List of Contributors xxiii Acronyms xxix Part I Fog Computing Systems and Architectures 1 1 Mobile Fog Computing 3Chii Chang, Amnir Hadachi, Jakob Mass, and Satish Narayana Srirama 1.1 Introduction 3 1.2 Mobile Fog Computing and Related Models 5 1.3 The Needs of Mobile Fog Computing 6 1.3.1 Infrastructural Mobile Fog Computing 7 1.3.2 Land Vehicular Fog 9 1.3.3 Marine Fog 11 1.3.4 Unmanned Aerial Vehicular Fog 12 1.3.5 User Equipment-Based Fog 13 1.4 Communication Technologies 15 1.4.1 IEEE 802.11 15 1.4.2 4G, 5G Standards 16 1.4.3 WPAN, Short-Range Technologies 17 1.4.4 LPWAN, Other Medium- and Long-Range Technologies 18 1.5 Nonfunctional Requirements 18 1.5.1 Heterogeneity 20 1.5.2 Context-Awareness 23 1.5.3 Tenant 25 1.5.4 Provider 27 1.5.5 Security 29 1.6 Open Challenges 31 1.6.1 Challenges in Land Vehicular Fog Computing 31 1.6.2 Challenges in Marine Fog Computing 32 1.6.3 Challenges in Unmanned Aerial Vehicular Fog Computing 32 1.6.4 Challenges in User Equipment-based Fog Computing 33 1.6.5 General Challenges 33 1.7 Conclusion 35 Acknowledgment 36 References 36 2 Edge and Fog: A Survey, Use Cases, and Future Challenges 43Cosmin Avasalcai, Ilir Murturi, and Schahram Dustdar 2.1 Introduction 43 2.2 Edge Computing 44 2.2.1 Edge Computing Architecture 46 2.3 Fog Computing 47 2.3.1 Fog Computing Architecture 49 2.4 Fog and Edge Illustrative Use Cases 50 2.4.1 Edge Computing Use Cases 50 2.4.2 Fog Computing Use Cases 54 2.5 Future Challenges 57 2.5.1 Resource Management 57 2.5.2 Security and Privacy 58 2.5.3 Network Management 61 2.6 Conclusion 61 Acknowledgment 62 References 62 3 Deep Learning in the Era of Edge Computing: Challenges and Opportunities 67Mi Zhang, Faen Zhang, Nicholas D. Lane, Yuanchao Shu, Xiao Zeng, Biyi Fang, Shen Yan, and Hui Xu 3.1 Introduction 67 3.2 Challenges and Opportunities 68 3.2.1 Memory and Computational Expensiveness of DNN Models 68 3.2.2 Data Discrepancy in Real-world Settings 70 3.2.3 Constrained Battery Life of Edge Devices 71 3.2.4 Heterogeneity in Sensor Data 72 3.2.5 Heterogeneity in Computing Units 73 3.2.6 Multitenancy of Deep Learning Tasks 73 3.2.7 Offloading to Nearby Edges 75 3.2.8 On-device Training 76 3.3 Concluding Remarks 76 References 77 4 Caching, Security, and Mobility in Content-centric Networking 79Osman Khalid, Imran Ali Khan, Rao Naveed Bin Rais, and Assad Abbas 4.1 Introduction 79 4.2 Caching and Fog Computing 81 4.3 Mobility Management in CCN 82 4.3.1 Classification of CCN Contents and their Mobility 83 4.3.2 User Mobility 83 4.3.3 Server-side Mobility 84 4.3.4 Direct Exchange for Location Update 84 4.3.5 Query to the Rendezvous for Location Update 84 4.3.6 Mobility with Indirection Point 84 4.3.7 Interest Forwarding 85 4.3.8 Proxy-based Mobility Management 85 4.3.9 Tunnel-based Redirection (TBR) 86 4.4 Security in Content-centric Networks 88 4.4.1 Risks Due to Caching 90 4.4.2 DOS Attack Risk 90 4.4.3 Security Model 91 4.5 Caching 91 4.5.1 Cache Allocation Approaches 91 4.5.2 Data Allocation Approaches 93 4.6 Conclusions 101 References 101 5 Security and Privacy Issues in Fog Computing 105Ahmad Ali, Mansoor Ahmed, Muhammad Imran, and Hasan Ali Khattak 5.1 Introduction 105 5.2 Trust in IoT 107 5.3 Authentication 109 5.3.1 Related Work 109 5.4 Authorization 113 5.4.1 Related Work 114 5.5 Privacy 117 5.5.1 Requirements of Privacy in IoT 118 5.6 Web Semantics and Trust Management for Fog Computing 120 5.6.1 Trust Through Web Semantics 120 5.7 Discussion 123 5.7.1 Authentication 124 5.7.2 Authorization 125 5.8 Conclusion 130 References 130 6 How Fog Computing Can Suppor Latency/Reliability-sensitive IoT Applications: An Overview and a Taxonomy of State-of-the-art Solutions 139Paolo Bellavista, Javier Berrocal, Antonio Corradi, Sajal K. Das, Luca Foschini, Isam Mashhour Al Jawarneh, and Alessandro Zanni 6.1 Introduction 139 6.2 Fog Computing for IoT: Definition and Requirements 142 6.2.1 Definitions 142 6.2.2 Motivations 144 6.2.3 Fog Computing Requirements When Applied to Challenging IoTs Application Domains 148 6.2.4 IoT Case Studies 152 6.3 Fog Computing: Architectural Model 154 6.3.1 Communication 154 6.3.2 Security and Privacy 156 6.3.3 Internet of Things 156 6.3.4 Data Quality 156 6.3.5 Cloudification 157 6.3.6 Analytics and Decision-Making 157 6.4 Fog Computing for IoT: A Taxonomy 158 6.4.1 Communication 159 6.4.2 Security and Privacy Layer 165 6.4.3 Internet of Things 170 6.4.4 Data Quality 173 6.4.5 Cloudification 179 6.4.6 Analytics and Decision-Making Layer 183 6.5 Comparisons of Surveyed Solutions 189 6.5.1 Communication 189 6.5.2 Security and Privacy 191 6.5.3 Internet of Things 193 6.5.4 Data Quality 194 6.5.5 Cloudification 195 6.5.6 Analytics and Decision-Making Layer 197 6.6 Challenges and Recommended Research Directions 198 6.7 Concluding Remarks 201 References 202 7 Harnessing the Computing Continuum for Programming Our World 215Pete Beckman, Jack Dongarra, Nicola Ferrier, Geoffrey Fox, Terry Moore, Dan Reed, and Micah Beck 7.1 Introduction and Overview 215 7.2 Research Philosophy 217 7.3 A Goal-oriented Approach to Programming the Computing Continuum 219 7.3.1 A Motivating Continuum Example 219 7.3.2 Goal-oriented Annotations for Intensional Specification 221 7.3.3 A Mapping and Run-time System for the Computing Continuum 222 7.3.4 Building Blocks and Enabling Technologies 224 7.4 Summary 228 References 228 8 Fog Computing for Energy Harvesting-enabled Internet of Things 231S. A. Tegos, P. D. Diamantoulakis, D. S. Michalopoulos, and G. K. Karagiannidis 8.1 Introduction 231 8.2 System Model 232 8.2.1 Computation Model 233 8.2.2 Energy Harvesting Model 235 8.3 Tradeoffs in EH Fog Systems 238 8.3.1 Energy Consumption vs. Latency 238 8.3.2 Execution Delay vs. Task Dropping Cost 239 8.4 Future Research Challenges 240 Acknowledgment 241 References 241 9 Optimizing Energy Efficiency of Wearable Sensors Using Fog-assisted Control 245Delaram Amiri, Arman Anzanpour, Iman Azimi, Amir M. Rahmani, Pasi Liljeberg, Nikil Dutt, and Marco Levorato 9.1 Introduction 245 9.2 Background 247 9.3 Related Topics 249 9.4 Design Challenges 250 9.5 IoT System Architecture 251 9.5.1 Fog Computing and its Benefits 252 9.6 Fog-assisted Runtime Energy Management in Wearable Sensors 253 9.6.1 Computational Self-Awareness 255 9.6.2 Energy Optimization Algorithms 255 9.6.3 Myopic Strategy 258 9.6.4 MDP Strategy 259 9.7 Conclusions 263 Acknowledgment 264 References 264 10 Latency Minimization Through Optimal Data Placement in Fog Networks 269Ning Wang and Jie Wu 10.1 Introduction 269 10.2 RelatedWork 272 10.2.1 Long-Term and Short-Term Placement 272 10.2.2 Data Replication 272 10.3 Problem Statement 273 10.3.1 Network Model 273 10.3.2 Multiple Data Placement with Budget Problem 274 10.3.3 Challenges 274 10.4 Delay Minimization Without Replication 275 10.4.1 Problem Formulation 275 10.4.2 Min-Cost Flow Formulation 276 10.4.3 Complexity Reduction 277 10.5 Delay Minimization with Replication 279 10.5.1 Hardness Proof 279 10.5.2 Single Request in Line Topology 279 10.5.3 Greedy Solution in Multiple Requests 280 10.5.4 Rounding Approach in Multiple Requests 282 10.6 Performance Evaluation 285 10.6.1 Trace Information 285 10.6.2 Experimental Setting 285 10.6.3 Algorithm Comparison 286 10.6.4 Experimental Results 287 10.7 Conclusion 289 Acknowledgement 289 References 290 11 Modeling and Simulation of Distributed Fog Environment Using FogNetSim++ 293Tariq Qayyum, Asad Waqar Malik, Muazzam A. Khan, and Samee U. Khan 11.1 Introduction 293 11.2 Modeling and Simulation 294 11.3 FogNetSim++: Architecture 296 11.4 FogNetSim++: Installation and Environment Setup 298 11.4.1 OMNeT++ Installation 298 11.4.2 FogNetSim++ Installation 300 11.4.3 Sample Fog Simulation 300 11.5 Conclusion 305 References 305 Part II Fog Computing Techniques and Applications 309 12 Distributed Machine Learning for IoT Applications in the Fog 311Aluizio F. Rocha Neto, Flavia C. Delicato, Thais V. Batista, and Paulo F. Pires 12.1 Introduction 311 12.2 Challenges in Data Processing for IoT 314 12.2.1 Big Data in IoT 315 12.2.2 Big Data Stream 318 12.2.3 Data Stream Processing 319 12.3 Computational Intelligence and Fog Computing 322 12.3.1 Machine Learning 322 12.3.2 Deep Learning 326 12.4 Challenges for Running Machine Learning on Fog Devices 328 12.4.1 Solutions Available on the Market to Deploy ML on Fog Devices 331 12.5 Approaches to Distribute Intelligence on Fog Devices 334 12.6 Final Remarks 340 Acknowledgments 341 References 341 13 Fog Computing-Based Communication Systems for Modern Smart Grids 347Miodrag Forcan and Mirjana Maksimović 13.1 Introduction 347 13.2 An Overview of Communication Technologies in Smart Grid 349 13.3 Distribution Management System (DMS) Based on Fog/Cloud Computing 356 13.4 Real-time Simulation of the Proposed Feeder-based Communication Scheme Using MATLAB and ThingSpeak 359 13.5 Conclusion 366 References 367 14 An Estimation of Distribution Algorithm to Optimize the Utility of Task Scheduling Under Fog Computing Systems 371Chu-ge Wu and Ling Wang 14.1 Introduction 371 14.2 Estimation of Distribution Algorithm 372 14.3 Related Work 373 14.4 Problem Statement 374 14.5 Details of Proposed Algorithm 376 14.5.1 Encoding and Decoding Method 376 14.5.2 uEDA Scheme 377 14.5.3 Local Search Method 378 14.6 Simulation 378 14.6.1 Comparison Algorithm 378 14.6.2 Simulation Environment and Experiment Settings 379 14.6.3 Compared with the Heuristic Method 381 14.7 Conclusion 383 References 383 15 Reliable and Power-Efficient Machine Learning in Wearable Sensors 385Parastoo Alinia and Hassan Ghasemzadeh 15.1 Introduction 385 15.2 Preliminaries and Related Work 386 15.2.1 Gold Standard MET Computation 386 15.2.2 Sensor-based MET Estimation 387 15.2.3 Unreliability Mitigation 388 15.2.4 Transfer Learning 388 15.3 System Architecture and Methods 389 15.3.1 Reliable MET Calculation 390 15.3.2 The Reconfigurable MET Estimation System 392 15.4 Data Collection and Experimental Procedures 394 15.4.1 Exergaming Experiment 394 15.4.2 Treadmill Experiment 395 15.5 Results 396 15.5.1 Reliable MET Calculation 396 15.5.2 Reconfigurable Design 402 15.6 Discussion and Future Work 404 15.7 Summary 405 References 406 16 Insights into Software-Defined Networking and Applications in Fog Computing 411Osman Khalid, Imran Ali Khan, and Assad Abbas 16.1 Introduction 411 16.2 OpenFlow Protocol 414 16.2.1 OpenFlow Switch 414 16.3 SDN-Based Research Works 416 16.4 SDN in Fog Computing 419 16.5 SDN in Wireless Mesh Networks 421 16.5.1 Challenges in Wireless Mesh Networks 421 16.5.2 SDN Technique in WMNs 421 16.5.3 Benefits of SDN in WMNs 423 16.5.4 Fault Tolerance in SDN-based WMNs 424 16.6 SDN in Wireless Sensor Networks 424 16.6.1 Challenges in Wireless Sensor Networks 424 16.6.2 SDN in Wireless Sensor Networks 425 16.6.3 Sensor Open Flow 426 16.6.4 Home Networks Using SDWN 426 16.6.5 Securing Software Defined Wireless Networks (SDWN) 426 16.7 Conclusion 427 References 427 17 Time-Critical Fog Computing for Vehicular Networks 431Ahmed Chebaane, Abdelmajid Khelil, and Neeraj Suri 17.1 Introduction 431 17.2 Applications and Timeliness Guarantees and Perturbations 434 17.2.1 Application Scenarios 434 17.2.2 Application Model 436 17.2.3 Timeliness Guarantees 436 17.2.4 Benchmarking Vehicular Applications Concerning Timeliness Guarantees 437 17.2.5 Building Blocks to Reach Timeliness Guarantees 440 17.2.6 Timeliness Perturbations 441 17.3 Coping with Perturbation to Meet Timeliness Guarantees 443 17.3.1 Coping with Constraints 443 17.3.2 Coping with Failures 448 17.3.3 Coping with Threats 448 17.4 Research Gaps and Future Research Directions 449 17.4.1 Mobile Fog Computing 449 17.4.2 Fog Service Level Agreement (SLA) 450 17.5 Conclusion 451 References 451 18 A Reliable and Efficient Fog-Based Architecture for Autonomous Vehicular Networks 459Shuja Mughal, Kamran Sattar Awaisi, Assad Abbas, Inayat ur Rehman, Muhammad Usman Shahid Khan, and Mazhar Ali 18.1 Introduction 459 18.2 Proposed Methodology 461 18.3 Hypothesis Formulation 463 18.4 Simulation Design 464 18.4.1 Results and Discussions 464 18.4.2 Hypothesis Testing 467 18.5 Conclusions 469 References 470 19 Fog Computing to Enable Geospatial Video Analytics for Disaster-incident Situational Awareness 473Dmitrii Chemodanov, Prasad Calyam, and Kannappan Palaniappan 19.1 Introduction 473 19.1.1 How Can Geospatial Video Analytics Help with Disaster-Incident Situational Awareness? 473 19.1.2 Fog Computing for Geospatial Video Analytics 474 19.1.3 Function-Centric Cloud/Fog Computing Paradigm 475 19.1.4 Function-Centric Fog/Cloud Computing Challenges 476 19.1.5 Chapter Organization 477 19.2 Computer Vision Application Case Studies and FCC Motivation 478 19.2.1 Patient Tracking with Face Recognition Case Study 478 19.2.2 3-D Scene Reconstruction from LIDAR Scans 480 19.2.3 Tracking Objects of Interest in WAMI 482 19.3 Geospatial Video Analytics Data Collection Using Edge Routing 484 19.3.1 Network Edge Geographic Routing Challenges 484 19.3.2 Artificial Intelligence Relevance in Geographic Routing 486 19.3.3 AI-Augmented Geographic Routing Implementation 487 19.4 Fog/Cloud Data Processing for Geospatial Video Analytics Consumption 490 19.4.1 Geo-Distributed Latency-Sensitive SFC Challenges 491 19.4.2 Metapath-Based Composite Variable Approach 492 19.4.3 Metapath-Based SFC Orchestration Implementation 495 19.5 Concluding Remarks 496 19.5.1 What Have We Learned? 496 19.5.2 The Road Ahead and Open Problems 497 References 498 20 An Insight into 5G Networks with Fog Computing 505Osman Khalid, Imran Ali Khan, Rao Naveed Bin Rais, and Asad Waqar Malik 20.1 Introduction 505 20.2 Vision of 5G 507 20.3 Fog Computing with 5G Networks 508 20.3.1 Fog Computing 508 20.3.2 The Need of Fog Computing in 5G Networks 508 20.4 Architecture of 5G 508 20.4.1 Cellular Architecture 508 20.4.2 Energy Efficiency 510 20.4.3 Two-Tier Architecture 512 20.4.4 Cognitive Radio 512 20.4.5 Cloud-Based Architecture 513 20.5 Technology and Methodology for 5G 514 20.5.1 HetNet 515 20.5.2 Beam Division Multiple Access (BDMA) 516 20.5.3 Mixed Bandwidth Data Path 516 20.5.4 Wireless Virtualization 516 20.5.5 Flexible Duplex 518 20.5.6 Multiple-Input Multiple-Output (MIMO) 518 20.5.7 M2M 519 20.5.8 Multibeam-Based Communication System 520 20.5.9 Software-Defined Networking (SDN) 520 20.6 Applications 521 20.6.1 Health Care 521 20.6.2 Smart Grid 521 20.6.3 Logistic and Tracking 521 20.6.4 Personal Usage 521 20.6.5 Virtualized Home 522 20.7 Challenges 522 20.8 Conclusion 524 References 524 21 Fog Computing for Bioinformatics Applications 529Hafeez Ur Rehman, Asad Khan, and Usman Habib 21.1 Introduction 529 21.2 Cloud Computing 531 21.2.1 Service Models 532 21.2.2 Delivery Models 532 21.3 Cloud Computing Applications in Bioinformatics 533 21.3.1 Bioinformatics Tools Deployed as SaaS 533 21.3.2 Bioinformatics Platforms Deployed as PaaS 535 21.3.3 Bioinformatics Tools Deployed as IaaS 535 21.4 Fog Computing 537 21.5 Fog Computing for Bioinformatics Applications 539 21.5.1 Real-Time Microorganism Detection System 541 21.6 Conclusion 543 References 543 Index 547

Summarizes the current state and upcoming trends within the area of fog computing Written by some of the leading experts in the field, Fog Computing: Theory and Practice focuses on the technological aspects of employing fog computing in various application domains, such as smart healthcare, industrial process control and improvement, smart cities, and virtual learning environments. In addition, the Machine-to-Machine (M2M) communication methods for fog computing environments are covered in depth. Presented in two parts—Fog Computing Systems and Architectures, and Fog Computing Techniques and Application—this book covers such important topics as energy efficiency and Quality of Service (QoS) issues, reliability and fault tolerance, load balancing, and scheduling in fog computing systems. It also devotes special attention to emerging trends and the industry needs associated with utilizing the mobile edge computing, Internet of Things (IoT), resource and pricing estimation, and virtualization in the fog environments. Includes chapters on deep learning, mobile edge computing, smart grid, and intelligent transportation systems beyond the theoretical and foundational conceptsExplores real-time traffic surveillance from video streams and interoperability of fog computing architecturesPresents the latest research on data quality in the IoT, privacy, security, and trust issues in fog computing Fog Computing: Theory and Practice provides a platform for researchers, practitioners, and graduate students from computer science, computer engineering, and various other disciplines to gain a deep understanding of fog computing.