Edge Computing - Shi, Weisong

. .

Résumé :

About the Authors xiii

Preface xv

About the Companion Website xvii

1 Why Do We Need Edge Computing? 1

1.1 The Background of the Emergence 1

1.2 The Evolutionary History 6

1.2.1 Technology Preparation Period 7

1.2.2 Rapid Growth Period 12

1.2.3 Intelligence Integration Period 14

1.3 What Is Edge Computing? 15

1.4 Summary and Practice 18

1.4.1 Summary 18

1.4.2 Practice Questions 18

1.4.3 Course Projects 18

2 Fundamentals of Edge Computing 23

2.1 Distributed Computing 23

2.1.1 Distributed Computing Technologies 24

2.1.2 Distributed System Platforms 25

2.2 The Basic Concept and Key Characteristics of Edge Computing 26

2.2.1 The Basic Concept 27

2.2.2 The Key Characteristics 29

2.3 Edge Computing vs. Cloud Computing 33

2.3.1 The Concept of Cloud Computing 34

2.3.2 The Big Data Era 35

2.3.3 Edge Computing vs. Cloud Computing 36

2.3.4 Advantages and Challenges of Edge Computing 39

2.4 Summary and Practice 42

2.4.1 Summary 42

2.4.2 Practice Questions 42

2.4.3 Course Projects 42

3 Architecture and Components of Edge Computing 47

3.1 Edge Infrastructure 47

3.1.1 Introduction to Edge Computing Architecture 47

3.1.2 Different Grades/Layers of Edge 49

3.1.3 Capabilities of Edge Infrastructure 51

3.1.4 New Progress of Edge Computing Architecture 53

3.1.5 Open Questions 54

3.2 Edge Computing Models 55

3.2.1 Overview and Definitions 55

3.2.2 Collaborative Edge Computing Models 57

3.2.3 Choosing the Right Model 61

3.2.4 Open Questions 64

3.3 Networking in Edge Computing 65

3.3.1 Introduction and Development Process of Edge Computing-Network Integration 65

3.3.2 Edge Computing-Network Architectures 68

3.3.3 Current Progress and Future Trend 69

3.4 Summary and Practice 71

3.4.1 Summary 71

3.4.2 Practice Questions 71

3.4.3 Course Projects 71

4 Toward Edge Intelligence 77

4.1 What Is Edge Intelligence? 77

4.1.1 Formal Definition 78

4.2 Hardware and Software Support 80

4.2.1 Hardware 81

4.2.2 Software 86

4.2.3 Container 90

4.3 Technologies Enabling Edge Intelligence 91

4.3.1 Compression Techniques 91

4.3.2 Hardware-Software Codesign for Edge Optimization 101

4.3.3 Applying Deep Learning Models on Resource-Constrained Edges 102

4.4 Edge Intelligent System Design and Optimization 104

4.4.1 Training on Edge 104

4.4.2 Model Inference on Edge 107

4.5 Summary and Practice 111

4.5.1 Summary 111

4.5.2 Practice Questions 112

4.5.3 Course Projects 112

5 Challenges and Solutions in Edge Computing 123

5.1 Programmability and Data Management 123

5.1.1 Programmability 123

5.1.2 Automatic Program Partitioning 125

5.1.3 Naming Conventions 126

5.1.4 Data Abstraction 128

5.2 Resource Allocation and Optimization 130

5.2.1 Scheduling Strategies 130

5.2.2 Data Offloading and Load Balancing 131

5.2.3 Optimization Metrics 133

5.3 Security, Privacy, and Service Management 136

5.3.1 Privacy Protection and Security 136

5.3.2 Edge Service Management 140

5.4 Deployment Strategies and Integration 142

5.4.1 Edge Nodes Deployment 142

5.4.2 Deployment of AI Models on Resource-Constrained Edge Devices 143

5.4.3 Integration with Vertical Industries 145

5.4.4 Hardware and Software Selection 146

5.5 Foundations and Business Models 147

5.5.1 Theoretical ...

Sommaire:

Lanyu Xu, PhD, is Assistant Professor in the Department of Computer Science and Engineering, Oakland University, Michigan, where she leads the Edge Intelligence System Laboratory. Her research intersects edge computing and deep learning, emphasizing the development of efficient edge intelligence systems. Her work explores optimization frameworks, intelligent systems, and AI applications to address challenges in efficiency and real-world applicability of edge systems across various domains.

Weisong Shi, PhD, is an Alumni Distinguished Professor and Chair of the Department of Computer and Information Sciences at the University of Delaware, where he leads the Connected and Autonomous Research Laboratory. He is an internationally renowned expert in edge computing, autonomous driving, and connected health. His pioneer paper, Edge Computing: Vision and Challenges, has been cited more than 8000 times in eight years. He is an IEEE Fellow....