 The Drive for Improved Agility and Efficiency 36
2.10.1 DevOps and Continuous Integration and Continuous
Delivery 36
2.11 Separation between Data Plane and Control Plane 37
2.11.1 The 5G User Plane Function and Data Network 38
2.11.2 5G Standalone and Non-Standalone Deployments 39
2.12 3GPP as the Leading Standard Body for the Mobile
Network 40
2.13 Introducing the International Telecommunication Union 41
2.14 Other Standards Bodies 42
2.15 Open RAN’s Role in Virtualizing 5G 43
2.16 Venture Capital Investments 44
2.17 Summary 45
References 46
3 Virtualization Concepts for Networks 49
3.1 The Virtualization of the Network 49
3.1.1 What Is Virtualization? 50
3.2 Managing the Virtual Resources: Resource Control and
Efficiency 51
3.3 A Brief History of Virtualization Concepts 52
3.4 Virtualization Through the Ages 53
3.4.1 The Early Years: Computer and OS Virtualization 53
3.4.2 The Second Decade of Virtualization: Virtualization
Leaves the Research Labs 55
3.4.3 Smaller Computers Join the Fray 56
3.4.4 Processes Start Talking to Each Other 56
3.4.5 Democratizing Computing in the 1980s 57
3.4.6 1990s: Universality and Independence 58
3.4.7 2000: The Era of Hardware Efficiency 58
3.4.8 2010: Control Efficiency 59
3.5 Cloud Computing 60
3.5.1 1970–1980: The Embryonic Phase 60

x Virtualizing 5G and Beyond 5G Mobile Networks
3.5.2 1990: Distributed and Bundling 61
3.5.3 2000: The Cloud Becomes a Commercial Offering 61
3.5.4 2010s: Control, Automation, Orchestration, and
Application Engineering 62
3.6 Network Virtualization 63
3.6.1 1960–Mid-1980: Roots and Programmability of
Distributed Computing 63
3.6.2 Mid-1980–2000: The Internet Boom 64
3.6.3 2000–2005: Powerful Application Overlays and
Ossification of the Internet 64
3.6.4 2005–2010: Network Virtualization and Network Slices 65
3.6.5 2010: Programmability of the Network 66
3.7 Basic Objects and Data Structures for Network
Virtualization 67
3.7.1 Network Topology 68
3.7.2 Addressing 68
3.7.3 Routing 68
3.7.4 Resource Management 69
3.8 Summary 69
References 69
4 Data Plane Virtualization and Programmability for
Mobile Networks 71
4.1 Data Plane Acceleration with OpenFlow and P4 71
4.1.1 Context for Acceleration 71
4.2 OpenFlow 74
4.2.1 Flows 75
4.2.2 Configuration 75
4.2.3 System Model and Pipeline 75
4.2.4 Ports 76
4.2.5 Group, Meters, and Counters 77
4.2.6 Forwarding Abstraction 77
4.2.7 Instructions and Actions 80
4.2.8 Header and Match Fields 81
4.2.9 Examples for Matching Headers 81
4.2.10 OpenFlow Protocol 81
4.2.11 Distributed Controllers and Flow Visor 83
4.2.12 Evaluation of the OpenFlow Concept 85

Contents xi
4.2.13 The Importance of OpenFlow in 5G 86
4.3 P4 86
4.3.1 Domain-Specific Programmability 87
4.3.2 The P4 Language 87
4.3.3 P4 Concept 88
4.3.4 Data Plane Forwarding and P4 Enhancements 89
4.3.5 Portable Switch Architecture 90
4.3.6 Programming a P4 Device 90
4.3.7 The P4 Language 92
4.3.8 P4 Runtime Architecture 96
4.3.9 Evaluation of P4 96
4.4 Conclusion 97
References 97
5 Performance of Infrastructures for Virtual Network
Functions 99
5.1 Performance and Security Considerations 99
5.1.1 Virtualization Modes and Requirements 100
5.1.2 Sharing, Aggregation, and Emulation in Virtualization 100
5.2 Performance Evaluation Concepts for the Sharing of
Resources 103
5.2.1 Networking Scenario 103
5.2.2 Mathematical Concept 104
5.2.3 Mathematics Model 105
5.2.4 A More Realistic Description of the Impact 106
5.2.5 Smallest Timescale and Timescale Analysis 108
5.2.6 Capabilities and Conclusion 110
5.3 Performance Evaluation Concepts for the Aggregation
of Resources 113
5.3.1 Foundations 113
5.4 CPU Pinning 116
5.5 Non-Uniform Memory Access 119
5.6 Conclusion 123
References 123

xii Virtualizing 5G and Beyond 5G Mobile Networks
Part II
Engineering of Virtualized 5G and B5G Systems 125
6 Transforming and Disaggregation in 5G and B5G
Networks 127
6.1 The Transforming and Disaggregation of the Network 127
6.1.1 Challenges to Transforming the Telco Network 128
6.2 DevOps: A Method to Improve System Management 129
6.3 Telco DevOps 131
6.4 Transforming the Operations in the Network 133
6.5 Rolling out 5G in the Network 135
6.5.1 5G Non-Standalone and Standalone Considerations 135
6.6 Private LTE and Private 5G 137
6.7 The Cost of 4G and 5G Is Changing 138
6.7.1 Regulatory Considerations 139
6.8 Security in the Disaggregated Network 140
6.9 Transforming Operations: A Use Case Example 141
6.10 Beyond 5G Market Drivers 141
References 142
7 Designing Virtualized RAN 143
7.1 Virtualizing the 5G RAN 143
7.1.1 It All Begins with the Standards 144
7.1.2 Operating Systems of Choice 145
7.1.3 Supplementation of the OS 146
7.2 The Continuing Evolution of the Standards 147
7.3 Attaching the UE to a Network 148
7.3.1 The Roaming UE 150
7.3.2 The UE Detailed Signaling Flow 150
7.4 Initialization of the DU to CU Connection 152

Contents xiii
7.4.1 Back to the UE Attachment 153
7.5 The 80/20 Rule 153
7.6 Splitting the RAN: Revisited 154
7.6.1 FEC Processing and More in the RAN 154
7.7 Enhanced Common Public Radio Interface: The
Fronthaul Interface Transformation 159
7.8 Summary 162
References 163
8 vRAN Performance Engineering 165
8.1 Network Performance Engineering 165
8.1.1 5G Drivers 165
8.1.2 5G Usage Scenarios 165
8.1.3 5G Spectrum Bands 167
8.2 5G Functional Split 167
8.2.1 5G Functional Split Origin 167
8.2.2 eCPRI 168
8.2.3 Functional Split Options 169
8.2.4 Functional Splits Trade-Off 169
8.2.5 How to Select and Additional Functional Split Options 170
8.3 5G Deployment Options: SA and NSA Architecture 172
8.3.1 SA and NSA Deployment Options 173
8.3.2 Technical and Cost Comparison 174
8.3.3 Migration Path from 4G LTE to 5G 176
8.4 5G Roadmap 178
8.4.1 3GPP Release of 5G NR 178
8.4.2 5G Services in North America 179
8.4.3 4G-5G Interworking Architecture 180
8.4.4 User Plane and Control Plane Deployment
Considerations 182
8.5 Key Challenges in 5G Rollout 183
8.5.1 System Security 183
8.5.2 Service Performance and Availability 184
References 185

xiv Virtualizing 5G and Beyond 5G Mobile Networks
9 Building the vRAN Business: Technologies and
Economical Concerns for a Virtualized Radio
Access Network 187
9.1 What Are the Costs and Opportunities of 5G? 187
9.2 The 5G Business Outcome 189
9.3 New Models to Address the TCO 191
9.4 The oRAN Model Introduces a RAN Intelligent
Controller 192
9.5 Features of the One-Socket Server 196
9.6 Open Source Remains a Critical Element to the
Virtualization Effort 197
9.6.1 Open-Source Community in the RAN 197
9.7 Asymmetry in 5G and the Previous Gs 197
9.8 5G Market Drivers in Asia 198
9.9 Business Considerations of Virtualization 199
9.10 Pro and Cons of White Boxes, Which Are Truly
SHVSs, in the vRAN 199
9.11 Bright Boxes: Standard High-Volume Servers with
One or Two Customized Features 200
References 201
10 Designing Virtualized 5G Networks 203
10.1 Successfully Designing Virtualized 5G Networks 203
10.1.1 What Is Success for a Virtual System Design? 204
10.1.2 Overall Aim 204
10.1.3 Efficient Virtualization 204
10.1.4 Separation and Portability 205
10.1.5 Open-Source Software 205
10.2 Open-Source Software for 5G 206
10.2.1 Why Open-Source Software? 207
10.2.2 Flexibility and Agility 207
10.2.3 Speed of Development and Deployment 207

Contents xv
10.2.4 Low Licensing Efforts 208
10.2.5 Cost-Effectiveness 208
10.2.6 Ability to Start Small 209
10.2.7 Software Security 209
10.2.8 Shared Maintenance Costs 210
10.2.9 Enabling Future Development and Attract Better Talent 210
10.3 5G Open-Source Efforts 211
10.3.1 Open-Source 5G Core Network Elements 211
10.4 Design and Performance Criteria for Virtualized 5G
Systems 211
10.4.1 Computer Systems and Software Engineering
Concepts for Virtualized 5G Systems 212
10.5 Computer Systems and Software Engineering
Concepts for 5G Functions 213
10.6 Performance Criteria for 5G Systems 215
10.6.1 Scenarios and KPIs 217
10.7 Summary 218
References 218
11 Scaling Disaggregated vRANs  221
11.1 The Disaggregated vRAN 221
11.1.1 RAN Disaggregation 221
11.2 RAN Intelligent Controller Overview 223
11.2.1 Interfaces 223
11.2.2 RIC Design Principles and Components 225
11.2.3 Policy Guidance 225
11.2.4 ML/AI Role in the RIC 226
11.3 Security Challenges 227
11.3.1 Key Security Threats 227
11.3.2 Key Security Pillars 229
11.4 5G Resiliency 233
11.4.1 Network Resiliency 233
11.4.2 VNF Resiliency 234
11.4.3 Dynamic Rerouting with Live Migration Support 235
References 236

xvi Virtualizing 5G and Beyond 5G Mobile Networks
Part III
Future Developments in the Mobile Network 239
12 Private 5G Networks and the Edge 241
12.1 The Privatization of the Network with p5G 241
12.1.1 Usage Scenario and Objectives 242
12.1.2 Service Objectives and Attributes for Private 5G 243
12.2 Technology Overview 244
12.2.1 Deployment Scenarios 245
12.3 Multiaccess Edge Computing and Private 5G Systems 252
12.3.1 MEC Overview 252
12.3.2 MEC Architecture Elements 254
12.3.3 Future MEC Solutions for Private 5G Systems 254
12.4 Business Issues with Private 5G and MEC Systems 256
12.4.1 Enabling Private 5G Benefits for Applications 257
12.4.2 SIM, eSIM, iSIM 259
12.4.3 MEC and Hyperscalers at the Edge 259
12.5 Summary 260
References 261
13 Open-Source Software Development and
Experimental Activities 265
13.1 Introduction 265
13.2 5G Open-Source Software Packages 265
13.2.1 Open-Source 5G Core Network Elements 266
13.2.2 Open-Source Evolved Packet Core 267
13.2.3 Open-Source Radio Access Network Elements 268
13.2.4 Open SDR Devices 269
13.2.5 Open-Source Control and Orchestration 270
13.3 5G Experimental Networks for US-EU Collaboration 270
13.3.1 POWDER 271
13.3.2 Colosseum 271
13.3.3 COSMOS 272
13.3.4 AERPAW 272
13.3.5 NITOS 273
13.3.6 R2lab 273

Contents xvii
13.3.7 Open Experimental Sites in 5G-EVE 274
13.3.8 Open Experimental Sites in 5GENESIS 277
13.3.9 Open Experimental Sites in 5G-VINNI 278
13.4 Summary 280
References 282
14 Summary of Virtualization of 5G and Beyond 285
14.1 Where It All Began 285
14.2 New Markets 289
14.3 6G Is on the Horizon 290
14.4 Summary of Some Key Factors 291
14.4.1 A Cloudy Crystal Ball 292
14.5 Conclusion 293
14.5.1 Possible Research Areas 293
References 294
Glossary of Acronyms and Common Terms 295
About the Authors 305
Index 309

xix
Acknowledgments
No effort of this magnitude is accomplished alone, nor can all those who
contributed be properly thanked to the extent they deserve, yet some indi-
viduals merit special recognition. Apologizes in advance for any oversight that
have been made in omission. First, to my coauthors, Andrea, Kurt, and Priya,
thanks, we made it, and hopefully have created a friendship that will remain
well into the future. We must also thank the reviewers, editors, and members
of the staff at Artech House: Natalie, Isabel, and some whom remain unknown
to us. Thanks for granting us the opportunity to put some of our work down
into words; without you this would simply remain scattered thoughts in papers,
slide decks, and side conversations over the course of our careers. I must also
thank my family for their encouragement and support during the creation of
this work. Finally, I need to thank by name Elizabeth A. Q. for her efforts, en-
couragement, and design work, your name is equal to those found on the title
page in my view.
Kurt’s work on this book was funded partly by the Knowledge Foun-
dation, Sweden, through the Human-Centered Intelligent Realities (HINTS)
Profile Project (contract 20220068).
Larry J. Horner

Part I
Fundamentals of Virtualization in
Communication Service Provider
Networks

3
1
Virtualizing of the 5G Radio Access and
Core Network
1.1 Introduction to Virtualizing the Mobile Network
Network virtualization has evolved as the preferred method for realizing the ef-
ficient and future-proof design, engineering, and operation of communication
networks. To those new to this topic there is significant expansion needed in
this opening statement, which will be provided shortly. For those already well
versed in this topic we hope to provide additional insights into both the busi-
ness and technical aspects of the current and future work guiding the ongoing
transformation underway in this field.
1.1.1 The Beginning of Network Function Virtualization
Before getting too deep, a short review of how the term and vision of network
virtualization came to being.
A team of authors from 13 companies [1] published a white paper titled
“Network Function Virtualization, an Introduction, Benefits, Enablers, Chal-
lenges & Call for Action.” This relatively short paper, numbering only 16 pages,
was first published October 22–24, 2012, at the SDN and OpenFlow World
Congress held in Darmstadt, Germany. Those not familiar with this white pa-
per and interested in this topic would be well served to have a quick read of
this foundational paper. The authors gifted the industry with both a framework
to transform the way networks are built and operated and new terminology in

4 Virtualizing 5G and Beyond 5G Mobile Networks
which to discuss the nature of the transformation. Prior to this white paper, the
terms network function virtualization (NFV) and virtualized network func-
tions (VNFs) had not been coined or used in the context of transforming the
communication network. A note of caution: new members of this community
may overload or mix NFV and VNF, they are not the same but are related to
each other. NFV represents a concept where a network function has been dis-
aggregated (virtualized) from the underlying infrastructure, whereas a VNF is
a specific realization of a network function in a virtualized environment. The
authors of the 2012 paper paid special attention to position the NFV concept
alongside another technology—software-defined networking (SDN)—that had
recently been gaining favor. The concept of SDN today is often overshadowed
in the context of NFV discussions, which may be viewed as simply a consolida-
tion of the two concepts into one, although technically they remain separate
concepts. SDN is the separation of a traditional control function from the data
plane function on an appliance. For example, with switches or routers in the
classical model the control function runs on the same systems where the links
are terminated. With SDN, new routing or switching decisions are managed
by a centralized controller, and the routing or switching tables in the router or
switch are updated via an application programming interface (API) between
these nodes. The IEEE currently maintains the term NFV-SDN in their confer-
ence proceedings [2]. Significant work has been underway by leading network
operators and by the vendors supplying equipment into the network in the en-
suing years to implement the vision laid out in the 2012 NFV paper. The origi-
nal authors were also very sensitive to the need to include the existing ecosystem
solution providers, the telecommunication equipment manufacturers (TEMs),
also called network equipment providers (NEPs), in their overall strategy to
transform the industry. As one should fully expect in any field of technology,
there has been some evolution and expansion in the procedures employed over
time. This is the case with NFV in building the interesting workloads that re-
alize the network. (There are a large number, sometimes well over a hundred
processes, running on a modern system in the network. Many are important
and critical but benign, placing little or no load on the system, and a smaller
number perform the majority of the functional processing that realized the core
intent of the service. The latter are the “interesting workload.”) Nevertheless,
the underlying concept laid out in the paper has proven transformative to the
network and to the vendors delivering products into that network.
Figure 1.1 is from that now-famous 2012 white paper. It shows the con-
ceptual separation of the various layers of the NFV infrastructure (NFVi) (of-
ten called a middleware layer of software), virtualization, compute, storage and
networking, and the element management system (EMS), which is a nod to the
TEMs, on the left, and the management and network orchestration (MANO)
on the right. The EMS is one area where time has challenged the original view;

Virtualizing of the 5G Radio Access and Core Network 5
the legacy EMS was often a vendor-specific separate system or set of systems
that was used to monitor and manage the network functional elements. This
would include collecting performance and log or alarm data from the integrated
system. Today these EMS functions rarely are part of the NFV platform and in-
stead reside on other nodes in the network, either with MANO functions or on
other elements. The MANO too has evolved from the original and continues
to evolve today to meet the needs of the operators and developer community.
Since the time of writing, there has been significant development work
and additional abstraction to this design. Today we find the basic model is well
applied and significant progress has been made to the underlying concept of
separating the functionality of hardware from software for solutions in the tele-
communications industry, which is currently called disaggregation.
Today the idea of disaggregating the network functions from the underly-
ing hardware to achieve the first stated goal found in the white paper, which is
“reduce equipment cost and reduce power consumption,” is having a positive
impact on the capital expenditure (CapEx) (the cost of adding assets or adding
value to existing assets; these expenses are depreciated over time) and opera-
tional expenditure (OpEx) (the costs incurred to run and operate a business)
for the network operators. The NFV paper also emphasized a goal of reducing
the time to market (sometimes called agility), both for bringing existing ser-
vices into scope and introducing possibly new features and services. The goal
Figure 1.1 Introducing NFV for the first time.

of continuously improving the speed and agility of the network operations re-
mains as much a focus today as it was when NFV was first envisioned. It is com-
mon for the disaggregation of the network functions to be spoken of in terms
of virtualization. Indeed, it is fair to consider these terms synonymous in many
cases; the intended goal was the separation and then later recombination of the
underlying hardware and the upper layer application along with the supporting
middle layers and associated system integration.
1.2 Expanding on the First Vision of Virtualization
There will be more on the evolution of virtualization in the communication
network in later chapters, but for now we return to the opening statement. To
begin the expansion of the opening statement, let’s dissect this from the right to
the left, starting with what a communication network is. Generally, the com-
munication network should be considered the global network that allows the
ubiquitous exchange of data in multiple forms in near real time. If we accept
this, then the commonly well understood internet is a subset of the commu-
nication network. One view shown in the center of Figure 1.2 provides an ab-
straction of the communication network. In the center, the telecommunication
network reaches close to the edge where the users (consumers and enterprises)
connect. This reach may take several physical forms, which include both over
the air and those with physical media (e.g., fiber or metal wires). There are other
Figure 1.2 The global network, shown in the center.

networks that we will not be discussing, which may be multinational in scope
or controlled by national entities. For simplicity, the network we are discussing
here is the global network that is still capable today of connecting two people
anywhere on the globe who have a phone number and/or broadband connectiv-
ity. This of course is not the full extent of the capabilities of the telecommunica-
tions network, but rather a foundational element that is still supported today.
This network consists of both wired and wireless segments, and while
both are undergoing the transformation to virtualization it is the mobile net-
work (also known as the wireless network) that has proven to be the leading
candidate in this space. The global communication network is made up of well
over 600 mobile network operators. These operators are defined here as those
that hold a license to use mobile network spectrum and offer services that rely
on the utilization of that spectrum as a part of their business. There are an even
larger number of network operators that may not have license for spectrum,
and others that use the spectrum of the licensee. This global communication
network is comprised of the equipment that allows both people and enterprises
(e.g., any entity or thing that is not an individual will be referred to as an en-
terprise, which includes different types of governmental and nongovernmental
users) to exchange data over a common infrastructure. The variety of methods
of data exchange, the rates, and the format is large, and often the network
doesn’t care about the format. The network that realizes the mobile portion of
this global network originally evolved from the legacy phone network where
the data exchanged was founded in the requirement to exchange human voice
in full duplex (e.g., both parties could be talking simultaneously). The network
continues to support this fundamental capability today, the ability to carry a
voice call in full duplex. The network today carries a significantly greater vol-
ume of data that is not voice calls; nevertheless, they are still an important
aspect of the design considerations. Additionally, the ability to carry voice calls
and the service uptime may be heavily regulated by the license granting author-
ity (e.g., the ability to call emergency services over the network).
1.3 Breaking Down the Fundamentals Driving Virtualization
Returning to the opening statement, the engineering and operation of the com-
munication network touches both the CapEx and OpEx of the network. CapEx
in this case is the accumulation of the cost paid by the operator to purchase
and install all the materials associated with building the network infrastruc-
ture. This includes the purchase cost of any systems (hardware) and in many
cases the initial licensing of any associated software and may also include the
cost associated with installing (“racking and stacking”) of the systems into a
specific site. The OpEx cost is the recurring costs of running the network, this

would include the cost of power (power to operate the systems and the cost of
cooling) and any direct labor cost associated with maintaining the operation
of the systems in the network, along with any recurring hardware or software
maintenance fees. The vendor of the hardware and/or software often includes
some level of ongoing support, also known as maintenance, and on an annual
basis requires the operator to remit payment for the next year’s maintenance.
There is a new category that is starting to emerge in some markets and by some
vendors that is solely based on collecting an ongoing maintenance fee, deferring
the initial and usually larger CapEx, and relying solely on an OpEx model of
continuously recurring usage fees. In the case of CapEx, the operators can often
take investment credits on their tax obligations in some jurisdictions where the
OpEx often does not qualify for this accounting. While both expense types
represent a significant expense to the operators, we often see models where the
trade-off of one expense verses the other can change the rate at which an opera-
tor is able to introduce new technology into their network. This case arose with
the introduction of 5G and will be discussed later when the topic of standalone
and non-standalone 5G core is discussed.
By some estimates, the total cost to operate a modern system in today’s
network is nearly evenly divided into thirds. Unfortunately, data of this nature
is closely protected by the operators, so our estimates are based on the distil-
lation of several conversations protected under nondisclosure agreements be-
tween global communication service providers (CSPs) and one of the authors.
The first third is the CapEx to bring the function into the network. The
second third is the power to operate and cool the systems during their opera-
tional life. The final third is the operator’s labor and recurring maintenance fees
associated with the solution. One third of the total cost of ownership (TCO)
is CapEx and the other two thirds of the TCO are OpEx. There are a few ge-
ographies where abundant hydroelectric power may have a significant impact
in reducing the energy cost in the OpEx portion, but these are rare, and one
might think of areas vast in land and sparse in human population to find ex-
amples where this holds true. In addition, it is estimated (again, usually a closely
guarded metric) that some network operators with both wireless and wireline
networks allocate 70% of their CapEx to the mobile network, and as a result
a fair share of the following OpEx. With this foundational understanding it
should be clear why these network operators would be very interested in pursu-
ing the goal of the 2012 white paper as pointed out in the opening.
1.4 Applying This Discussion to the Mobile Radio Network
The foundation of applying this discussion to the virtualization of the 5G and
Beyond 5G (B5G) mobile network brings the opportunity to introduce to those

new to this subject the latest on the work as well as to both the technology and
the industry. For those well versed in either the current technology and/or the
business motivation, we hope to stimulate additional work on the transforma-
tion ongoing in this global network.
1.5 Transforming the Mobile Network One G at a Time
One very common trend when discussing the mobile network is to describe
the steps in terms of the Gs (generations), 1G, 2G, 3G, 4G, 5G, and Beyond
5G. Here 1G is the first generation, which was not called 1G initially—readers
with long memories may still refer to this as Advanced Mobile Phone System
(AMPS). Here we simply give it the first counting number for convenience and
find little reason to define or discuss 1G except to simply reference it as the
starting point for the network we have today.
This progress of Gs is shown in Figure 1.3.
The caution here is that while the industry speaks of the Gs incrementally,
there have also been many sub-Gs along the way, as shown in Figure 1.4, and
network operators often have multiple variants active in their network simulta-
neously, so that in many cases the instance of more than one G can be realized
on the same hardware with other Gs at the same time. This has sometimes been
described as “stretching the metal” around multiple G specifications at the same
Figure 1.3 The Gs of the mobile network.

time (the computer chassis are made of metal, and if one collection of these
computers are running, for example, 4G and 5G on the same systems, we’ve
stretched the metal across 4 and 5G). Said differently, for example, a Mobile
Management Entity (MME) function (MME is an element found in the 4G
network specification) in 4G may be realized on the same computer hardware
as the access mobility management function (AMF) for the 5G network (for
now MME and AMF can be thought of as an application workload on a com-
puter; there will be more to say about these functions in later chapters). One
may presume simply from the expansion of the acronym that these two might
have some commonality in function and in fact they do, and this is often the
case from one G to the next; there is an evolution of functionality and often a
name change in what are otherwise similar modules. With 5G we find better
separation of functions that allows for the virtualization (NFV and SDN) de-
sign considerations to be realized in the software.
5G has three well defined enhancements above the capabilities found in
4G, which are shown in Figure 1.5: enhanced mobile broadband (eMBB or
EMBB), ultra-reliable low-latency communications (URLLC), and massive
machine-type communications (mMTC). The terms eMBB, URLLC, and
mMTC are preferred, but from time to time we find that some authors may
choose to use all caps to align with the URLLC format. 5G was specified to
support use cases where massive bandwidth and very-low-power devices could
Figure 1.4 Evolution of the mobile network versions.

coexist in one technology, but not necessarily on the same user equipment (UE)
device. Design goals speak of up to 5 years or more of battery life for some de-
vices, and, at least on paper, up to 10 Gbps download speed for other devices;
the 10 Gbps of 5G is 100× the “on paper” 4G download speed. Additional de-
sign considerations in the 5G specifications allow for the separation of network
user traffic and management functions in addition to the ability to migrate
from 4G to 5G while reusing portions of the 4G network. 5G also introduced
concepts allowing traffic steering where not all traffic would receive the same
treatment. Some traffic may be processed closer to the source, often called the
edge, while other traffic from the same origination point might be transported
deeply into the network. Previous generations of the Gs required that all data
be treated exactly the same way. This opens 5G to wider use cases beyond the
consumer space, and when coupled with innovative spectrum licenses in some
jurisdictions, has the potential to open new markets and drive innovation.
The reality is that the evolution of the G has taken several smaller steps
from time to time, a more realistic view of this stepwise evolution. Many his-
torical network operators have been running mobile networks since at least the
2G era, and the reality is that once hardware systems have been deployed in the
Figure 1.5 5G feature drivers.

network, for example at the start of the 2G rollout, these systems might not be
capable of running the 3G workloads. The operators and the supplier ecosys-
tem then had two approaches when 3G rolled around and the 2G network was
(and in many places still is) operational. Like the sides of a coin choosing heads
over tails or tails over heads, there were simply two options to consider without
clear advantages one way or the other: continuing to operate the 2G network
equipment and bring in new equipment to operate the 3G network, or when
bringing in the 3G network equipment ensuring that it was also capable of run-
ning the 2G workloads.
For example, choosing to deploy separate systems for 3G would increase
the number of computers that must be managed in the network and may re-
quire more rack space in the installed location, and also possibly require more
power and interface ports. But while upgrading the system to a single platform
that could support both 2G and 3G may bring some advantages in space and
power savings while reducing the number of nodes that need to be managed,
it would possibly incur increased hardware cost. In this case, the operator may
have already purchased the 2G hardware and licenses and might have to repur-
chase some of the compute resources as a minimum.
1.6 Evolving Small Steps on the Gs
There are several points of interest when comparing the differences in Figures
1.3 and 1.4. First, the introduction of the next G does not terminate the previ-
ous G in the network, noting that 2G appears to have a longer life than 3G. At
the time of writing, some network operators are terminating their 3G and will
likely repurpose the spectrum for usage in 5G deployments. Briefly, a mention
of where spectrum comes from may be required; this is a portion of the elec-
tromagnetic spectrum known as the radio frequency (RF) portion. Each nation
claims sovereignty to this limited resource and in many cases issues licenses for
the use of specific bands or portions of the spectrum. Often these licenses are
obtained for a fee. Just how much of a fee depends; for example, according to a
recent article in IEEE Spectrum, the total value of the spectrum auction in the
United States over the past 30 years was about 230 billion USD for all licenses.
This includes not only the mobile network spectrum but also other licenses as
well, including TV, satellite, classical radio, and others. Clearly, spectrum has
high value for both the licensing jurisdiction and to the licensee [3].
A keen eye looking at Figure 1.4 will notice that 2G lingers past the 3G
timeline. Why keep 2G around longer than 3G? 2G has a number of advan-
tages to some legacy enterprise use cases and while the typical consumer may
no longer be a 2G user, it is likely that it will be many more years before all
usage of 2G expires. Regulators may also play a role in the decommissioning

of a particular service; operators were not able to just turn down their 3G and
abandon subscribers who had not upgraded their devices. In many cases regula-
tors required operators to provide compatible devices to these 3G consumers at
no cost to the consumer. In some jurisdictions, portions of the CSP network,
if not all if it, are considered critical infrastructure and as such, regulators have
significant say about how a business can be operated and they take regulatory
compliance very seriously.
There are a number of sub-Gs as shown in Figure 1.4, such as 4.5G and
4.9G. In some cases, equipment installed at the beginning of a G’s timeline may
not be capable of running the last iteration of the latest G. Consider the case
when two Gs are intended to be run together on one set of systems prior to the
actual turnup of the latest G. A set of systems in the network currently running
3G, for example, may have to be upgraded prior to the introduction of 4G into
the network if these systems are intended to support both 3G and 4G. As time
passes and new subvariants of 4G emerge, it can be possible that this equipment
is incapable for a number of reasons from running the latest iteration of 4G.
In this case the operator has to either install new equipment for the latest sub
of 4G, or forgo that upgrade and risk that competitors will outcompete them.
For example, systems installed in the late 2008/9 time frame to support both
3.95G and the soon-to-be-turned-on 4G may not have been capable of run-
ning the 4.9G release in the late 2010s. This is the crux of the problem network
operators were facing—these purpose-built network computers were constantly
requiring upgrades to support the current technology in the network and posi-
tioned for the next release. This, and even more significantly the ever-increasing
traffic load on the telco network, required that the operators take a new look
at how they build, deploy, and operate their networks, hence NFV. Telco, for
those unfamiliar, is a nickname for a telecommunication service provider, and is
often used to refer to a class of service providers today, providing at a minimum
connectivity into the wider global communication network. Telecom is often
used in a wider sense than telco, and is used to refer to the wider ecosystem
rather than just to the service providers themselves (e.g., a telco is an operator,
telecom is an industry and includes more than just the operators themselves).
1.7 Which Network Is This Exactly?
Some readers will already know that the terms “telecom” and “communica-
tion networks” are sometimes used interchangeably. This is intentional because
“telecom” can be a noun for an industry and “communication networks” can be
a qualified noun for the physical realization that enables a business to operate in
this domain. Webster’s [4] provides some insight here. While it is expected that

the concept of a network is understood by anyone interested or involved in this
area, new entrants into this field may benefit from a more thorough treatment.
1.8 Acronyms and Domain-Specific Terms Abound
Telcos, CSPs (also known as CoSPs by some silicon venders in their literature),
network operators, operators, carriers, service providers, multiple systems op-
erators (MSOs, also known as cable operators), and a few other terms are used
as names for the same thing. While we will strive to be consistent in our usage
in this book, note that different regions of the globe use different terms for this
industry and for the companies that are generating revenue from the use of the
communication network. Furthermore, even within some communities, the
names are used interchangeably. As an additional note of caution, some enti-
ties have chosen to abbreviate this even further: the acronym CSP is common
in some press and publications covering this industry, while at least one large
influential silicon manufacturer prefers CoSP, reserving CSP for cloud service
providers (also known as hyperscalers like Alibaba, Amazon, Azure, Google,
etc.). We will continue to use CSP to mean communication service provider, as
it is more generally accepted at this time. At times, however, other terms for the
telcos or CSPs may be used, such as “operators,” where either the context and
flow fit the usage better, or where they can serve as a reminder of the variations
that will be encountered in the industry at large. There is an entire ecosystem of
solution providers into this industry that are also known by various names and
associated acronyms that are used interchangeably, such as NEPs and TEMs
(NEP is generally preferred but TEM still can appear in the literature). To de-
termine the context of the three-letter acronym (TLA) or term in use, the reader
may from time to time have to consider the scope of the material. One can
consider learning the various overloaded synonyms that will be encountered as
part of the criteria for entry into this field.
1.9 Telecom Providers Go by Many Names
Terms like mobile network operator (MNO) and mobile virtual network op-
erator (MVNO) will also be encountered in the literature, presentations, and
conversations. Today, most MNOs also have wireline business segments, and
our hope is that this term begins to fade as we view the network as a unified en-
tity. One last note of caution in the naming conventions—the V in MVNO is
not “virtualization,” which refers to the business of reselling possibly rebranded
service on an existing operator’s network. In some markets, MVNOs evolved
to provide a lower-cost service possibly based on a prepaid subscription model,
while in others it was market brand driven; for example, standing up a wholly-

owned subsidiary with a younger generational name and possibly a different
level of service than found on a premium-branded offering.
These discussions will primarily center around the businesses and tech-
nology that have evolved from the legacy wired communication network. There
are two other domains of existing networks that are now intersecting with what
can be called the historical mobile network operators: the cable and satellite op-
erators. The intent is to include them in the discussion on green-field entrants
into the mobile operator space rather than attempt to include the evolution of
their networks to grow into the traditional telco space. This is not to discount
their evolution or their contribution to the global communication network,
but to confine the overall volume of this work to a manageable size. These new
entrants into the mobile communication network as defined here have evolved
their networks with different technologies and with a different cadence than
that of the traditional mobile network operator. While the underlying services
have similar capabilities today (namely the transportation in near real time of
end-to-end data including voice), the names of the nodes, the protocols used,
and the vendors supplying equipment into their business sector are different.
Their entry into the MNO space will very likely see their network elements (at
least for the mobile portion) align with the standards bodies and ecosystems we
will be discussing. The final new entrants as discussed above are those coming
in without a legacy network of any kind. We have seen three to date, one in
India entering initially as a 4G provider, one in Japan entering as a late 4G and
early 5G provider, and one in the United States coming from broadcast satellite.
Each of these entrants came into the space with an initial vision to virtualize
as much of their first network as possible at the time, bringing the cost and ef-
ficiency savings envisioned in the 2012 white paper from the start.
1.10 Addressing the Various Audiences
This book is intended to aid new entrants into this domain (e.g., students),
practicing engineers (those developing, designing, building, and operating
networks), members of business teams working in this domain, and academ-
ic researchers. It will provide fundamental and theoretical insights as well as
hands-on discussion for these practitioners. For those new to the communica-
tion network domain, some introduction to this industry is presented in early
chapters. A seasoned veteran of this industry will find some sections useful and
possibly gain fresh insight. The review of where this industry has evolved from
may aid in seeing the most likely vector toward the future. As we have shown,
this industry is rife with TLAs. In addition, there are other abbreviated terms
and borrowed concepts that are deeply rooted in the day-to-day language used
and at first this might be a distraction to the newly initiated. To understand

the full context of the material being discussed, every attempt will be made to
provide context and clarity when a new term or concept is introduced. Not all
concepts or terms will be fully defined here due to the limited space available.
See the discussion in the Glossary of Acronyms and Common Terms for more
on this subject.
1.11 To Those New to This Industry
Welcome to telecom! This industry is very much like an ancient forest—there is
much to explore, some of it very old and in good condition, some possibly in a
more declining state, and even new life if you are willing to have a deeper look.
Virtualization represents some of the new life in this old forest.
Our goal is to introduce the reader to the technical and business funda-
mentals driving the virtualization of 5G and B5G in mobile networks.
1.12 Structure of the Remaining Chapters
This book is organized into three major sections with multiple chapters in each
section. Chapters 1–5 cover the fundamentals, Chapters 6–11 cover engineer-
ing considerations, and Chapters 12–14 cover future developments. The book
ends with a list of acronyms and nomeclature used in the field.
1.12.1 The Fundamentals: Chapters 1–5
The remainder of this chapter continues with the structure of the book and
provides guidance on the usage of the subsequent chapters.
Chapter 2 discusses the benefits of NFV for 5G and B5G networks. We
begin with a holistic and techno-economic discussion of the anticipated ben-
efits when using virtualization techniques in mobile networks. We address what
the technical and business case is that motivates the transformation that com-
prises NFV, and why is this being discussed in the 5G use case and B5G? The
discussion of virtualization technologies begins with a discussion on the accel-
eration of packets in the data plane using SDN techniques.
The historical fundamentals of virtualization are presented briefly, lead-
ing to a discussion on how the performance of virtual mobile networks might
be achieved and eventually differ from nonvirtualized networks, and how to
measure a successful design of virtual 5B/B5G networks. The chapter concludes
with a discussion on the business impact and increasing interest of private 5G/
B5G networks for industrial users and outlines their implementation and ap-
plication pathways.

This chapter will provide the reader with a solid foundation on under-
standing the why in addition to the what and how, which will be covered in the
subsequent chapters.
Chapter 3 discusses the current state and practices in virtualization of the
CSP network. This includes coverage on the interesting workloads in the net-
work. The control plane and data plane are discussed along with the their chal-
lenges; for example, the control plane is stateful, and the data plane has signifi-
cant performance requirements (e.g., throughput and latency). To address the
needs for performance, acceleration of the data packets is required, which takes
the conversation to field-programmable gate arrays (FPGAs), smart network in-
terface cards (NICs), and new languages such as P4 for smart switch hardware.
Also covered are data collection and analysis to measure and monitor the
server performance metrics.
This includes virtualization of the CSP workloads, core network design,
control plane virtualization, data plane virtualization, and packet acceleration.
Within packet acceleration both hardware and software acceleration techniques
are covered, and includes use of FPGAs, eASICs, and programmable ASICs.
eASIC is a new class of ASIC with some limited ability to be configured after
initial manufacturing (e.g., at power up). One key point is the ever-present
concern over the power budget (both in terms of kilowatts-hours consumed
and the actual cost). The role that power consumption plays in the network is
extremely important, both technically (e.g., dissipating heat) and economically
(e.g., energy cost).
Chapter 4 explains that data plane virtualization consumes the most sig-
nificant portions of the operational system and there are techniques and trade-
offs that must be considered to maximize the performance and minimize the
cost, both in capital hardware and in the power consumed in operating the
systems. This chapter covers the current implementation options available to
meet the demands of the 5G network. One technique is to use SDN to bring
control and flexibility to the network operations.
Hardware acceleration remains a critical component in the network to-
day. This can be realized in several ways: NICs, smart NICs, separate FPGA
cards, programmable ASICs (also known as eASICs), and modern switching
fabrics that are programmable using the P4 language. The chapter concludes
with a detailed discussion on P4. Specialized NIC interface software techniques
and packet acceleration in the core with techniques like DPDK (today it’s just
DPDK, but expanded it was Data Plane Development Kit [5], single root I/O
virtualization (SRIOV) [6], and eXpress Data Path (XDP) [7]) are covered in
later chapters.
Chapter 5: Performance remains a key topic in virtualization. Drivers in-
clude how TCO is minimized while providing a consistent SLA. In this chapter

the key performance measurements are mathematically defined and how they
are determined in a virtualized network element are considered. CPU pinning
(ensuring a fixed workload is anchored to a fixed CPU core) is discussed. �Non-
uniform memory access (NUMA) awareness both at the server and within a
CPU are also covered. In addition, how this might differ from a cloud workload
is considered. Developers and users of the network require deep insight into the
performance of the systems. This requires instrumentation that enables detailed
measurements of the various elements of the system, from the hardware up to
the application layer and may include system controls that are aware of work-
load placement. This is important because CPU pinning (e.g., the ability to
specify exactly which core in a die a threat is required to be executed on is fixed
and it cannot be moved or rescheduled to another core on the CPU) can have
implications to network traffic engineering that requires NUMA-aware [8].
1.12.2 Engineering of Virtualized 5G and B5G Systems: Chapters 6–11
Chapter 6: Many mobile operators today have significant investments in 4G
network infrastructure, and the introduction of 5G in many cases will overlay
with existing 4G networks for the foreseeable future. In some cases, even the
nodes that are deployed will support both 4G and 5G functionality in the same
system. Transforming these systems to a pure virtualized network will take con-
siderable effort from both the NEPs and the telcos.
Nearly every operator today has legacy networks and thus teams of en-
gineers practicing in this space that use or have built existing internal systems.
Thus, there is internal competition that bring what some have described as
“institutional antibodies” who resist any change.
Chapter 7 goes into the design and implementation of the modern virtual-
ized RAN, in particular the 5G RAN. The standards from 3GPP [9], the main
standards body, and the ORAN alliance [10] and vRAN design principles will
be covered in detail. Splits (partitioning) in the design will also be covered in
this chapter. There are a number of possible considerations when building the
virtualized RAN, one of which is the choice of splits that may be deployed.
There is no one design, but rather a large number of possible configurations
driven by deployment considerations. Some of these are based on geographical
considerations, while others may be driven by power and space considerations.
This chapter will investigate the various implementation requirements in depth.
Chapter 8 will show how the metal is stretched around the specifications
found in the standards. We have already introduced the concept of stretching
the metal, and we expand on that in this chapter. Stretching the metal refers
to two things. First, the computers that realize the various functions in the
network are cased in metal chassis and the silicon devices (fixed function or

programmable) are constructed using metal. Thus, both the computer hard-
ware and chassis have metal stretched around them. Second, it is an expres-
sion of the combination of functions from various generations (e.g., 4G and
5G functionality) realized in hardware or software that in some cases may be
found executing on the same system. Thus, we have stretched the functionality
of a system to include multiple functions and possibly even within the same
generation.
While much of the discussion in this work spans the 5G network, not all
networks are pure 5G. In some cases, an analysis of components reveals that real
networks today may support 4G, 5, and Wireline all wrapped into one. Many
network operators and network equipment providers have 4G functionality
running within the same software and on the same hardware as 5G elements.
This is important for several technical and business reasons. In this chapter we
will look at the desire to have standalone and non-standalone functionality in
the network, or, what can be referred to as “how the metal is stretched around
the specifications.”
Chapter 9 expands on the work in the previous chapter and breaks the
RAN into bite-sized pieces. The design and implementation considerations of
the various splits will be discussed in terms of both technical and business con-
cerns. The focus on the disaggregation of the modern RAN and how design
splits are realized are discussed. The chapters also examines business consider-
ations in virtualization as it implies changing the business model of how the
networks are built and operated.
Chapter 10 addresses the question “How does the industry measure suc-
cess of virtualization?” There are multiple cost drivers. One is the cost of devel-
oping and maintaining software. Another is the OpEx associated with operat-
ing the network (e.g., how many people are required to maintain the mobile
network). To address the first cost driver, the chapter gives particular attention
to open-source software (OSSW), its capability to reuse software development
and maintenance efforts, and available 5G. The chapter also includes a discus-
sion of performance and operational criteria for 5G systems.
Chapter 11 looks at how operating the network at scale is a significant
concern to mobile network operators. To tackle and manage the growing com-
plexity with scaling, it is critical to automate the process of deploying, optimiz-
ing, and operating the RAN while also taking full advantage of newly available
data-driven technologies (for example artificial intelligence (AI) and machine
learning (ML)) to ultimately improve the end-user quality of experience. The
RAN intelligent controller (RIC) is an area of significant modernization in to-
day’s RAN. This chapter will discuss disaggregation, RIC, security, support, and
the lifecycle management of the nodes and operational considerations of the
virtualized network. One key consideration in that operation and scale includes
reconfigurability and migration of virtualized modules.

1.12.3 Future Developments: Chapters 12–14
This part provides a more forward-looking view of what might be coming in
the industry as we continue to roll out 5G and begin considering the implica-
tions of what comes next.
Chapter 12 details how 5G brings a new market potential to the telcos.
One possible area is in private networks for enterprises—a potentially signifi-
cant opportunity for revenue generation for the telcos. This chapter will cover
the technical and business opportunity at the “the edge” where one finds the
private networks. One significant challenge here for the CSPs is to expand past
their legacy view of providing connectivity and security to the enterprise, and
pivot to providing domain-specific computing at the edge. The current trend is
that the services of private networking are being led by the NEPs, the systems
integrators (SIs), and the enterprises themselves, with the CSPs currently pro-
viding only 20% of the overall private network deployments.
Chapter 13: The scientific community at large has responded to the many
challenges of designing and developing high-performance 5G technologies in
two ways. On the one hand, it has focused on the development of open-source
3GPP software implementations and deployment frameworks that can be effec-
tively used to quickly build 5G proof-of-concept (PoC) designs. On the other
hand, it has pursued the construction of publicly available open experimental
labs that offer researchers time-shared access to various 3GPP technology ser-
vices. These ongoing efforts are reviewed in this chapter.
Chapter 14 offers final thoughts on the current and possibly future state
of the effort to disaggregate the operational mobile network. In addition, new
challenges that remain unaddressed, continued packet acceleration capabilities,
better power management, improved cloud native operational considerations
are put forth.
1.12.4 Acronyms and Terms
This section lists many of the acronyms that are part of the common language—
the alphabet soup—found in the communication industry. There is a very long
list of TLAs and terms that have been abbreviated or used extensively in the
industry, and by no means is this an attempt to list all that may be encountered.
A more complete list can be found in Telecom Dictionary, which is still available
from some sources [11, 12].
References
[1] https://portal.etsi.org/NFV/NFV_White_Paper.pdf.

[2] IEEE Conference on Network Function Virtualization and Software Defined Networks,
https://nfvsdn2022.ieee-nfvsdn.org.
[3] “Billionaires Battle for Global Spectrum,” IEEE Spectrum, October 2022.
[4] https://www.merriam-webster.com/dictionary/telecom.
[5] https://software.intel.com/content/www/us/en/develop/topics/networking/dpdk.html.
[6] https://docs.microsoft.com/en-us/windows-hardware/drivers/network/
overview-of-single-root-i-o-virtualization--sr-iov.
[7] https://www.kernel.org/doc/html/latest/networking/af_xdp.html.
[8] https://en.wikipedia.org/wiki/Non-uniform_memory_access.
[9] https://www.3gpp.org/about-3gpp/about-3gpp.
[10] https://www.o-ran.org/.
[11] https://www.barnesandnoble.com/w/newtons-telecom-dictionary-harry-newton/
1119676342.
[12] http://www.telecomdictionary.com/index.asp.

23
2
Benefits of NFV for 5G and B5G
Networks and Standards Bodies
2.1 Why Use NFV for Networks?
This chapter explores the business and technical question: Why is the industry
changing the way the networks are built and operated today, and in doing so,
contrasting this new approach with the previous approaches? The technology
that exists today transforms the legacy structures to a large extent, and this can
improve the business model the telcos use to build and operate their networks.
An understanding of the “why” in addition to the “what” and “how” establishes
a firm foundation for the following chapters. Also covered here is the role of
several significant standards bodies in contributing technical requirements that
allow for the global interoperation of the mobile network we know today.
2.1.1 Transformation of a Large Legacy Business Is Difficult
There are significant areas that slow the transformation of the legacy commu-
nication service provider community. These areas lie in the vendor community
and within the telcos themselves. One is inertia both in the existing ecosystem
and in the telcos and a second is the difficulty of innovation in the industry.
New entrants into the industry face significant technical and financial barriers
to entry: long development times, longer procurement cycles, and strong ven-
dor relationships that go back decades in some cases with the telco customers.
Nevertheless, there are existing examples where new entrants have penetrated

the market, bringing innovation and new business models to the industry.
There have also been new entrants into the service provider space themselves,
and while rare, they have proven that the industry can innovate both in how
the network is built and operated and how they brought disruptive vendors into
the industry.
2.2 The Existing NEP Ecosystem of Vendors
The existing ecosystem of the TEMs (also known as NEPs) comprises a very
limited group of very large companies (see Figure 2.1).
While Figure 2.1 gives the total revenue for each for a few years back, the
significance is that with many above 20 billion USD in annual sales these by
no means are small companies. It should also be pointed out that several are
well over 100 years old (in order of their founding: Corning 1851, Nokia 1865,
Ericsson 1867, NEC 1899, Fujitsu 1935, Cisco 1984, Qualcomm 1985, ZTE
1985, and Huawei 1987).
The revenue profile of each of these companies is not within the same
domain; for example, Corning has a large fiber and glass product line and
Figure 2.1 Revenue per top CSP supplier (NEP/TEM).

Benefits of NFV for 5G and B5G Networks and Standards Bodies 25
pharmaceuticals along with other business areas that are not replicated by any
of the other eight. Qualcomm likewise has a user equipment (UE, e.g., a mobile
handset) modem business unit supplying the largest share of handset modem
chips. One might also ask why companies such as Samsung and Apple are not
included. This table was generated from a query of the top revenue-generating
companies in telecom equipment manufacturing. The other two, while very
large indeed, do not generate as much revenue in supplying equipment and
services into the network operators directly. We will see some of the Apple and
Samsung impacts represented shortly.
These ecosystem partners in some cases have been building and deploying
equipment and solutions into this industry for over a century. Their business
has grown to include the ability to not only build the equipment that is con-
sumed by the network operator, but to also perform systems engineering and
operation of the system once installed. In addition, some of the TEMs provide
complete lifecycle management of the systems once installed. This work would
include providing and applying software upgrades and if a piece of hardware
fails the TEM is engaged to replace it. Software upgrades may include enhance-
ments that cure defects, close security holes, or introduce new features devel-
oped by the TEM. These business relationships are normally covered under
maintenance contracts where the CSP (or CommSP, also known as the telco)
is paying an annual fee to the TEM for this service in addition to the normal
ongoing software license and hardware maintenance costs.
2.3 Changing Business Models Midstream
One of the transformation challenges lies in the size and organization of these
large and complex companies. This of course is the challenge many technolo-
gies companies face—how to transform an existing business model and remain
viable during the process. A prime example is the former U.S. company East-
man Kodak (Kodak) founded in 1892. We stepped outside the communication
field so as not to cause too much disagreement here. For those unfamiliar, Ko-
dak for most of the 1900s was the dominant photographic film company. Ko-
dak developed the first self-contained digital camera, and yet in 2012 filed for
bankruptcy. Today it is a shadow of its former self. Kodak could not transform
from selling chemicals and film quickly enough when the digital transforma-
tion hit them. This is not an attempt to predict any future in the CSP industry,
but rather to point out that technology does not wait for market trends to be
favorable. Disruption forces can exist inside a company and there still may not
be enough forces to counteract internal resistance. One colleague calls this be-
ing attacked by the institutional antibodies.

One of the companies on the list above, Cisco, is a counterstudy in their
ability to transform and direct resources and thrive as a result. Cisco in the late
1990s and prior to the internet bubble burst in March 2000 used a model of
acquiring a number of promising startups and quickly integrating their tech-
nology and staff into the Cisco products and culture. This method allowed
Cisco to direct their research and development (R&D) funds to their existing
products and markets while nascent technologies were developed and tested out
by startups, with Cisco then acquiring the startups at just the right time for the
Voice over IP (VoIP) era.
2.4 Independent Software Vendors as NEPs
Today there are a number of growing independent software vendors (ISVs) that
have had some market success in the CSP space, and yet there is over little over-
all investment in new startups targeting CSP networks and operations and the
mobile network in particular. One key reason is the very long lead times to de-
velop product and existing vendor lock-in with CSPs. There are, however, three
large-scale new entrants in the mobile 4G/5G space. These completely new
entrants as service providers in major markets in the past decade are Reliance
Jio in India, Rakuten in Japan, and Dish in the United States. We introduced
the term “green field” in Chapter 1; for those unfamiliar, green field refers to a
new entrant into a business or industry that has no legacy customers or legacy
infrastructure to support [1]. Green field can be easier to transform (both hard-
ware systems and staff), as one is starting without any existing technology, staff,
or procedures that need to be supported and transformed.
2.5 Green-Field Entrants into the CSP Business
Jio stood up a 4G network without any previous generations to support and
brought an innovative market-winning approach to attract customers in their
first few years of operation. This was accomplished with a significant invest-
ment that in two years grew from zero to over 130 million subscribers.
Rakuten entered the market in Japan with a green-field 5G network built
with the largest proportion of virtualized network possible at the time, and has
subsequently acquired one of their RAN providers, Altiostar. Rakuten remains
an interesting transformative force as they have the ability to license their tech-
nology and 5G solution to other CSPs.
There is a third entity that is quasi green field. Dish Networks is a satel-
lite service provider that is turning up a terrestrial 5G network, and while Dish
has some capabilities that might be associated with a traditional CSP, we would

view them as also coming in as green field because they have no land-based
legacy or prior Gs to support.
There are several more modestly sized ISVs developing and delivering
solutions into the CSP 4G and 5G RAN today, two of which were acquired
by Microsoft—Affirmed Networks and Metaswitch in 2019. The hyperscalers
clearly have interest in the CSP markets and are developing and acquiring tech-
nology and staff knowledge in this space. Their largest barrier today is access to
spectrum licenses. Spectrum licenses will be discussed in more detail later.
2.6 Transformation from Hardware-Centric to Software-Centric
Networks
Returning for the moment to companies in Figure 2.1, a decade or so ago each
of these, possibly with the exception of Corning, were vendors of purpose-built
hardware solutions. A CSP would procure a particular solution from the ven-
dor, and the systems that realized that solution would have been branded with
the logo or name of the vendor. Today, while some still offer physical devices
(servers possibly with enhanced capabilities), many also offer at least a portion
of their solution as software only that has been validated to perform well and is
supported on specific servers from original equipment or design manufacturers
(OxMs; the x means either equipment or design). These would include brand
names like Dell, HPE, SuperMicro, Silicom, and Lenovo, and are more com-
monly known as industry standard high-volume servers or white boxes. This
transformation, in some small part, has been due to the pull from the CSPs as a
direct result of the virtualization envisioned in the 2012 white paper mentioned
in Chapter 1.
Let’s return to economic drivers. One way to consider this is that for
every bit that moves through the network there is a finite cost borne by the
operator in equipment, power, and support costs. Since the introduction of the
first Apple iPhone in 2007 the profile of the traffic through the network has
changed dramatically, and as a direct result so have the economics of building
and operating the network.
2.6.1 Data Traffic Dominates the Network
Figure 2.2 shows the growth in the CSP network from 2001 to 2014. This in-
terval is chosen to clearly show the impact that the introduction of the iPhone
and subsequent mobile devices has had on the usage pattern of the network.
Prior to 2007 the majority of the traffic through the network was voice and
the CSPs were able to monetize the use of the network by charging for voice
minutes irrespective of the origination or termination on their own or another
network. As long as one leg of the call was one of their subscribers, they were

able to charge for the use. Just two years after the introduction of the iPhone,
data usage on the network surpassed voice traffic and the race was on. Today,
voice traffic through the network is barely measurable and some services carry
voice as data and counted it as data traffic. Today we still call them phones, but
the meaning as a sound-transmitting device is nearly lost in this context.
2.6.2 There Is a Fixed Cost to Moving Bits
There is a crossover point where it is estimated that the cost to build and oper-
ate a CSP network exceeds the revenue generated from the users of the network.
This assessment is based on the “fixed cost to move a bit” analysis. When the
cost to grow or expand the network is linear and the growth in the demand
(capacity of traffic through the network) is greater than linear, there is a point
where the cost to grow and operate the network crosses over the revenue curve.
This would push a commercial enterprise into a losing position financially.
A very closely guarded financial analysis by some CSPs estimates that in
the near future this crossover point—to expand the network capacity and to de-
ploy new 5G nodes—may be reached. To prevent this from occurring there are
only a few options: stop expanding the network, increase the fees charged to us-
ers at a rate at least equal to the demand growth rate, find new revenue sources,
or change the cost curve associated with expanding and operating the network.
There is always going to be a fixed cost to move a bit through any given
network and the revenue to sustain that movement of traffic must be borne by
the subscribers of the network, either directly or indirectly.
Figure 2.2 Exponential growth of data in the CSP network from 2001 to 2014.

It was with these realities in mind that individuals from the R&D arms
of some of the leading CSPs met in 2012 to discuss how virtualization could be
used to address the problem. This resulted in publication of the famous NFV
white paper that was mentioned in Chapter 1.
The vision was simple to articulate but by no means easy to achieve. Hav-
ing a clear understanding that the hyperscalers (cloud providers such as Micro-
soft, Google, Facebook, Amazon, Alibaba, and Tencent, etc.) had already faced
this problem the CSPs agreed that their technologies should be leveraged. The
hyperscalers had already decoupled their data center compute and networking
from purpose-built systems and migrated to standard high-volume servers with
virtualization. In fact, some of them had developed much of the necessary sup-
porting software. Furthermore, in addition to transforming the cost model to
build the network, they had transformed the model of providing the support
and lifecycle management of their data centers. This transformation allowed
them to apply a �development and operations (DevOps) approach to manag-
ing the systems. This model significantly transformed the number of support
engineers necessary to maintain the systems. By some estimates the hyperscalers
were nearly two orders of magnitude more efficient than the CSPs in this area.
Where a CSP engineer may on average maintain a few hundred systems, the hy-
perscalers would manage tens of thousands. In addition, the hyperscalers could
upgrade software and apply patches nearly continuously, whereas a CSP may be
spending months on applying a single upgrade to the systems in their network.
This difference is significant if you consider the OpEx expense implications.
However, and fundamentally, the regulatory model that the CSPs operate un-
der is significantly different than that of the hyperscalers that informs how they
operate and how they introduce new technology. For example, a CSP may have
a significant financial penalty from a regulator or customer for failing to meet a
SLA, failing to complete a call for emergency services that could lead to injury
or loss of life, or billing errors that result in customers being overcharged. This
in many cases drives their operational behaviors.
2.6.3 A Tale of Two Models
Table 2.1 contains a small sample of data from the end-of-year 2020 financial
reports of two cloud and telco companies. In considering some of what may be
implied in Table 2.1 note that this is not exactly an equal-to-equal comparison
across companies for several reasons, one being there is no indication of the
number of systems under management or the amount of power consumed to
operate these systems. One nevertheless might see relationships that provide
insights that can be used to argue that hyperscalers may be more efficient than
CSPs for reasons worthwhile to explore.

Random documents with unrelated
content Scribd suggests to you:

96. PURPLE ASTER
Once started it spreads very rapidly along the highways or in
cultivated areas.
The plant is suspected to be toxic to livestock, but is rarely eaten. It
is used in the treatment of asthma in humans. In addition, external
use is made of it to relieve the irritation caused by Poison Ivy.
The plants have probably migrated to this area from the central
plains states.
Curlycup Gumweed
Sunflower Family
Aster sp.

97. FLEABANE
The Asters and Fleabanes are sometimes confused, but they can
generally be recognized by the difference in the number of ray
flowers. That is, Asters have only about half as many ray flowers as
do the Fleabanes.
Species of Purple Aster form an important part of the late summer
floral display at Cedar Breaks and Bryce Canyon. They come on after
the Lupine, Columbine, Indian Paintbrush and other early bloomers
have faded.
Purple Aster
Sunflower Family
Erigeron speciosus

63
98. MOUNTAIN SUNFLOWER
There are numerous species of Fleabane in this region. Some
particular kind may be found at any time of the growing season, for
certain species bloom very early and others continue late in autumn.
Some species of Fleabane grow in dense masses and, in early
spring, carpet the meadows and roadsides. The ray flowers of the
Fleabanes are generally twice as numerous per head as are the ray
flowers of the Asters. The plants are quite similar in other respects.
Fleabane
Sunflower Family
Helianthella uniflora
Sunflowers abound in these parks during the late summer. As the
early summer flowers, mostly in blues and purples, fade, the yellow

99. COMMON SUNFLOWER
and red flowers come on. This is especially true in the alpine
meadows of Cedar Breaks in August when the predominant species
are the Sunflowers.
The seeds of the Sunflowers supply abundant food for many birds
and small mammals.
Mountain Sunflower
Sunflower Family
Helianthus annuus
The very large flowers of these plants sometimes present a colorful
display as they take over the roadways or abandoned fields.
Members of this group are generally considered weeds because of
their habit of crowding out the more desirable species. Certain

100. GAILLARDIA
species of Sunflowers are now being developed for commercial
purposes and produce oil for cooking and meal for livestock feed.
Indians in some areas of North America cultivated sunflowers for
food and for trade.
Common Sunflower
Sunflower Family
Gaillardia parryi
This handsome and conspicuous plant is found growing in the
Sonoran Zones of Zion. It has a slender, rough stalk, about a foot
tall, and stiff, rather hairy, dull-green leaves growing mostly from the
root. The beautiful flowers, about three inches across, have golden-
yellow rays which are three pointed. The center of the flower is a
shaded maroon and yellow, very velvety, and becoming an attractive

65
101. DESERT MARIGOLD
fuzzy, round, purplish head when the rays drop off. This plant
blooms mostly in May and June.
Gaillardia
Sunflower Family
Baileya multiradiata
These golden-yellow flowers, measuring about three inches across,
are fairly common along the trails and roadways of Zion Canyon and
in other low-elevation areas of the park. They bloom during May and
June. The ray flowers become bleached and papery as they mature,
thus accounting for the name Paper Daisy. This attractive composite
is also known as Desert Baileya. In California this plant is cultivated
for the flower trade. It has been found poisonous to sheep, although
horses crop the flower heads, apparently without harmful effects.

102. WESTERN YARROW
Desert Marigold
Sunflower Family
Achillea lanulosa
Western Yarrow is very widespread and can be found in much of
Western America. It is more common at elevations above 5,000 feet
in these areas. It grows to be 12 to 20 inches high, and the fernlike
leaves, which have a pungent odor when crushed, and the umbrella-
shaped clusters of flowers are characteristics of this plant that help
to identify it.
Since ancient times the plant has been highly regarded for its
healing properties. Legend ascribed the discovery of this virtue to
Achilles, in whose honor the plant is named.

103. HEARTLEAF ARNICA
Western Yarrow
Sunflower Family
Arnica cordifolia
A common flower in the Pine and Spruce forest of Cedar Breaks and
the alpine areas of Zion and Bryce Canyon is the Heartleaf Arnica.
The yellow ray flowers are few, while the disk or central flowers of
the flower head are numerous. The flowers measure about three
inches across and are often mistaken for sunflowers. The heart-
shaped leaves help distinguish this flower from its close relatives.
Tincture of arnica is obtained from certain species of Arnica.

67
104. MEADOW SALSIFY
Heartleaf Arnica
Sunflower Family
Tragopogon pratensis
This interesting plant, known also as Oyster Plant, has been
naturalized from Europe and is now quite common in the West. It
has a smooth, stout hollow stem about 2 feet tall, rather dark-green,
smooth leaves clasping at the base, and handsome flowers from 2 to
4 inches across. The flowers open early in the morning, closing at
midday, to remain closed until the next morning.
Meadow Salsify is most commonly found along the roadways or in
other places where the native soil has been disturbed.
See Figure 105 for description of the seed of this flower.

105. MEADOW SALSIFY
Meadow Salsify
Sunflower Family
Tragopogon pratensis
This habitat view of the Meadow Salsify illustrates one of the
important ways in which plants scatter their seeds about. As the
flower matures into seeds in a conspicuous and very large,
dandelion-like head, each seed is equipped with a perfect parachute
of silky fibers. Winds often sweep these flight-equipped seeds for
many miles and result in wide dissemination of this species, which
was introduced into this country from Europe not very long ago.
The seeds of many plants are scattered about in various ways—
some by wind, others by water and many by the birds and animals.

106. ARROWLEAF BALSAMROOT
Meadow Salsify Fruit
Sunflower
Family
Balsamorhiza sagittata
This plant, with its large, showy yellow flowers, is often found on the
southern exposures of steep hillsides or in the Sagebrush flats. It
was first discovered by Lewis and Clark on their expedition across
the continent in 1804-1806.
The rind of the root contains a turpentiny balsam, but the heart of
the root is edible and was used by the Indians and early pioneers.
The plant is called Mormon Biscuit in Utah. The seeds of the plant
were used by the Indians to make “Pinole” or meal, and the stems
and leaves were eaten as greens.

69
107. WESTERN WALLFLOWER
Arrowleaf Balsamroot
Mustard Family
Erysimum capitatum
There are two kinds of Wallflowers in Zion National Park. Their
bright-yellow flowers, which grow on stalks taller than those of most
other mustards, make them among the most attractive members of
this family. They are usually found on rather dry slopes in the Upper
Sonoran and Transition Zones.
Notice how the petals are arranged as a cross which is a
characteristic of all members of the Cruciferae or Mustard Family.

108. BITTERCRESS
Western Wallflower
Mustard Family
Cardamine hirsuta
You may find this plant blooming during April and May in the
Sonoran Zones of Zion National Park. Its habitat is generally the dry
sandy hillsides rather than the deep canyons.
The wide-spreading, circular, doom-shaped clumps present an
attractive display in pure white flowers. The petals of four are
arranged like a cross.
Being a perennial, the clumps seem to expand from year to year and
often reach a spread of four to five feet across. The plants are useful
in building soil and in preventing erosion.

109. HUMMINGBIRD TRUMPET
Bittercress
Evening-
primrose Family
Zauschneria garrettii
One of the late blooming plants in Zion National Park is the
Hummingbird Trumpet, also called Fire-chalice, or sometimes the
Wild Fushia. It can often be found on the Canyon Overlook Trail or
on the West Rim Trail at elevations near 6,000 feet.
It can be identified by the narrow oval leaves pointed and toothed,
and the fushialike flowers, narrowly funnel-shaped, with the pistil
and stamens extending beyond the petals.
The brilliant scarlet of this flower in fairly dense clusters makes a
very attractive display in late August and September.

INDEX
Common Name Scientific Name
Figure
Number
A
Alfilera Erodium circutraium 48
American Harebell Campanula rotundifolia 92
Antelope Bitterbrush Purshia tridentata 34
Arnica, Heartleaf Arnica cordifolia 103
Arrowleaf Balsamroot Balsamorhiza sagittata 106
Aster, Purple Aster sp. 96
B
Balsamroot, Arrowleaf Balsamorhiza sagittata 106
Baneberry, Western Actaea arguta 20
Bearberry Honeysuckle Lonicera involucrata 90
Beavertail Cactus Opuntia basilaris 61
Bitterbrush, Antelope Purshia tridentata 34
Bittercress Cardamine hirsuta 108
Bitterroot Lewisia rediviva 14
Blackbrush Coleogyne ramosissima 35
Blazingstar, Desert Mentzelia multiflora 57
Bluebells, Mountain Mertensia ciliata 78
Bluedicks Dichelostemma
pulchellum
4
Buckhorn Cholla Cactus Opuntia acanthocarpa 62
Buckwheat, Wild Eriogonum umbellatum 8
Buffaloberry, Roundleaf Shepherdia rotundifolia 65

Bush Cinquefoil Potentilla fruticosa 36
Buttercup, Sand Ranunculus juniperinus 21
Buttercup Ranunculus sp. 16
Butterfly Milkweed Asclepias tuberosa 75
C
Cactus Fruit Opuntia engelmannii 64
Calypso Orchid Calypso bulbosa 11
Cardinalflower, Western Lobelia splendens 93
Cinquefoil, Bush Potentilla fruticosa 36
Chokecherry Prunus virginiana 40
Cliffrose, Stansbury Cowania stansburiana 33
Columbine Aquilegia sp. 17
Coneflower Rudbeckia occidentalis 25
Creosotebush Larrea tridentata 50
Curlycup Gumweed Grindelia squarrosa 95
D
Deathcamas, Mountain Zigadenus elegans 7
Desertbeauty Dalea Dalea johnsoni 46
Desert Blazingstar Mentzelia multiflora 57
Desert Globemallow Sphaeralcea ambigua 53
Desert Marigold Baileya multiradiata 101
Desert Princesplume Stanleya pinnata 26
Desert Sage Salvia carnosa 80
E
Eaton Penstemon Penstemon eatoni 83
Elder, Red-berried Sambucus racemosa 31
Elephanthead Pedicularis Pedicularis groenlandica 91
Elk Thistle Cirsium foliosum 24
Engelmann Pricklypear
Cactus
Opuntia engelmannii 64
Ephedra, Green Ephedra viridis 9

Euphorbia, Whitemargin Euphorbia albomarginata 51
Evening-primrose, White Oenothera caespitosa 66
Evening-primrose, Yellow Oenothera strigosa 67
F
Fineleaf Yucca Yucca angustissima 6
Firechalice Zauschneria garrettii 109
Fireweed Epilobium angustifolium 54
Flax, Lewis Linum lewisii 49
Fleabane Erigeron speciosus 97
Four-O’Clock Mirabilis multiflora 12
Fourwing Saltbush Atriplex canescens 10
Fremont Geranium Geranium fremontii 47
Fremont Barberry Berberis fremonti 23
Fringed Gentian Gentiana thermalis 74
Fritillary, Purplespot Fritillaria atropurpurea 5
G
Gaillardia Gaillardia parryi 100
Gentian, Fringed Gentiana thermalis 74
Gentian, Green Frasera speciosa 70
Geranium, Fremont Geranium fremontii 47
Gilia, Skyrocket Gilia aggregata 77
Globemallow, Scarlet Sphaeralcea coccinea 52
Globemallow, Desert Sphaeralcea ambigua 53
Green Ephedra Ephedra viridis 9
Green Gentian Frasera speciosa 70
Greenleaf Manzanita Arctostaphylis patula 69
Gumweed, Curlyleaf Grindelia squarrosa 95
H
Harebell, American Campanula petiolata 92
Heartleaf Arnica Arnica cordifolia 103

Hedgehog Cactus Echinocereus coccineus 58
Honey Mesquite Prosopis juliflora 41
Honeysuckle, Bearberry Lonicera involucrata 90
Hummingbird Trumpet Zauschneria garrettii 109
I
Indianpotato Orogenia linearifolia 72
Indian Paintbrush Castilleja coccinea 88
L
Larkspur Delphinium sp. 18
Lewis Flax Linum lewisii 49
Ligusticum, Porter Ligusticum porteri 68
Littleleaf
Mountainmahogany
Cercocarpus intricatus 32
Loco Astragalus sp. 43
Locust, New Mexico Robinia neomexicana 45
Lupine Lupinus sp. 42
M
Manzanita, Greenleaf Arctostaphylis patula 69
Mariposa, Segolily Calochortus nuttallii 1
Mariposa, Yellow Calochortus nuttallii var.
aureus
2
Marigold, Desert Baileya multiradiata 101
Marshmarigold Caltha leptosepala 15
Meadow Salsify Tragopogon pratensis 104-105
Mesquite, Honey Prosopis juliflora 41
Milkweed, Butterfly Asclepias tuberosa 75
Monkeyflower Mimulus cardinalis 87
Monkshood Aconitum columbianum 19
Mountain Bluebells Mertensia ciliata 78
Mountain Deathcamas Zigadenus elegans 7

Mountainmahogany,
Littleleaf
Cercocarpus intricatus 32
Mountain Sunflower Helianthella uniflora 98
Mullein Verbascum thapsus 81
N
New Mexico Locust Robinia neomexicana 45
O
Oregon Grape Berberis repens 22
Oysterplant - Meadow
Salsify
Tragopogon pratensis 104-105
P
Palmer Penstemon Penstemon palmeri 85
Parry Primrose Primula parryi 71
Penstemon, Eaton Penstemon eatoni 83
Penstemon, Palmer Penstemon palmeri 85
Penstemon, Royal Penstemon speciosus 82
Penstemon, Thickleaf Penstemon pachyphyllus 84
Phlox, Pink Phlox canescens 76
Pinedrops, Woodland Pterospora andromedea 55
Poisonvetch Astragalus sabulosus 44
Porter Ligusticum Ligusticum porteri 68
Prairiesmoke Geum triflorum var.
ciliatum
37
Prairie Spiderwort Tradescantia occidentalis 3
Pricklepoppy Argemone platyceras 28
Pricklypear Cactus Opuntia rhodantha 60
Princesplume, Desert Stanleya pinnata 26
Puccoon, Narrowleaf Lithospermum incisum 79
Purplespot Fritillary Fritillaria atropurpurea 5
Purple Torch Cactus Echinocereus engelmanii 59
R

Rabbitbrush Chrysothamnus sp. 94
Rocky Mountain Beeplant Cleome serrulata 29
Rose, Wild Rosa sp. 38
Roundleaf Buffaloberry Shepherdia rotundifolia 65
Royal Penstemon Penstemon speciosus 82
S
Sacred Datura Datura meteloides 86
Sage, Desert Salvia carnosa 80
Saltbush, Fourwing Atriplex canescens 10
Sand Buttercup Ranunculus juniperinus 21
Scarlet Globemallow Sphaeralcea coccinea 52
Segolily Mariposa Calochortus nuttalli 1
Serviceberry Amelanchier alnifolia 39
Shootingstar Dodecatheon pauciflorum 73
Skyrocket Gilia Gilia aggregata 77
Spiderflower, Yellow Cleome lutea 30
Snowberry Symphoricarpos utahensis 89
Spiderwort, Prairie Tradescantia occidentalis 3
Springbeauty Claytonia lanceolata 13
Stansbury Cliffrose Cowania stansburiana 33
Stonecrop Sedum stenopetalum 27
Sunflower, Common Helianthus annuus 99
Sunflower, Mountain Helianthella uniflora 98
T
Thickleaf Penstemon Penstemon pachyphyllus 84
Thistle, Elk Cirsium foliosum 107
V
Violet, Yellow Viola praemorsa 56
W
Western Baneberry Actaea arguta 20

74
Western Cardinalflower Lobelia splendens 93
Western Wallflower Erysimum capitatum 107
Western Yarrow Achillea lanulosa 102
Whitemargin Euphorbia Euphorbia albomarginata 51
Wild Buckwheat Eriogonum umbellatum 8
Wild Rose Rosa sp. 38
Woodland Pinedrops Pterospora andromedea 55
Y
Yarrow, Western Achillea lanulosa 102
Yellow Mariposa Calochortus nuttallii var.
aureus
2
Yellow Spiderflower Cleome lutea 30
Yellow Violet Viola praemorsa 56
Yucca, Fineleaf Yucca angustissima 6

Welcome to our website – the perfect destination for book lovers and
knowledge seekers. We believe that every book holds a new world,
offering opportunities for learning, discovery, and personal growth.
That’s why we are dedicated to bringing you a diverse collection of
books, ranging from classic literature and specialized publications to
self-development guides and children's books.
More than just a book-buying platform, we strive to be a bridge
connecting you with timeless cultural and intellectual values. With an
elegant, user-friendly interface and a smart search system, you can
quickly find the books that best suit your interests. Additionally,
our special promotions and home delivery services help you save time
and fully enjoy the joy of reading.
Join us on a journey of knowledge exploration, passion nurturing, and
personal growth every day!
ebookbell.com

Virtualizing 5g And Beyond 5g Mobile Networks Larry J Horner

More Related Content

Similar to Virtualizing 5g And Beyond 5g Mobile Networks Larry J Horner (20)

Recently uploaded (20)

Virtualizing 5g And Beyond 5g Mobile Networks Larry J Horner