AI-Based Advanced Optimization Techniques for Edge Computing
Description
The book offers cutting-edge insights into AI-driven optimization algorithms and their crucial role in enhancing real-time applications within fog and Edge IoT networks and addresses current challenges and future opportunities in this rapidly evolving field.
This book focuses on artificial intelligence-induced adaptive optimization algorithms in fog and Edge IoT networks. Artificial intelligence, fog, and edge computing, together with IoT, are the next generation of paradigms offering services to people to improve existing services for real-time applications. Over the past few years, there has been rigorous growth in AI-based optimization algorithms and Edge and IoT paradigms. However, despite several applications and advancements, there are still some limitations and challenges to address including security, adaptive, complex, and heterogeneous IoT networks, protocols, intelligent offloading decisions, latency, energy consumption, service allocation, and network lifetime.
This volume aims to encourage industry professionals to initiate a set of architectural strategies to solve open research computation challenges. The authors achieve this by defining and exploring emerging trends in advanced optimization algorithms, AI techniques, and fog and Edge technologies for IoT applications. Solutions are also proposed to reduce the latency of real-time applications and improve other quality of service parameters using adaptive optimization algorithms in fog and Edge paradigms.
The book provides information on the full potential of IoT-based intelligent computing paradigms for the development of suitable conceptual and technological solutions using adaptive optimization techniques when faced with challenges. Additionally, it presents in-depth discussions in emerging interdisciplinary themes and applications reflecting the advancements in optimization algorithms and their usage in computing paradigms.
Audience
Researchers, industrial engineers, and graduate/post-graduate students in software engineering, computer science, electronic and electrical engineering, data analysts, and security professionals working in the fields of intelligent computing paradigms and similar areas.
More details
Other editions
Additional editions
Persons
Mohit Kumar, PhD, is an assistant professor in the Department of Information Technology at Dr. B.R. Ambedkar National Institute of Technology, Jalandhar, India. He has published more than 60 research articles in reputed international journals and conferences and served as a session chair and keynote speaker for many international conferences and webinars in India. His research interests include cloud computing, soft computing, fog and edge computing, optimization algorithms, artificial Intelligence, and Internet of Things.
Gautam Srivastava, PhD, is a professor at Brandon University, Manitoba, Canada with over eight years of academic experience. He has published more than 150 papers in various international journals and conferences and serves as an editor for several international journals. In addition to his written work, he has delivered guest lectures in Taiwan and the Czech Republic. His research interests include data mining, big data, cloud computing, Internet of Things, and cryptography.
Ashutosh Kumar Singh, PhD, is an assistant professor in the Department of Computer Science and Engineering, United College of Engineering and Research Allahabad, India. He has published over 25 papers in reputed international journals and conferences and is a reviewer for various reputed journals, conferences, and books. His research interests include network optimization, software-defined networking, machine learning, Internet of Things, and edge computing.
Kalka Dubey, PhD, is an assistant professor in the Department of Computer Science and Engineering, Rajiv Gandhi Institute of Petroleum Technology, Amethi, India. He has published more than 20 articles in international journals and conferences. His research interests include task scheduling, virtual machine placement and allocation in cloud-based systems, quantification and monitoring of security metrics, soft computing, and enforcing security in cloud environments.
Content
Preface xv
Acknowledgement xvii
1 Navigating Next-Generation Network Architecture: Unleashing the Power of SDN, NFV, NS, and AI Convergence 1
Monika Dubey, Snehlata, Ashutosh Kumar Singh, Richa Mishra and Mohit Kumar
1.1 Introduction 2
1.2 Revolutionizing Infrastructure with SDN, NFV, and NS 4
1.2.1 SDN: Definition and Architecture 6
1.2.2 NFV: Definition and Architecture 9
1.2.3 NS: Conceptual Abstractions 11
1.3 Realizing NS Potential with SDN and NFV 13
1.4 Artificial Intelligence: Pivotal Role in Networking Transformation 15
1.4.1 Supervised Learning 16
1.4.2 Unsupervised Learning 18
1.4.3 Reinforcement Learning 18
1.4.4 Deep Learning 21
1.5 Navigating Challenges and Solutions 23
1.5.1 Performance Issues in Network Structure 23
1.5.2 Management and Orchestration Issues 24
1.5.3 Security and Privacy 24
1.5.4 New Business Models 25
1.6 Conclusion 26
Disclosure Statement 26
References 26
2 OctoEdge: An Octopus-Inspired Adaptive Edge Computing Architecture 35
Sashi Tarun
2.1 Introduction 36
2.1.1 Edge Computing as Resource Manager 36
2.1.2 Edge Computing Hurdles 37
2.1.3 Edge Computing and the Need for Adaptability 38
2.2 Problem Statement 39
2.3 Motivations 40
2.4 Related Work 41
2.5 OctoEdge Proposed Architecture 45
2.5.1 OctoEdge Working Principles 48
2.5.2 Benefits of OctoEdge 49
2.6 OctoEdge Architecture Functional Components 53
2.7 Results and Discussion 59
2.8 OctoEdge Architecture: Scope and Scientific Merits 60
2.9 Use Cases and Applications 64
2.10 Challenges and Future Directions 68
2.11 Conclusion 68
References 69
3 Development of Optimized Machine Learning Oriented Models 71
Ratnesh Kumar Dubey, Dilip Kumar Choubey and Shubha Mishra
3.1 Introduction 72
3.1.1 NSL-KDD Dataset 75
3.2 Literature Review 76
3.3 Problem Definition 78
3.4 Proposed Work 80
3.4.1 Machine Learning 82
3.5 Experimental Analysis 86
3.6 Conclusion 90
3.7 Future Scope 91
References 91
4 Leveraging Multimodal Data and Deep Learning for Enhanced Stock Market Prediction 93
Pinky Gangwani and Vikas Panthi
4.1 Introduction 94
4.1.1 Motivation and Contribution 96
4.1.2 Rationale for Selecting the Methods 98
4.2 Literature Review 100
4.3 Proposed Design of an Efficient Model that Leverages Multimodal Data and Deep Learning for Enhanced Stock Market Prediction 107
4.3.1 Discussion on Selection Criteria 114
4.4 Statistical Analysis and Comparison 116
4.5 Acknowledging Limitations and Potential Challenges 122
4.6 Mitigation Strategies and Future Directions 123
4.7 Conclusion 124
4.8 Future Scope 125
References 125
5 Context Dependent Sentiments Analysis Using Machine Learning 129
Mahima Shanker Pandey, Bihari Nandan Pandey, Abhishek Singh, Ashish Kumar Mishra and Brijesh Pandey
5.1 Introduction 130
5.1.1 Motivation 131
5.2 Literature Review 131
5.2.1 Text Sentiment 132
5.2.2 Audio Sentiment 132
5.2.3 Video Sentiment 133
5.3 Methodology 135
5.3.1 System Architecture 135
5.4 Proposed Model 137
5.4.1 Proposed Algorithm 137
5.4.2 Data Set Sources 138
5.4.3 Text Sentiment 140
5.4.4 Audio Sentiment 141
5.4.5 Video Sentiment 142
5.5 Implementations and Results 142
5.5.1 Results 142
5.5.2 Text Sentiment 143
5.5.3 Audio Sentiment 144
5.5.4 Video Sentiment 146
5.5.5 Applications 149
5.6 Conclusion 149
References 150
6 Thyroid Cancer Prediction Using Optimizations 153
Swati Sharma, Vijay Kumar Sharma, Punit Mittal, Pradeep Pant and Nitin Rakesh
6.1 Introduction 154
6.2 Background and Related Work 155
6.3 Proposed Methodology 160
6.4 Architecture 165
6.5 Materials and Methods 169
6.6 Results and Discussion 171
6.7 Conclusion 175
References 177
7 An LSTM-Oriented Approach for Next Word Prediction Using Deep Learning 181
Nidhi Shukla, Ashutosh Kumar Singh, Vijay Kumar Dwivedi, Pallavi Shukla, Jeetesh Srivastava and Vivek Srivastava
7.1 Introduction 182
7.2 Related Work 184
7.3 Design and Implementation 186
7.3.1 Background 186
7.4 Proposed Model Architecture 190
7.4.1 Experimental Setup 192
7.4.2 Dataset Specification 192
7.5 Results and Discussions 193
7.6 Conclusion 198
References 199
8 Churn Prediction in Social Networks Using Modified BiLSTM-CNN Model 203
Himanshu Rai and Jyoti Kesarwani
8.1 Introduction 204
8.2 Customer Behavior in Social Networks 209
8.3 Proposed Methodology 218
8.3.1 Churn Dataset Acquisition 218
8.3.2 Data Preprocessing 220
8.3.3 Proposed Model 220
8.4 Result 221
8.5 Conclusion 225
References 226
9 Fog Computing Security Concerns in Healthcare Using IoT and Blockchain 231
Ruchi Mittal, Shikha Gupta and Shefali Arora
9.1 Introduction 232
9.1.1 Types of Security Concerns in Healthcare 236
9.2 Related Work 239
9.3 Open Questions and Research Challenges 241
9.4 Problem Definition 242
9.5 Objectives 242
9.6 Research Methodology 243
9.6.1 The Three-Tier Blockchain Design 243
9.6.2 System Architecture 243
9.6.3 Workflow in Different Scenarios 245
9.7 Conclusion and Future Work 249
References 249
10 Smart Agriculture Revolution: Cloud and IoT-Based Solutions for Sustainable Crop Management and Precision Farming 253
Shrawan Kumar Sharma
10.1 Introduction 255
10.1.1 IoT in Agriculture 257
10.1.2 Cloud Computing in Agriculture 259
10.1.3 Precision Farming 263
10.1.4 Sustainable Agricultural and Remote Sensing 265
10.2 Data Analytics and Decision Support 267
10.2.1 Remote Monitoring 269
10.3 Challenges and Solutions Smart Agriculture 270
10.3.1 (AI) Approach in Agriculture and Needs 270
10.3.2 Needs of AI Farm 273
10.3.3 Role of AI in Agriculture 274
10.4 AI for Soybean (Glycine max) Crop 275
10.4.1 Soybean Disease Image Acquisition and Pretreatment 276
10.5 Result Discussion 281
10.5.1 Emerging Trends and Technologies in Smart Agriculture 281
10.6 Conclusion 283
References 285
11 Greedy Particle Swarm Optimization Approach Using Leaky ReLU Function for Minimum Spanning Tree Problem 289
Ashish Kumar Singh and Anoj Kumar
11.1 Introduction 290
11.1.1 Goal 291
11.1.2 Research Contribution are Below Listed 292
11.2 Background 292
11.2.1 Minimum Spanning Tree 294
11.2.2 Particle Swarm Optimization 296
11.2.3 Firefly Algorithm 297
11.2.4 Leaky ReLU Activation Function 298
11.3 Population-Based Proposed Optimization Approach 298
11.3.1 Motivation 299
11.3.2 Greedy Particle Swarm Optimization Using Leaky ReLU (LR-GPSO) 300
11.4 Experimental Setup and Result Analysis of Proposed Work (LR-GPSO) 307
11.4.1 Complexity 307
11.4.2 Simulation Experiments 308
11.4.3 Convergence Curve 311
11.5 Conclusion and Future Work 313
References 314
12 SDN Deployed Secure Application Design Framework for IoT Using Game Theory 317
Madhukrishna Priyadarsini and Padmalochan Bera
12.1 Introduction 318
12.1.1 IoT Overview 318
12.1.2 SDN Overview 319
12.1.3 Game Theory Overview 321
12.2 Background Study 322
12.2.1 IoT Security Using SDN 322
12.2.2 IoT Security Using Game Theory 323
12.3 SDN-Deployed Design Framework for IoT Using Game-Theoretic Solutions 324
12.3.1 Trust Verification 324
12.4 Case Study: SDN Deployed Design Framework in Robot Manufacturing Industry 334
12.4.1 Working Procedure of a Robot Manufacturing Industry 334
12.4.2 Integration of SDN-Deployed Design Framework in Robot Manufacturing Industry 335
12.4.3 Experimental Results 336
12.5 Discussion 338
12.6 Conclusion 339
References 339
13 Framework for PLM in Industry 4.0 Based on Industrial Blockchain 341
Ali Zaheer Agha, Rajesh Kumar Shukla, Ratnesh Mishra and Ravi Shankar Shukla
13.1 Introduction 342
13.1.1 What is Blockchain? 343
13.1.2 Blockchain Technology's Integration with Industry 4.0 343
13.1.3 Blockchain Applications in Industry 4.0 343
13.1.4 A Consensus Algorithm 344
13.1.5 Product Lifecycle Management 345
13.1.6 Benefits of Smart Contracts in Addressing PLM Challenges 347
13.2 Related Work 348
13.2.1 Product Lifecycle Management 349
13.2.2 Industrial Blockchain 351
13.2.3 The On-Chain vs. Off-Chain Principle 353
13.3 The Recommended Architecture's Methodology 354
13.3.1 The Suggested Platform's Architecture 354
13.3.2 The Suggested Platform's Technological Solution 358
13.4 Key Services That are Suggested 360
13.4.1 A Co-Creation Service Enabled by Blockchain 360
13.4.2 Blockchain-Enabled QAT2 Service 363
13.4.3 Proactive Upkeep Service Facilitated by Blockchain 364
13.4.4 Smart Recycling Program Driven by Blockchain 365
13.5 Modelling and Assessment 366
13.5.1 Overview of the Investigation 366
13.5.2 Experimental Evaluation and Comparison 368
13.5.3 Discussion 372
13.6 Conclusion and Future Work 373
A Statement of Competing Interests 374
References 375
14 Machine Learning Enabled Smart Agriculture Classification Technique for Edge Devices Using Remote Sensing Platform 381
Priyanka Gupta, Suraj Kumar Singh, Neetish Kumar and Bhavna Thakur
List of Abbreviations 382
14.1 Introduction 382
14.2 Related Works 384
14.3 Methods and Dataset 386
14.3.1 Research Area and Dataset 386
14.3.2 Pre-Processing and Image Dataset 387
14.3.3 Classifiers 390
14.4 Proposed Algorithm 391
14.5 Results and Discussions 392
14.5.1 Classified Crop Map 394
14.6 Conclusion 395
References 396
15 A Lightweight Intelligent Detection Approach for Interest Flooding Attack 401
Naveen Kumar, Brijendra Pratap Singh and Rohit
15.1 Introduction 402
15.2 NDN Background 405
15.2.1 NDN Architecture 405
15.2.2 NDN Security 408
15.3 Related Work 409
15.4 IFA Feature Selection and Detection 411
15.4.1 IFA Modelling 412
15.4.2 Data Collection 413
15.4.3 Balancing the Dataset 414
15.4.4 Feature Selection 415
15.4.5 Dimensionality Reduction 421
15.4.6 Classification 424
15.5 Conclusion 428
References 429
16 An Internet of Vehicles Model Architecture with Seven Layers 433
Sujata Negi Thakur, Manisha Koranga, Sandeep Abhishek, Richa Pandey and Mayurika Joshi
16.1 Introduction 434
16.2 Literature Review 435
16.3 Proposed Architecture of Internet of Vehicles 439
16.4 Applications, Characteristics, and Challenges of the Internet of Vehicles (IoV) 451
Conclusion 455
References 455
Index 457
1
Navigating Next-Generation Network Architecture: Unleashing the Power of SDN, NFV, NS, and AI Convergence
Monika Dubey1*, Snehlata2, Ashutosh Kumar Singh2, Richa Mishra1 and Mohit Kumar3
1Department of Electronics & Communication, University of Allahabad, Prayagraj, U.P., India
2Department of Computer Science and Engineering, United College of Engineering & Research, Prayagraj, U.P., India
3Department of Information Technology, National Institute of Technology Jalandhar, Punjab, India
Abstract
The framework for existing legacy network architecture is massive and complex. It mainly relies on inflexible and expensive equipment, typically constructed from a massive number of switches, routers, firewalls, and hubs. Moreover, this vendor-specific network configuration and complex control protocols are not flexible enough to offer customized quality of services (QoS). Provisioning of next-gen (Next Generation, 5G, and beyond) technologies, software-defined networking (SDN), network function virtualization (NFV), and network slicing (NS) work as catalysts to offer simplified, customized, and clever networking. To provide centralized positioning, SDN decouples the control plane (CP) and data plane (DP) from the traditional router. In the SDN architecture, decision making and network control are now done at a centralized place known as the controller. However, DP is still intact with the routing device. This arrangement privileges the network administrators to control, manage, and alter network behavior dynamically. To contrast the vender-specific networking, NFV allows network functions (NFs) to run on generic hardware. In this direction, NS pioneers QoS-specific use cases as a new business model. NS involves the slicing of a single physical network in the form of multiple slices. It not only supports the customization of QoS services for diverse use cases, but it also improves isolation, independence, multitenancy, dynamic resource allocation, and end-to-end service provisioning. In this chapter, we first delved into NexGen's promising technologies and explored their intertwined role and impact on the modern networking framework. We accessed various SDN and NFV architectures and discussed network-slicing framework. Secondly, we have shed light on the importance of AI-driven automated network management over traditional network approaches. In this sequence, we conducted a comparative analysis of AI-driven machine learning (ML) and deep learning (DL) approaches in the context of NextGen technologies. In this chapter, we intend to systematically and intricately navigate the multifaceted landscape of NexGen technologies. This chapter will offer researchers, industry stakeholders, and practitioners a timely and deeper understanding of transformative technology and its impact on modern network paradigms.
Keywords: Next-generation technology, SDN, NFV, QoS, NS
1.1 Introduction
The evolution of network technologies has marked pivotal advancements in the telecom sector. It spans from the radiant stage of ARPANET to modern networking. The existing legacy network architecture is based upon un-flexible and costly network equipment comprising switches, hubs, routers, and firewalls [1]. These proprietary hardware-based traditional networks grapple with the demands of modern networking. The surge of extensive data traffic, dynamic network conditions, and the need for real-time decision-makers pose challenges that traditional networks are not capable of addressing efficiently [2]. Traditional methods, such as Static Routing, Ethernet, Transmission Control Protocol (TCP), and Internet Protocol (IP), are built on manual configuration and static protocols. With the surge of diverse applications, customized QoS, high volume, and unpredicted traffic necessitate a paradigm shift. To address these limitations of the traditional approach, Next-Gen (Next Generation, 5G, and beyond) technologies, Software Defined Networking (SDN), Network Function Virtualization (NFV), and NS act as catalysts for redefining the network paradigm. SDN [3] disrupts traditional decentralized architecture by decoupling the Control Plane (CP) and Data Plane (DP) from conventional routers. This centralized control and decision-making entity is known as the controller. This architectural shift empowers the network controller to dynamically manage, control, and modify the network behavior. Concurrently, NFV [4] revolutionizes network functionality by enabling them to run on generic hardware instead of proprietary hardware, offering cost-effectiveness, flexibility, and simplified maintenance. With the advancement of the network landscape, customize QoS-specific servers are the new business model. In this direction, NS [5] has become a revolutionary approach, involving the partitioning of a single physical network into multiple slices. It not only offers customized QoS requirements to modern applications but also enhances isolation, dynamic resource allocation, multi-tenancy, and security [6].
This book chapter also explored the NextGen promising technologies and their intertwined role and impact on modern networking. Traditional networking approaches are static and require human intervention during changes in the network. The increase in network size and the unpredictable nature of network traffic make them more time-consuming and complex. Therefore, AI emerges as a key driver for NextGen networking. It introduced the level of intelligence with its learning and capability of predictive analysis. This chapter also sheds light on how AI-driven approaches complement and enhance the functionalities of SDN, NFV, and NS.
The contributions and highlight of this book chapter are as follows:
- Initially, we present a concise overview of the evolutionary history of network technologies and the key phases that shaped the modern networking landscape.
- To explore the transformative NexGen technologies (SDN, NFV, and NS), we highlight the influence and intertwining role of NexGen technologies.
- This paper systematically highlights the importance of AI over traditional methods. In this sequence, we conducted a comparative analysis of AI-driven Machine Learning (ML) and Deep Learning (DL) approaches in the context of NextGen technologies.
- Finally, we identify challenges associated with NexGen Technologies and with the integration of these modern technologies.
In a nutshell, this chapter will offer researchers and industry stakeholders a timely and deep understanding of transformative NexGen technologies and the impact of their combination on modern technology. It also includes the contribution and comparative analysis of AI-driven algorithms in the context of NexGen technologies.
1.2 Revolutionizing Infrastructure with SDN, NFV, and NS
Due to increasing day-to-day network traffic, networking technologies have undergone a continuous evolution, and based on this, they can be categorized into several phases, such as traditional networking, Wireless Sensor Networking (WSN), client-server networking, and more. Before discussing NexGen technologies and its specifications, it is crucial to examine the evolutionary changes of networking technologies and the key developments that have been influenced by traditional networking. Concise overview is given as follows:
- ARPANET and Early Networking:
- ARPANET: The Advanced Research Projects Agency Network (ARPANET) [7], established in the 1960s, conducted early experiments for linking computer systems over short distances. It laid the foundation for modern networking. However, these networks remained restricted to research institutions.
- Packet Switching: The development of packet switching [8], a key innovation during this era, allowed data to be broken into packets, transmitted independently, and reassembled at the intended destination.
The pioneering work and packet switching laid the fundamental groundwork for the internet.
- Emergence of the Internet:
- Standardization (TCP/IP): During the 1980s, the TCP [9] and IP underwent standardization, forming the backbone of the modern Internet.
- Commercialization: The Internet underwent a pivotal shift from being primarily dedicated to research and academia to a commercial platform, leading to the rise of the World Wide Web (WWW). It establishes the fundamental framework for the contemporary Internet.
- Emergence of Client-Server Architecture and LANs:
- Client-Server Model: In 1980s, the paradigm of computing is shifting from centralized mainframes to distributed systems with the client-server model [10].
- The rise of Local Area Networks (LANs): The internet and other LAN technologies emerged, allowing computers to share resources within confined spaces.
- Wireless Networking and Mobility:
- Wi-Fi Standardization: In the 2000s, the standardization of wireless technologies, particularly Wi-Fi adoption [11], empowered enhanced mobility and flexibility in network access.
- Expansion of Mobile Networks: The surge in mobile device usage during this era empowered enhanced mobility and flexibility in network access [12].
- Cloud Computing and Virtualization:
- Evolution of Cloud Services: The 2010s witnessed a transformative shift with the advent of cloud computing [13], fundamentally changing the way data and applications...
