Software-defined Infrastructure E-Book

Tran Tuan Chu

• Herausgeber: John Wiley & Sons
• Kategorie: Wissenschaft und neue Technologien
• Serie: ISTE Invoiced
• Sprache: Englisch

Seitenzahl: 207

Veröffentlichungsjahr: 2025

Beschreibung

In the era of the Internet of Things (IoT) and Digital Twins (DT), network infrastructures are rapidly evolving to meet industrial demands. Thus, we were motivated to explore the changing landscape of network deployment, management and utilization, driven by the rise of connected devices and emerging challenges.

Software-defined Infrastructure focuses on the cutting-edge shift in hardware deployment and communication management methods, which enable unified, scalable and adaptive network management to support Healthcare IoT (H-IoT) communication, where real-time data transmission is essential to system success. The book presents a novel network concept and solutions for tackling the challenges in key areas such as 5G, and beyond, in network management, multipath transport protocols and edge computing.

This book aims to simplify network management, improve remote patient monitoring communication and enhance patient outcomes. Through case studies and theoretical models, the book offers insights into the transformation of advanced networks in the H-IoT context and lays the foundation for innovative ideas in this research domain.

Leseprobe

Table of Contents

Cover

Table of Contents

Dedication Page

Title Page

Copyright Page

Preface

List of Acronyms

Introduction

I.1. H-IoT context

I.2. Scope identification

I.3. Main contributions

I.4. Content organization

Chapter 1: Healthcare Internet of Things: The State of the Art

1.1. H-IoT network landscape

1.2. Technology emergence in H-IoT

1.3. Learned lessons and neglected opportunities

1.4. Conclusion

Chapter 2: A Novel Network Infrastructure Concept

2.1. Introduction

2.2. Related work

2.3. The evolution of SDI

2.4. Unified functional model formalization

2.5. Description of the experiments

2.6. Improvement evaluation approaches

2.7. Conclusion

Chapter 3: SMART Connection Migration

3.1. Introduction

3.2. Related work

3.3. Solution design

3.4. Performance evaluation

3.5. Conclusion

Chapter 4: Generic Adaptive Deep Learning-based Multipath Scheduler Selector

4.1. Introduction

4.2. Related work

4.3. Prototype design

4.4. Simulated evaluation

4.5. Practical evaluation

4.6. Limitations

4.7. Conclusion

Conclusions and Perspectives:

C.1. Summary of contributions

C.2. Perspectives and future work

References

Index

Other titles from ISTE in Networks and Telecommunications

End User License Agreement

List of Tables

Chapter 1

Table 1.1. Healthcare IoT and generic IoT comparision

Table 1.2. Short range communication technologies parameters

Table 1.3. Long-range communication technologies parameters

Table 1.4. Edge computing in healthcare IoT

Table 1.5. Software-defined networking in healthcare IoT

Chapter 2

Table 2.1. Correspondence between a RL system and SDI’s ambulance

Table 2.2. SDI Ambulance’s Q-table structure

Table 2.3. Experimental approach context summary

Chapter 4

Table 4.1. Path’s characteristic metric

Table 4.2. Variations of networks conditions

List of Illustrations

Introduction

Figure I.1. Contribution roadmap.

Chapter 1

Figure 1.1. Three-tier H-IoT architecture: Things, Communication and Processing ...

Figure 1.2. H-IoT communication infrastructure.

Figure 1.3. eHealth dedicated slice technological landscape.

Figure 1.4. Edge computing’s taxonomy in Health-IoT

Figure 1.5. Software-defined taxonomy in Health-IoT

Figure 1.6. Explored research topic generalization in network-centric healthcare...

Figure 1.7. Proposed research directions in network-centric healthcare IoT.

Chapter 2

Figure 2.1. First contribution’s prospect in the overall framework.

Figure 2.2. General network context.

Figure 2.3. SDI-empowered modern network infrastructure.

Figure 2.4. SDI’s practical integrating prospect

Figure 2.5. Unified functional model of each SDI component

Figure 2.6. Block diagram of the model structure

Figure 2.7. SDI ecosystem unified functional model

Figure 2.8. Scenario 1: opportunistic coverage enhancement.

Figure 2.9. Scenario 2: connection recovery.

Figure 2.10. Scenario 3: self-assisted coverage deployment and data transportati...

Figure 2.11. Scenario 4: priority orchestration in high-density network.

Figure 2.12. Network metrics.

Figure 2.13. Network bandwidth in connection recovery experiment.

Figure 2.14. Network bandwidth in self-assisted coverage enhancement.

Figure 2.15. Network throughput in priority orchestration in high-density networ...

Chapter 3

Figure 3.1. Second contribution’s prospect in the overall framework.

Figure 3.2. Sequence diagram comparison.

Figure 3.3. Connection migration module’s design and interactions

Figure 3.4. Simulated experimental environment

Figure 3.5. Impact of the source server delay on connection migration time.

Figure 3.6. Impact of consecutive connection migration on connection migration t...

Chapter 4

Figure 4.1. Third contribution prospect in the overall framework.

Figure 4.2. The scheduler selector

Figure 4.3. Proposed deep neural network architecture

Figure 4.4. Proposal architecture.

Figure 4.5. Experiment topology

Figure 4.6. Best performance histogram.

Figure 4.7. Worst-performance histogram.

Figure 4.8. Class distribution in the training, validation and testing set.

Figure 4.9. KNN’s accuracy performance

Figure 4.10. Random forest’s accuracy performance

Figure 4.11. Accuracy on the training and validation set.

Figure 4.12. Loss value on the training and validation set.

Figure 4.13. Confusion matrix of the testing set

Figure 4.14. GADaM’s prototype scheduler suggestion mechanism

Figure 4.15. Selected metro line for the first mobile scenario.

Figure 4.16. Trial run architecture.

Figure 4.17. Round-robin versus GADaM scheduler performances.

Figure 4.18. MinRTT versus GADaM schedulers performances.

Figure 4.19. BLEST versus GADaM scheduler performances.

Figure 4.20. ECF versus GADaM scheduler performances.

Figure 4.21. Peekaboo versus GADaM scheduler performances.

Figure 4.22. Experimental setup.

Figure 4.23. Static scenario scheduler performances.

Figure 4.24. Subway scenario scheduler performances.

Figure 4.25. Vehicle scenario scheduler performances.

Guide

Cover Page

Table of Contents

Dedication Page

Title Page

Copyright Page

Preface

List of Acronyms

Introduction

Begin Reading

Index

Other titles from iSTE in Networks and Telecommunications

WILEY END USER LICENSE AGREEMENT

Pages

ii

iii

iv

ix

x

xi

xii

xiii

xiv

xv

xvi

xvii

xviii

xix

xx

xxi

xxii

xxiii

xxiv

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

To my beloved mother, Aïcha Salah, who died in my arms onDecember 28th, 2024, leaving behind an immeasurable void.You have been an example to me, and always will be.You live on in each of us.Abdelhamid Mellouk

To my two grandmothers, my beloved parents, my dear sister, and the one whoquietly stood by my side throughout this journey: my deepest appreciation for yourunconditional love, unwavering support, and endless encouragement.Tran-Tuan Chu

To my family, my rock. To Alma, who changed everything.Mohamed-Aymen Labiod

To my beloved family.Brice Augustin

New Generation Networks Set

coordinated byAbdelhamid Mellouk

Volume 4

Software-defined Infrastructure

A Novel Concept of Advanced Network Infrastructure

Tran-Tuan Chu

Mohamed-Aymen Labiod

Brice Augustin

Abdelhamid Mellouk

First published 2025 in Great Britain and the United States by ISTE Ltd and John Wiley & Sons, Inc.

Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act 1988, this publication may only be reproduced, stored or transmitted, in any form or by any means, with the prior permission in writing of the publishers, or in the case of reprographic reproduction in accordance with the terms and licenses issued by the CLA. Enquiries concerning reproduction outside these terms should be sent to the publishers at the undermentioned address:

ISTE Ltd27-37 St George’s RoadLondon SW19 4EUUK

www.iste.co.uk

John Wiley & Sons, Inc.111 River StreetHoboken, NJ 07030USA

www.wiley.com

© ISTE Ltd 2025The rights of Tran-Tuan Chu, Mohamed-Aymen Labiod, Brice Augustin and Abdelhamid Mellouk to be identified as the authors of this work have been asserted by them in accordance with the Copyright, Designs and Patents Act 1988.

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s), contributor(s) or editor(s) and do not necessarily reflect the views of ISTE Group.

Library of Congress Control Number: 2025935238

British Library Cataloguing-in-Publication DataA CIP record for this book is available from the British LibraryISBN 978-1-83669-053-5

Preface

After two decades of development history, the number of Internet of Things (IoT) devices is growing dramatically. Along with the dramatic growth of the general IoT paradigm, a crucial branch has received significant research attention from the community, which is the healthcare Internet of Things (H-IoT). This particular branch results from the IoT’s emergence in the eHealth domain, which allows for innovation in humanity’s healthcare prospects.

In a broad field such as eHealth, H-IoT covers healthcare data transportation, data collection and storage. Notably, delivering vital healthcare data in real-time is an essential factor behind the success of applications built around the H-IoT paradigm, especially in emergencies. Hence, this particular characteristic demands the most attention from the communications infrastructure, with consideration toward an architecture specific to this type of application. In this respect, the evolution of the Internet operating model, impacted by new approaches such as network function virtualization (NFV), network slicing and the technological advances offered by 5G/6G networks, represents new challenges and complexities to be considered. Furthermore, such an evolution also introduces promising prospects in the network redeployment context of crisis or natural disaster situations. Part of this book was initially based on the work done in the framework of Tran-Tuan Chu’s PhD thesis, under the direction of Professor Abdelhamid Mellouk. It was then elaborated on to create a book that can provide a cutting-edge framework that can advance the use of the IoT, especially prioritizing supporting healthcare providers in urgent situations.

After studying the state of the art and addressing the current limitations in the literature, we designed a coherent framework based on an incremental approach, integrating three levels from the infrastructure to the terminal node. Our framework concerns the network infrastructure required to carry all traffic types in a mobile and constrained environment. At the top level, we consider the cutting-edge network infrastructure as a unified ecosystem – software-defined infrastructure – which can also deploy and extend its infrastructure. Thanks to this approach, we could adopt a unified optimization model built around each hardware block within this ecosystem and overcome the hardware heterogeneity imposed by the network equipment diversity. We explore our framework deeper by addressing communication infrastructure’s dynamicity issues, particularly in migrating service for a mobile object, which could be a moving emergency vehicle, such as an ambulance or similar, to provide pre-hospital care for a patient. Last but not least, at the bottom level, we present a standalone solution embedded within the mobile object to optimize end-to-end latency. More precisely, we proposed a deep learning-based approach to select the optimal multipath scheduler that manages multiple network interfaces based on current network characteristics to choose the most suitable interface for sending the traffic. The proposed framework and the solutions it offers constitute essential components for our use cases. Thus, three enhancement levels facilitate the overall performance improvement of all functional blocks within the context of the H-IoT field, whose uncertainty and complexity characteristics are distinguished.

Beside introducing H-IoT literature landscape, this book also aims to inspire the research community toward a coherent development road map by sharing a common goal, developing software-defined infrastructure altogether.

It would have been impossible to give a complete bibliography and a historical account of the research that led to the present form of the subject. It is thus inevitable that some topics have been explored in less detail than others. The choices made reflect, in part, personal taste and expertise, and, in part, a preference for very promising research and recent developments in the fields covered by the book, highlighting the ongoing need for research and development in this vital area.

This book is a starting point, but also leaves many questions unanswered. We hope that it will inspire a new generation of investigators and investigations.

Finally, the authors hope that you will find this book interesting, while sparking novel and helpful ideas for your future work.

Tran-Tuan ChuMohamed-Aymen LabiodBrice AugustinAbdelhamid MelloukApril 2025

List of Acronyms

5G

Fifth Generation Technology Standard.

AdaBoost

Adaptive Boosting.

AI

Artificial Intelligence.

ANN

Artificial Neural Network.

API

Application Programming Interface.

ASs

Autonomous Systems.

BER

Bit Error RATE.

CDF

Cumulative Distribution Function.

CNN

Convolutional Neural Network.

DDoS

Distributed Denial-of-Service.

DL

Deep Learning.

DRL

Deep Reinforcement Learning.

GADaM

Generic Adaptive Deep Learning-Based Multipath Scheduler Selector.

H-IoT

Healthcare Internet of Things.

HoL

Head of Line blocking.

HTTP

Hypertext Transfer Protocol.

HTTPS

Hypertext Transfer Protocol Secure.

ICMP

Internet Control Message Protocol.

IETF

Internet Engineering Task Force.

IoT

Internet of Things.

IPsec

Internet Protocol Security.

ISP

Internet Service Provider.

KNN

K-Nearest Neighbors.

LLDP

Link-Layer Discovery Protocol.

MAC

Message Authentication Code.

MEC

Multi-access Edge Computing.

ML

Machine Learning.

MLP

Multi-Layer Perceptron.

MNO

Mobile Network Operator.

MOS

Mean Opinion Score.

MPTCP

MultiPath Transmission Control Protocol.

NFV

Network Function Virtualization.

NOs

Network Operators.

OF

OpenFlow.

P2P

Peer-to-Peer.

PDR

Packet Delivery Ratio.

QoE

Quality of Experience.

QUIC

Quick UDP Internet Connections.

RF

Random Forest.

RTT

Round Trip Time.

SCM

Smart Connection Migration.

SDI

Software-defined Infrastructure.

SDN

Software-defined Networking.

SDWBAN

Software-defined Wireless Body Area Network.

SINR

Signal-to-Interference-plus-Noise Ratio.

SNR

Signal Noise Ratio.

SS

Scheduler Selector.

SSH

Secure Shell Protocol.

SSL

Secure Sockets Layer.

TCP

Transmission Control Protocol.

TLS

Transport Layer Security.

TLS/SSL

Transport Layer Security/Secure Sockets Layer.

UDP

User Datagram Protocol.

VPN

Virtual Private Network.

WBAN

Wireless Body Area Network.

Introduction

Digital change has beenparticularly challengingin healthcare...

da Fonseca et al. (2021)

I.1. H-IoT context

The Internet of Things (IoT) is the technology that enables the link between the world’s objects and their identities in the digital space. After two decades of development, 15 billion things will be connected to the Internet by 2025, according to CISCO. This indicates the IoT’s development capability and exposes various potential issues and opportunities with such dramatic growth. The upward trend in the IoT’s number of devices is predicted to continue increasing even further because of the technology’s broad adoption in various industries: manufacturing, logistics, smart driving, smart homes, smart cities, etc.

Among numerous domains influenced by the IoT paradigm, there is a crucial domain that consistently improves the quality of human life and also receives significant research attention from the community, which is eHealth. The eHealth paradigm describes the healthcare services supported by digital processes, technology and communication. Throughout its history, the eHealth transition has significantly improved healthcare services – automating redundant tasks, unifying digital health records, increasing patient satisfaction, lowering costs, etc. – by digitizing healthcare procedures. However, various trivial tasks that need to be included require the next level of innovation, such as large-scale health monitoring, telemedicine and telehealth. The emergence of the IoT paradigm in the eHealth domain thus leads to a branch formulation in the literature: healthcare Internet of Things (H-IoT). The emergence of this domain promises the deployment of ubiquitous multilayer connected smart devices that enable potential solutions and open new opportunities for future use cases.

To investigate existing research challenges in the field, especially toward the eHealth use case, we conduct a thesis and present our work in this book.

I.2. Scope identification

In the H-IoT branch, the critical factor that fuels the success of the ecosystem is delivering vital healthcare data in real time, especially in urgent circumstances. The main reason is that the prehospital stage is vital to maintaining the patients’ conditions before treating them at the healthcare facilities. During this phase, time and communication continuity are essential to allow on-the-scene healthcare providers to deliver proper first aid to patients. Therefore, we oriented our solutions first to cope with such critical circumstances to satisfy its strict communication requirements and preserve the holism of our proposed solutions. However, their applicable prospects are feasible and unlimited in the general IoT context or even in less competitive scenarios and other domains.

Generally, the system architecture in the H-IoT domain comprises three layers: Things, Communication and Processing. Despite many existing studies, most have focused on the healthcare-related computing tasks at the Things and Processing layers. Some might investigate the Communication layer to improve the network performance. Such improvement might happen at the gateway endpoint but rarely influences the network ecosystem. The main reason for such limitation comes from a factor that could easily be ignored by researchers. Most of the research has always been conducted with a common presumption: the communication infrastructure is static throughout the communication process and could not be influenced by a third-party actor.

However, this assumption might no longer fully reflect the typical network conditions. The principal factors leading to this situation are the complex Internet operations, the software-defined network’s advancements, the rise in mobile users and the global deployment of the 5G network. To be more explicit, the complex Internet operations and the rise in mobile users cause variations in the network infrastructure while the transmission is happening. On the other hand, the global 5G deployment and SDN developments enable advanced features such as network function virtualization and network slicing. These features allow third-party actors to create a separate network from the Internet infrastructure. They could also join, operate, manage and share their assigned infrastructure with other participants. In short, the prospect of the Internet and the features it brought with it have evolved through time and invalidated the presumed legacy of a static H-IoT communication process. Acknowledging the situation, we are motivated to update the H-IoT’s Communication layer with the latest network advancements (multiple network interfaces and cutting-edge transport protocol – QUIC). The following points specify our objectives in this work:

We confirm the statements above by studying the H-IoT state of the art (SOTA).

We present to the community our studied perspective, point out the existing limitations in the literature and later navigate our research directions.

We also present our proposed solutions to deal with the existing drawbacks and fill the pointed-out research gap in the domain’s most challenging conditions.

I.3. Main contributions

As mentioned in the previous discussion, our objective is to renovate the Communication layer in the H-IoT domain. Therefore, this book studies the latest opportunities offered by cutting-edge network technology. Briefly introducing our novel solutions, the following top-down view summarizes our contributions in this book:

– Software-defined infrastructure (SDI): a novel paradigm attempts to deal with network hardware heterogeneity by unifying the network infrastructure into one coherent logical view. The paradigm can provide dedicated solutions for critical network use cases, especially emergency H-IoT, by its central management capability and universal accessibility to the network devices. We formulated a unified functional model that allows the proposed paradigm to optimize the overall network performance or prioritize critical traffic in urgent situations. We also present four experimental approaches to evaluate the proposed paradigm’s efficiency. Our obtained simulation results also confirmed the paradigm’s network orchestrating capabilities and its effectiveness in improving network performances for vital data transmission.

– SMART connection migrations (SCM): an approach that takes advantage of the unified logical network overview empowered by the SDI paradigm to maintain a seamless network experience for emergency vehicles to transmit vital healthcare data. The proposed method not only enables the ability to ensure vital transmission continuity and optimize overall latency for onboard devices to edge computing servers, but also succeeds in reducing the required duration to execute the mobile device’s connection handover in a simulation environment by 50%.

– Generic adaptive deep-learning-based multipath scheduler selector (GADaM): an intelligent approach that manages multiple network interfaces to provide the best network performance for the network terminal in highly dynamic network conditions without the SDI controller’s assistance. In the simulation and practical environments, the “scheduler selector” deep-learning model achieved a high performance in selecting the most suitable scheduler based on multiple interfaces’ network characteristics. In particular, the practical results also proved the model’s capability to outperform the SOTA in some specific scenarios without prior knowledge.

I.4. Content organization

Having discussed the contributions of the book, we now present its organization. First of all, Chapter 1 covers the H-IoT SOTA. This chapter analyzes and classifies the surveyed papers to present the lessons learned and the opportunities neglected in the literature. Afterward, considering the given context and assisting the reader’s navigation, we structure our contributions into three dedicated chapters regarding a general system’s top-down view.

After the first chapter’s overview, Chapter 2 introduces the first contribution of SDI, which influences the whole healthcare network infrastructure with its top-down overview. Initially, we present the literature’s limited vision of the software-defined paradigm. Then, we investigate the research gap in Chapter 1. Later, we construct a new paradigm to fill the missing gap and enable future research. We also present a unified functional model with specified converged conditions and a simulation framework to concretize our proposed paradigm for prospective performance evaluations. Four distinguished experimental scenarios dedicated to supporting healthcare transport are presented to evaluate the novel paradigm’s performance. The results confirm SDI’s effectiveness by significantly improving the selected network metrics (bandwidth, delay and packet loss) versus the SOTA in healthcare vital data transmission scenarios.

Taking advantage of SDI’s accessibility and knowledge about the healthcare network ecosystem, Chapter 3 presents an intelligent approach toward seamless and transparent network experiences for a healthcare endpoint: SCM. The approach aims to bridge the gap in the SOTA Internet connection concept that primarily focuses on static infrastructure but not highly dynamic and flexible network infrastructure, whose criteria are mobile healthcare transport scenarios. By employing SDI’s network overview, our proposed approach promises to harness the cutting-edge Internet connection concept to maintain a seamless network experience and the feasibility of migrating service to optimize overall latency to the edge computing server for the terminal node. We also provide initial evaluation results in this chapter to confirm the improvements of the new approach versus the SOTA approach.

Modern network devices are equipped with multiple network interfaces thanks to cutting-edge technology development. However, these advantages have yet to exploit the offered opportunities fully and are adaptable in highly dynamic and complex environments, especially in healthcare transport scenarios. Hence, we present a novel concept that deals with such dynamicity by suggesting the most suitable scheduler based on the device’s current network characteristics. Chapter 4 discusses the proposed scheduler selector paradigm and its proof-of-concept instance (GADaM) toward an independent solution for end-user devices in high dynamic environments. This chapter presents the arguments behind the paradigm and various simulation (accuracy metrics) and practical experimental (network metrics) results to validate our idea’s efficiency. The results proved the proposed paradigm’s capability to enhance the multi-network interface solutions in the mobile network environment.

Figure I.1.Contribution roadmap.

For a better illustration, we also categorize our solutions by their influenced positions in a healthcare emergency framework. Figure I.1 describes the critical healthcare communication situation where an ambulance carries its patient to the hospital while transmitting vital information to the doctor and receiving the best instructions. Consider the first contribution: since it is essential to implement an SDI solution in each infrastructure’s network element to enable its maximum capability, we classify it as an infrastructure solution. Next, the SCM feasibility only requires the implementation of the two specified endpoints. Hence, we refer to it as an end-to-end solution. Finally, because of GADaM’s