Studies in Big Data 76

Sudeep Tanwar   Editor

Fog Data Analytics for IoT ApplicationsNext Generation Process Model with State of the Art Technologies

Studies in Big Data

Volume 76

Series Editor

Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland

The series “Studies in Big Data” (SBD) publishes new developments and advancesin the various areas of Big Data- quickly and with a high quality. The intent is tocover the theory, research, development, and applications of Big Data, as embeddedin the fields of engineering, computer science, physics, economics and life sciences.The books of the series refer to the analysis and understanding of large, complex,and/or distributed data sets generated from recent digital sources coming fromsensors or other physical instruments as well as simulations, crowd sourcing, socialnetworks or other internet transactions, such as emails or video click streams andother. The series contains monographs, lecture notes and edited volumes in BigData spanning the areas of computational intelligence including neural networks,evolutionary computation, soft computing, fuzzy systems, as well as artificialintelligence, data mining, modern statistics and Operations research, as well asself-organizing systems. Of particular value to both the contributors and thereadership are the short publication timeframe and the world-wide distribution,which enable both wide and rapid dissemination of research output.

** Indexing: The books of this series are submitted to ISI Web of Science, DBLP,Ulrichs, MathSciNet, Current Mathematical Publications, Mathematical Reviews,Zentralblatt Math: MetaPress and Springerlink.

More information about this series at http://www.springer.com/series/11970

http://www.springer.com/series/11970

Sudeep TanwarEditor

Fog Data Analytics for IoTApplicationsNext Generation Process Model withState of the Art Technologies

123

EditorSudeep TanwarDepartment of Computer Scienceand EngineeringInstitute of Technology, Nirma UniversityAhmedabad, Gujarat, India

ISSN 2197-6503 ISSN 2197-6511 (electronic)Studies in Big DataISBN 978-981-15-6043-9 ISBN 978-981-15-6044-6 (eBook)https://doi.org/10.1007/978-981-15-6044-6

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer NatureSingapore Pte Ltd. 2020This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whetherthe whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse ofillustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, andtransmission or information storage and retrieval, electronic adaptation, computer software, or by similaror dissimilar methodology now known or hereafter developed.The use of general descriptive names, registered names, trademarks, service marks, etc. in thispublication does not imply, even in the absence of a specific statement, that such names are exempt fromthe relevant protective laws and regulations and therefore free for general use.The publisher, the authors and the editors are safe to assume that the advice and information in thisbook are believed to be true and accurate at the date of publication. Neither the publisher nor theauthors or the editors give a warranty, expressed or implied, with respect to the material containedherein or for any errors or omissions that may have been made. The publisher remains neutral with regardto jurisdictional claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,Singapore

https://doi.org/10.1007/978-981-15-6044-6

Preface

Through the exponential growth of sensors and smart gadgets (collectively referredto as smart devices or Internet of Things (IoT) enabled devices), a significantamount of heterogeneous and multi-modal data, termed as Big Data (BD), is beinggenerated. To deal with such BD, we require efficient and effective solutions suchas data mining, analytics, and reduction to be performed at the edge of fog deviceson a cloud environment. Existing research and development efforts generally focuson performing BD analytics, overlooking the difficulty of facilitating fog dataanalytics (FDA). This book discusses the unique nature and complexity of FDA anddevelops a comprehensive taxonomy (divided into different chapters) abstractedinto a process model. The proposed model addresses various research challenges,such as accessibility, scalability, fog node communication, nodal collaboration,heterogeneity, reliability, and quality of service (QoS) requirements. To demon-strate the proposed process model, we have included some case studies. The mainfeature of this book is to consider all aspects required to manage the complexity ofFDA for IoT applications and also develops a comprehensive taxonomy. This bookfocuses on FDA in IoT and the requirements related to Industry 4.0. It provides acomprehensive taxonomy for FDA abstracted into a novel process model reflectingFDA over IoT. This taxonomy helps the readers to know about the sources andfeatures of FDA. The main benefits to the readers are as follows:

• It contains case studies to demonstrate the process model, which makes thereaders aware of future challenges associated with the FDA, especially for IoTapplications.

• It includes the layered architecture of FDA and also compares the life cycle ofboth big data and FDA.

The book is organized into five sections. The first section is focused on theintroduction and background of the FDA for IoT applications, which includes fivechapters. The second section discusses the emerging technologies and architectureof the FDA, which has three chapters. The third part illustrates the role of IoT

v

applications in the FDA with well-structured four chapters. The fourth sectionhighlights security issues, research challenges, and opportunities, which has fourchapters. Finally, the last section focuses on FDA application in Healthcare 4.0 withfour chapters.

Part: Introduction and Background of FDA

The chapter “Introduction” presents an overview of the FDA. The major aim of thischapter is to provide a bird’s eye view of the usage of cloud computing and fogcomputing technologies used for the FDA. This chapter also gives a comparativeanalysis of the existing data analytics techniques used in the FDA. Moreover, itdiscusses different researchers’ views about the FDA in detail. This chapter alsodiscusses the importance of the FDA in several IoT-based applications such ashealth care and smart grid and highlights the future research challenges for theresearchers/readers working in the same field.

The chapter “Introduction to Fog Data Analytics for IoT Applications” high-lights the role of fog computing to improve the IoT/Cloud paradigm. This chapterintroduces the origin of fog computing, the role of fog computing in IoT applica-tions, why to use fog computing, and the architecture of fog computing. Then, itemphasizes how fog computing works, characteristics of fog computing, compar-ison of fog computing with cloud computing, the difference between fog and edgecomputing, and advantages and limitations of fog computing with its applications.This chapter provides insights into possible research directions with an overallconcept of fog computing in the context of IoT applications.

The chapter “Fog Data Analytics: Systematic Computational Classification andProcedural Paradigm” discusses the characteristic attributes and analyzes thealgorithmic complexities in FDA. A systematic computational classification isgenerated to develop a paradigmatic procedure for FDA to process, store, andanalyze data efficiently and effectively. It also develop an ideal platform for theproliferating sensor-based devices and services on IoT applications. In the end, afew case studies have been discussed, which benefit from the proposed model, andthe study is concluded with future scope for the researchers.

The chapter “Fog Computing: Building a Road to IoT with Fog Analytics”discusses the relevance of the FDA in the area of IoT with its issues and challenges.This chapter majorly focuses on the basics of the FDA. Then, the different char-acteristics of cloud and fog computing platforms are explained. Also, a detailedarchitecture of both the platforms is introduced with a comparative analysis. On thefog server, the FDA tool performs data localization. All the methods of applicationmanagement, such as resource coordination technique, distributed applicationdeployment, and distributed data flow method, are discussed. Further, research

vi Preface

direction using Deep Learning to Big Data is discussed in detail to improve theformulation of data abstractions and dimensionality reduction, along with theirpossible solutions.

The chapter “Data Collection in Fog Data Analytics” highlights the collectiontechniques and management of data. It discusses how data differs in scenarios andthe various methods of data collection in the FDA like node-based segregation,which reduces the requirement of a large number of fog nodes to be set up and theoverloading of these nodes. This study explores the techniques wherein raw andpassive forms of data can be made to evolve and become meaningful with reducedsize, indulge on how Bluetooth low-energy technology can be used to processcollected data through gateways, and use data collectors with wireless low-poweredsensing systems. Finally, this chapter discusses various case studies related toMoving Vehicles, Industrial Automation, Underwater Data Collection, WaterConservation in Agriculture, Indoor Air Quality Monitoring, Health MonitoringSystem, Telehealth Big Data, and Healthcare 4.0 related to FDA.

Part: Emerging Technologies and Architecture for FDA

The chapter “Mobile FOG Architecture Assisted Continuous Acquisition ofFetal ECG Data for Efficient Prediction” presents an overview of FDA and dis-cusses the architecture for continuous monitoring of Fetal Electrocardiogram(fECG) from maternal ECG to avoid any kind of acute condition caused to thenewborn child at the time of birth. The continuous acquisition of fECG will lead toa very large amount of data to send over the cloud for further examination by thedoctor; this data has to be preprocessed before storing it to the cloud for much fasterand efficient evaluation. The proposed architecture has the potential to extend andvirtualize new and efficient healthcare processes for fetal health monitoring, alongwith the Healthcare 4.0 environment and mobile fog computing.

The chapter “Proposed Framework for Fog Computing to Improve Quality-of-Service in IoT Applications” enlightens a framework that aims to improvequality of service (QoS) by providing reduced latency and load balancing at the foglayer. This improvement in QoS is achieved with the help of data aggregation andload balancing. In the proposed framework, an overburdened fog node requests itsneighboring node to share its load. Additionally, it suggests implementing varioustechniques to aggregate data ahead of transmission. Resultantly, the study improvesQoS by outperforming the existing approaches by preventing bottlenecks in the fognetwork.

The chapter “Fog Data Based Statistical Analysis to Check Effects of Yajna andMantra Science: Next Generation Health Practices” provides a case study for fogdata-based statistical analysis and verifies the effect of Yajna and Mantra. For thisstudy, Havan Samgri is specially designed for Asthmatic patients with a 3:1 ratio.Then, Surya Gayatri Mantra chanting has been performed 24 times, and Nadi

Preface vii

Shodhan Pranayam has been executed for half an hour duration daily. Subjects(Users) have taken kwath of Hawan Samagri twice in a day. Lung Function Test(LFT) has been experimented to check the efficacy of Yagyopathy with threeparameters, i.e., FVC-Forced vital capacity (calculates the capacity of humanlungs), FEV-Forced expiratory volume, and FEV1/FVC known as MER-measuredexpiratory rate (directly related to the proportion of lung size exhaled per second).This study demonstrated a significantly improved performance in lung functionusing FDA.

Part: Role of IoT in FDA

The chapter “Process Model for Fog Data Analytics for IoT Applications” presentsa process model that describes the process flow of data analytics using fog com-puting and various modules as used in the fog computing architecture. It discussesthe use case of the FDA for IoT-based Healthcare applications and concludes with acase study, which highlights the current challenges of fog computing for a betteradoption of the technology.

In the chapter “Medical Analytics Based on Artificial Neural Networks UsingCognitive Internet of Things”, a Cognitive Radio network is simulated for opti-mization of spectrum sensing and energy detection. Moreover, two effective clas-sification methods are evaluated on remotely measured physiological parameters,such as blood pressure and heart rate, of patients with two types of diseases—chronic kidney disease and heart disease. Using the proposed framework, thepatients’ blood pressure values, after being measured, can be used by doctors andhospitals to predict the heart rate for heart disease patients and blood glucose(sugar) for chronic kidney patients remotely. This type of remote patient monitoringwith machine-learning-based disease state prediction can be beneficial for deter-mining patient’s disease remotely using their real-time bio-signal measurements.

The chapter “Application of IoT-Based Smart Devices in Health Care Using FogComputing” explores the field of fog computing, cloud computing, and IoTapplications and the importance of the quality of service in the healthcare industry.It presents the working environment and an integrated architecture of fog com-puting with IoT. The author highlights the importance of fog computing with IoT inthe healthcare sector with the help of various services and applications. At last, acase study discusses various issues and challenges in the adoption of IoT-baseddevices in health care.

The chapter “Data Reduction Techniques in Fog Data Analytics for IoTApplications” details the fundamental issues related to the FDA for IoT applica-tions. Then, it explores the various data reduction techniques with fog computingfor IoT applications. These techniques includes; Missing Values Ratio, LowVariance Filter, High Correlation Filter, Principal Component Analysis (PCA),Random Forest/Ensemble Trees, Backward Feature Elimination, and Forward

viii Preface

Feature Construction. Further, a case study with the PCA method for FDA isdiscussed in the end of the chapter, which demonstrates the effectiveness of datareduction methods in FC.

Part: Security Issues, Research Challenges, and Opportunities

The chapter “Background and Research Challenges for Fog Data Analytics and IoT”introduces FDA and describes the challenges and issues related to it. This studypresents an explanation of the advantages and disadvantages of the existing system,for instance, cloud computing, which motivates to incorporate Fog and its methodsto solve cloud computing issues. Then, it describes the various research workscarried out in FDA to overcome the problems of resource discovery, sharing, pathimprovement, and self-organization. FDA depends on other technologies for stor-age, communication, and processing to develop rapidly like a communication net-work, i.e., 4G and 5G. So, it still requires research work and attention to reach its fullpotential.

The chapter “Behavior-Based Approach for Fog Data Analytics: An ApproachToward Security and Privacy” discusses advanced security mechanisms for FDA.This chapter assures security and privacy in fog architecture by reducing the errorrate in the proposed security strategy. There are a lot of biometric-based techniquesthat have been developed using face, palm print, fingers, eyelids, and so on. Butthere are very few works available in the field of typing behavior characteristics. Toaddress the aforementioned issue, a security strategy for fog computing is designedby analyzing a user’s typing behavior pattern. For this study, nine behaviorparameters are deployed and the error rate is evaluated at each step. Results alsoshow that Crossover Error Rate (CER) reduces to 2% for the final stage using theproposed strategy. The proposed scheme is validated by a simulator designed forregistering a new users and identifying their request to get the access of services.

The chapter “Data Security and Privacy Functions in Fog Data Analytics”highlighted the reasons for the susceptibility of fog nodes. The security systems andconcerns in cloud and fog computing are compared in the chapter. The chapter alsoexamines the various types of attacks to which the fog network is vulnerable suchas man in the middle attacks, authentication threats, distributed denial of service,and others. Finally, the chapter aims at investigating the different methods to handlethese types of attacks. The first approach is the prevention of security attacks, whichincludes techniques like identity authentication, access control, and cryptographicschemes. The second approach for data privacy and security handling is to thedetection of attacks along with their various methods like intrusion detection, dataintegrity check, and network traffic analysis to handle the aforementioned issues.The last approach covered is recovery from attacks, which covers recoveryschemes. Thus, the chapter intends to provide a multi-faceted understanding of datasecurity and privacy in the FDA.

Preface ix

The chapter “Data Security and Privacy Functions in Fog Computing forHealthcare 4.0” highlights data security and privacy issues of fog computing inHealthcare industry 4.0. Then, it discusses various services provided by fog com-puting such as encryption, data availability, and traffic analysis using IoT. It alsoemphasizes relevant cloud/fog security design principles, software requirements forFDA, and various risk issues, for example, confidentiality, privacy, and compliancerisks. This chapter is important to understand the concepts of Data Security andPrivacy Functions in Fog Computing for Healthcare 4.0.

Part: FDA Applications in Health Care

The chapter “Fog Data Analytics and Healthcare 4.0” presents a deep insight intocloud computing, fog computing, and implications of fog computing acrossdomains with the imminent prevalence of the IoT devices. Mostly, the benefits ofcloud computing have been reaped by many of the large and small technologicalfirms, by providing services like data and file storage, hosting websites, etc., butwith the advent of IoT devices and Fog Computing, the doors are open for a widevariety of disciplines such as smart cities and health care. This chapter discusses thebenefits of fog computing over Cloud Computing and its applicability in Healthcare4.0 by presenting a case study. Finally, it concludes the study with future researchchallenges of IoT and fog computing.

The chapter “Fog Data Processing and Analytics for Health Care-Based IoTApplications” discusses the Healthcare-based IoT application, cloud computing, fogdata processing, FDA, and their integration and importance. A Literature surveyinvolving all the works that include fog and IoT is discussed. Case studiesinvolving fog and IoT in healthcare systems are also presented to provide light onhow fog and IoT eliminate pressures on healthcare systems that require real-timeprocessing. FDA deployment can be effective in remote monitoring system,equipment monitoring, and smart equipment maintenance. Hence, the researchissues and challenges in FDA are some prominences at last.

The chapter “The Importance of Fog Computing for Healthcare 4.0-Based IoTSolutions” highlights the amalgamation of cloud computing, fog computing, andIoT in Healthcare 4.0. The fog environment is created to address issues, which areoverlooked by the cloud computing model. It is an extension to cloud computing tokeep the devices close to the data center and cloud storage. The devices acting as anintermediary between the fog and cloud layers are known as fog nodes. Theyprovide limited information to the end-user or client for the appropriateness of hazegadgets and portals in the automation of any real-time application such as healthcare and smart cities.

x Preface

The last chapter “Conclusion” wraps up the FDA with the findings, futureresearch challenges, and directions in IoT applications, for instance, health care andsmart cities. This chapter shows that FDA critically contributes to thedecision-making of real-time IoT-based applications.

The editor is very thankful to all the members of Springer Private Limited,especially Mr. Aninda Bose, for the given opportunity to edit this book.

Ahmedabad, Gujarat, India Dr. Sudeep Tanwar

Preface xi

Contents

Introduction and Background of FDA

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3Aparna Kumari, Rajesh Gupta, and Sudeep Tanwar

Introduction to Fog Data Analytics for IoT Applications . . . . . . . . . . . . 19Puneet Kansal, Dilip Sharma, and Manoj Kumar

Fog Data Analytics: Systematic Computational Classificationand Procedural Paradigm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39D. Pradeep Kumar, R. Hanumantharaju, B. J. Sowmya, K. N. Shreenath,and K. G. Srinivasa

Fog Computing: Building a Road to IoT with Fog Analytics . . . . . . . . . 59Avinash Kaur, Parminder Singh, and Anand Nayyar

Data Collection in Fog Data Analytics . . . . . . . . . . . . . . . . . . . . . . . . . . 79S. R. Mani Sekhar, Snehil Tewari, Haaris Rahman, and G. M. Siddesh

Emerging Technologies and Architecture for FDA

Mobile FOG Architecture Assisted Continuous Acquisition of FetalECG Data for Efficient Prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107Anupam Bhardwaj, Pooja Khanna, and Sachin Kumar

Proposed Framework for Fog Computing to ImproveQuality-of-Service in IoT Applications . . . . . . . . . . . . . . . . . . . . . . . . . . 123Rakhi Akhare, Monika Mangla, Sanjivani Deokar, and Vaishali Wadhwa

Fog Data Based Statistical Analysis to Check Effects of Yajnaand Mantra Science: Next Generation Health Practices . . . . . . . . . . . . . 145Rohit Rastogi, Mamta Saxena, D. K. Chaturvedi, Santosh Satya,Navneet Arora, Mayank Gupta, and Parul Singhal

xiii

Role of IoT in FDA

Process Model for Fog Data Analytics for IoT Applications . . . . . . . . . 175Anjali Modi, Shreena Jani, Karansingh Chauhan, and Jitendra Bhatia

Medical Analytics Based on Artificial Neural NetworksUsing Cognitive Internet of Things . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199Himani Bedekar, Gahangir Hossain, and Ayush Goyal

Application of IoT-Based Smart Devices in Health CareUsing Fog Computing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263Satyasundara Mahapatra and Anupam Singh

Data Reduction Techniques in Fog Data Analytics for IoTApplications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279Srinidhi Hiriyannaiah, Zaifa Khan, Aniket Singh, G. M. Siddesh,and K. G. Srinivasa

Security Issues, Research Challenges, and Opportunities

Background and Research Challenges for Fog Data Analyticsand IoT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313Ansh Riyal, Geetansh Kumar, and Deepak Kumar Sharma

Behavior-Based Approach for Fog Data Analytics: An ApproachToward Security and Privacy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341Urvashi, Lalit K. Awasthi, and Geeta Sikka

Data Security and Privacy Functions in Fog Data Analytics . . . . . . . . . 355Apoorva Bhagat, Srishty Mittal, Uzma Faiz, and Deepak Kumar Sharma

Data Security and Privacy Functions in Fog Computingfor Healthcare 4.0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387Darpan Anand and Vineeta Khemchandani

FDA Applications in Health Care

Fog Data Analytics and Healthcare 4.0 . . . . . . . . . . . . . . . . . . . . . . . . . 423Madhurima Hooda, Shashwat Pathak, and Shreyans Pathak

Fog Data Processing and Analytics for Health Care-Based IoTApplications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445Tarjni Vyas, Shivani Desai, and Anand Ruparelia

The Importance of Fog Computing for Healthcare 4.0-Based IoTSolutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471U. Hariharan and K. Rajkumar

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495Rajesh Gupta, Aparna Kumari, and Sudeep Tanwar

xiv Contents

About the Editor

Dr. Sudeep Tanwar is an Associate Professor in the Computer Science andEngineering Department at Institute of Technology, Nirma University, Ahmedabad,Gujarat, India. He is visiting Professor in Jan Wyzykowski University in Polkowice,Poland and University of Pitesti in Pitesti, Romania. He received B.Tech in 2002from Kurukshetra University, India, M.Tech (Honor’s) in 2009 from Guru GobindSingh Indraprastha University, Delhi, India and Ph.D. in 2016 with specialization inWireless Sensor Network. He has authored or coauthored more than 130 technicalresearch papers published in leading journals and conferences from the IEEE,Elsevier, Springer, Wiley, etc. Some of his research findings are published in topcited journals such as IEEE TNSE, IEEE TVT, IEEE TII, IEEE Access, ComputerCommunication, Applied Soft Computing, Journal of Parallel and DistributedComputing, Emerging Transactions on Telecommunication, Journal of Network andComputer Application, Pervasive and Mobile Computing, International Journal ofCommunication System, Telecommunication System, Computer and ElectricalEngineering and IEEE Systems Journal. He has also contributed 10 edited/authoredbooks with International/National Publishers like IET, Springer. He has guided manystudents leading to M.E./M.Tech and guiding students leading to Ph.D. He isAssociate Editor of IJCS,Wiley and Security and Privacy Journal, Wiley. His currentinterest includes Wireless Sensor Networks, Fog Computing, Smart Grid, IoT, andBlockchain Technology. He was invited as Guest Editors/Editorial Board Membersof many International Journals, invited for keynote Speaker in many InternationalConferences held in Asia and invited as Program Chair, Publications Chair, PublicityChair, and Session Chair in many International Conferences held in North America,Europe, Asia and Africa. He has been awarded best research paper awards fromIEEE GLOBECOM 2018, IEEE ICC 2019, and Springer ICRIC-2019.

xv

Introduction and Background of FDA

Introduction

Aparna Kumari, Rajesh Gupta, and Sudeep Tanwar

Abstract In the recent couple of years, smart devices and sensors in IoT applicationsare growing drastically and generating an extensive amount of multi-modal andheterogeneous data, designated as Big Data (BD). BD requires several intelligentcomputation systems such as data mining and data analytics to handle BD-relatedstorage challenges on a cloud. Presently, Cloud Computing (CC) is comprehensivelyused in the industry to handle BD challenges, but it raises various issues such aslatency, security, and high cost for data management. The promising technology FogComputing (FC) facilitates at the edge of a cloud to handle the aforementioned issues.The analytics on fog data generated by diverse IoT devices is one of the challengingtasks. In this chapter, we discourse on the complexity and uniqueness of Fog DataAnalytics (FDA). A detailed discussion on FDA architecture is abstracted with theinnovative process model. This chapter highlights the various attributes of the FDAfor IoT applications and discusses the FDA classification like fog data collection,storage, and analytics on it. The proposed FDA process model addresses numerousresearch challenges, such as scalability, accessibility, heterogeneity, reliability, nodalcollaboration, and quality of service (QoS) with future research directions.

Keywords Fog data analytics · Cloud computing · Big data · Data analysis ·Internet of Things · Fog computing

A. Kumari · R. Gupta · S. Tanwar (B)Nirma University, Ahmedabad, Gujarat, Indiae-mail: sudeep.tanwar@nirmauni.ac.in

A. Kumarie-mail: 17ftphde22@nirmauni.ac.in

R. Guptae-mail: 18ftvphde31@nirmauni.ac.in

© The Editor(s) (if applicable) and The Author(s), under exclusive licenseto Springer Nature Singapore Pte Ltd. 2020S. Tanwar (ed.), Fog Data Analytics for IoT Applications, Studies in Big Data 76,https://doi.org/10.1007/978-981-15-6044-6_1

3

http://crossmark.crossref.org/dialog/?doi=10.1007/978-981-15-6044-6_1&domain=pdfmailto:sudeep.tanwar@nirmauni.ac.inmailto:17ftphde22@nirmauni.ac.inmailto:18ftvphde31@nirmauni.ac.inhttps://doi.org/10.1007/978-981-15-6044-6_1

4 A. Kumari et al.

1 Introduction

Since the inception of the Information and Communications Technology (ICT) andInternet of Things (IoT), a fast increase of data is being generated. This is bettertermed as Big Data (BD); the aggregation and processing of BD have become a fore-most research concern. The IoT comprises various smart devices such as smart cars,smartphones, sensors (sensing physical environment), and others, which are inter-connected to each other through the communication network without any humaninteraction. The ballooning of IoT devices over cloud-based applications causedseveral BD challenges. The various data storage-associated challenges have beenhandled through Cloud Computing (CC) technology and it is extensively used inthe industry using the pay-as-you-go scheme [1]. It discourses the BD challenges,for instance, latency and high cost due to its data distribution management and scal-ability approach [2]. Even though CC provides a flexible and scalable system fordata analytics, security challenges between cloud and local assets raise the down-time and other factors [3, 4]. Nonetheless, to handle the aforementioned issues, theemergent technology fog computing (FC) (proposed by Cisco) plays a vital role dueto its decentralized architecture over the cloud [5]. A comparative analysis betweenCC and FC based on the numerous parameters like data communication cost andnetwork bandwidth requirement and many more are shown in Table1. This table hashighlighted the benefits of the usage of FC along with CC (FC+CC) over only the CCsystem for IoT applications. Unlike CC, FC handles the latency-sensitive real-timeapplications on the verge of a network [6]. FC possesses a substantial processingpower at edge nodes to perform the computation of a huge volume of data on theirown (without going through distant servers). The FC has Cloudlets (a small scaledata centers), which supports data-intensive IoT-based devices for low latency duringthe data processing and transfer. FC divests the cloud servers to enable large datastorage and increase the processing power of the entire organization. FC requiresnumerous intelligent solutions for decision-making like data analytics, data mining,and reduction in the edge of the fog devices over the cloud [7, 8].

Further, Fog Data Analytics (FDA) is a consolidation of the CC, FC, and analyt-ics on data at the cloud end [9]. The elementary functionality of Fog is its abilityto improve data distribution and reduce latency [10]. A distinctive FDA life cyclecomprises several functions like BD storage, BD distribution, and visualization ofBD, as shown in Fig. 1. In Fig. 1, the FDA architecture is divided into three layers;these layers are End User Layer, Fog Layer (FL), and Cloud Layer (CL). The firstlayer is the End User layer, where data is collected from various IoT devices likesensors and actuators to store at the edge devices over a cloud. The collected datais processed, and analytics performed on it;only after that, only relevant data is sentto the cloud to handle the issues of BD [11]. The bottom (End User) layer containsusers and sensor/IoT devices that are linked to the cloud using a communicationchannel. Data extracted here are forwarded to the next layer, which is an FL. FLcompromises various Fog Nodes (FNs) to handle the priority-based requests [12].

Introduction 5

Table 1 A comparative analysis between FC and CC

Features Cloud Fog

Network bandwidthrequirement

High Improved

Data communication cost High Low

Data storage capacity High Low

Data access latency (RTT) High Improved

Fault tolerance No Yes

Data flow All Relevant

System response time High Improved

System and data reliability Manageable Improved

Mobility support of smartdevices

Random Systematic

Security Low Improved

Privacy High Low

Battery lifetime Low Improved

Applications All areas Critical applications

IoT devices communicate to FL-based FNs using fog gateway. Then, FL to CL com-munication takes place through a cloud gateway, which is the utmost top layer in theFDA architecture.

1.1 Internet of Things (IoT) Applications

The development of ICT and digital devices revolutionized these digital devicesto smart devices. The IoT is essentially a network of devices, which can recorddata, share data, and even perform some computation on data using digital appara-tuses, software, sensors, and entrenched chips. Each IoT device has a unique InternetProtocol (IP) address, which helps to establish connections and verify the connec-tion between these smart devices. The communication between these IoT devicesdoes not need any predefined interactions, for instance, Human-to-Computer (H2C),Computer-to-Human (C2H), or Human to Human (H2H). The IoT applications havebeen really effective in a different section of the various industries like health care[13], smart cities, and smart grid with data storage and processing facility [14],although it has several limitations related to BD (generated from IoT devices) suchas small computation capacity and limited data analysis power, which need to bemitigated using appropriate approaches like FDA [7].

6 A. Kumari et al.

Fig. 1 Layered FDA architecture

1.2 Fog Computing and Its Role in FDA

Intel estimates the data generation capacity of the average automated vehicle, whichproduces around 40TB of data in every 8h [15]. In this scenario, FC infrastructure iscommonly provisioned to use the relevant data for specific tasks [16]. The remainingdata that are not timely for the specified task or process are forwarded to the cloud.The cloud consists of the extended computing resources for BD storing (edge devicesproduce but do not use) [17]. Then, the cloud provides supplementary computingresources for analytics that make it a complementary ecosystem for FC-based appli-cations. An FC infrastructure consists of various functions and components basedon its real-time application. It includes computing gateways, which accept BD from

Introduction 7

diverse collection endpoints like routers and switches (connecting assets within anetwork) or data sources. FC is a decentralized architecture that comprises com-puting resources and locating these resources near to the data-generation sources.The processing of data close to the edge decreases latency and uses less computingresources [18].

The steps to process BD through the FDA (using FC architecture) in an IoTapplication:

1. The signals from various IoT devices are read by an automation controller.2. The controller executes the system program required to automate the IoT devices.3. Then, the control system program sends data through to standard gateway pro-

tocols or Open Platform Communication (OPC) server. The OPC is basically aninteroperability standard for data exchange.

4. These data are converted into the format to be accepted by Internet-based serviceproviders like HTTP(S) and MQTT.

5. After conversion, data is sent to FNs or IoT gateway. These endpoints receive thedata for analysis or transfer the data to the cloud.

1.3 Process Model for FDA

In the FDA process model, as shown in Fig. 2, the multi-modal heterogeneous BDgenerated from various heterogeneous fog devices is collected and aggregated. Aftercollection and preprocessing, the BD is stored in the storage system and then for-warded for analytics using priority gateways. Then, historic data get processed andanalyzed with the help of routine events at the FNs. In the case of real-time data,the urgent event performs real-time processing and analysis at FNs as per the needof the IoT application [19]. Successively, the BD intelligence on the processed datatakes place, and useful information is retrieved for decision-making. Then, the BDwill be transmitted to other FNs or cloud for further analytics.

1.4 FDA Attributes for IoT Applications

The computational flow in FDA from cloud to fog is quite similar to the data analyticsin CC, with the single variation being the inclusion of the edge devices [20]. Thiscomputational approach for the data analytics defines various attributes of FDA asshown in Fig. 3, and it empowers the extensive development of IoT applications andservices as listed beneath.

8 A. Kumari et al.

Fig. 2 Process model of FDA

Fig. 3 FDA attributes forIoT applications

Introduction 9

1.4.1 Heterogeneity

Heterogeneity comprises hierarchial components that work as building blocks of adistributed architecture of the FDA. FC infrastructure provides the major FDA facil-ity, for instance, data storage, data computation, and networking services betweencore cloud and end devices.

1.4.2 Interoperability

To support the wide range of services to the IoT application, fog devices and FDAwork in an interoperating environment. These services could be real-time data ana-lytics, artificial intelligence, predictive decisions, and data streaming [21, 22].

1.4.3 Real-Time Interaction

The FDA has the competency to work in real time to achieve better QoS, for instance,energy distribution [23] and monitoring systems in smart grid and real-time trafficmonitoring in intelligent transportation systems (ITS).

1.4.4 Cognition

In this scenario, the goal is to become user-centric. The data accessing facility andanalytics based on the user requirements are focused by the FDA to better understandthe need of the user. It also provides the best place to control, store, and transmit thedata throughout the cloud to end devices located at the edge of the IoT application.The IoT application incorporated with FC provide better responses to the user due totheir nearness to end devices and more efficiently reproduce the users’ requirements[24].

1.4.5 Geo-Graphical Environment Distribution

The QoS can handle effectively dynamic and static edge-devices in IoT applications.Its application network consists of geographically dispersed FNs and sensors invarious environments like weather monitoring sensors, temperature monitoring, andhealthcare monitoring [25].

1.4.6 Edge Location with Low Latency

The recent research and development of applications for IoT-based smart deviceshave been recognized as insufficient due to the absence of vicinity between the

10 A. Kumari et al.

devices. For better QoS at the edge of the IoT application, low-latency servicesare required with the application like video streaming TV services and live gamingapplications [26].

1.5 Classification for FDA in IoT Application

1.5.1 Data Collection

There are various data collection systems and components that exist in the data ana-lytics ecosystem. These data collection systems provided duplicate data, erroneousdata, and missing values, which need a faithful preprocessing. Following are theprevalent components for the data collection:

– IoT Devices: The IoT has opened a door in the field of ICT with a combination ofthe computer and physical eco-sphere, i.e., IoT devices. It enhances proficiencyand accurateness as it reduces human intervention. A few of the examples aresmart cities, smart homes, smart grids, virtual power plants, ITS, and many more[27]; as each element is uniquely recognizable within the network, it provides theglobal connectivity.

– Sensory Devices: It is the core device used for the data collection when it refersto automatic data collection (without human intervention) for cloud-based IoTapplications. These data generated from various sources are characterized as voice,image, vibration, weather, pressure, temperature, voice, current, vehicles, and soon in a specific time interval.

These huge data (BD) are transferred primarily through a Local AreaNetwork (LAN)or a wireless network for BD collection, storage, and processing. The FDA pro-visions better communication and services by establishing a strong network con-nection through LAN/WAN (5G/4G) [28] or wireless network technologies (Blue-tooth/ZigBee) [29], and their edge capacities at FN are as shown in Figs. 4 and 5.Figure4 illustrates the various network technologies such as Wi-Fi and ZigBee toempower FDA communication between edge-devices. The sensory devices containAPI through which data interacts with IoT applications by using their unique IPaddress. There are other potential data sources, such as social media websites andsocial networking API.

1.5.2 Data Storage

Data storage is majorly categorized into three categories—Clustering, Indexing, andReplication. In clustering, a group of data is collected and stored into the fog storagedevices. In indexing, the data has comparative indexations for quick retrieval andfast access. In real-time IoT applications, complex data are streamed by combining

Introduction 11

Fig. 4 Empowering network technologies for FDA communication

Fig. 5 Edge capacity of wireless models in FDA

a sequential approach for newly arrived data and an indexing approach for old data.In the last category, replication, the same data is simulated over other machines tohandle fault tolerance.

1.5.3 Data Processing

The FC has a well-known distributed architecture for data exchange and processing.In this FC architecture, most of the systems are controlled remotely and others arehandled at the edge of the cloud. FC tends to reduce the volume of data that is passed

12 A. Kumari et al.

on to CC for storage, processing, and analysis [30]. The data processing in an FCsystem happens in the smart devices that lie at the edge of the application network.It also ensures data security and privacy, and only relevant data is transferred to thecloud [31, 32]. The most vital attribute of the FDA is its capability of filtering datathat needs to be processed at the cloud layer. There are three basic components: (i)IoT Verticals (inhabitant applications or products—Smart Devices such as commu-nicators, external interface, controllers, and so on), (ii) Orchestration Layer (consistsof predefined functions, for instance, data migration, data sharing, decision-making,and policy supervision), and (iii) Abstraction Layer (provides a uniform interface tothe end-user, e.g., generic API).

1.5.4 Data Analysis

The analysis part is different from conventional computing methods in the FDA asit combines analytic on historical data and real-time analytics as well. In the IoTrebellion, every industry is moving toward huge volumes of data for processing andreal-time analytics to achieve better QoS. This real-time analytics can be facilitatedusing the FC architecture. For example, the smart grid industry leverages deep BDanalysis for prediction of power failures, power consumption, energy prices, anddemand forecasting to achieve higher productivity and better QoS.

1.6 FDA Research Challenges and Future Direction

The FDA discussion for IoT applications in the context of real-time access, het-erogeneity, and predominantly interoperability requires BD storage and processing.Thoughmany solutions exist to handle this, they are not capable enough as comparedto the exponential data growth rate. Presently, minimal software and tools exist toaddress FDA issues for IoT applications. Hence, potential research challenges andissues are identified as delineated in Fig. 6, which demand the development of aprocess model for FDA life cycle. The hunt for these challenges will expedite theencroachment of the FDA.

1.6.1 Heterogeneity in Fog

Fog is located close (relatively at the edge) to the end devices which are connectedvia a fog gateway. The responsibility of the fog gateway is to manage and maintainthe heterogeneous connectivity between the fog and the end devices. Emergent pop-ular technologies like network virtualization (NV), network function virtualization(NFV), and software-defined network (SDN) can be used to maintain the networkefficiently and increase network reliability and scalability.

Introduction 13

Fig. 6 Research challenges in FDA communication

1.6.2 Data Reliability

Reliability is one of the backbone characteristics of the network, which ensuresmaximum network availability. To achieve this, a network needs to perform periodicchecks and locate the failure points, so that the failed components again start working.As the fog network is highly dynamic, it is not feasible to do checks and reschedulingof failed components periodically. The other reason for the same is the latency ofthe communication network. So, the other possible solution to make the systemfault-tolerant is to make the replicas of the fog nodes.

1.6.3 Storage Capacity

The storage capacity of the fog node is limited compared to the cloud capacity. Fognodes store the data whose requirement is urgent (or critical data). Nowadays, smartend devices are generating a huge amount of data called big data, so fog nodes arenot much capable to store this [33]. This limits the analysis capability also.

14 A. Kumari et al.

1.6.4 Data Communication Latency

A fog-based network needs to perform real-time analysis of data instead of batchprocessing. It requires all resources at the same time for processing the real-timedata, but the storage capacity of a fog node is limited. As all fog nodes are connectedto each other for data and resource sharing, latency is the major issue. High latencyreduces the quality of service (QoS) of the fog network.

1.6.5 Data Distribution

Most of the applications require a trusted data flow between the multiple fog devices,which requires pre-attribute definitions. Data distribution among the fog nodes isessential as the capacity of single fog is not quite limited and is not able to storeample amount of data. It is challenging because of the communication networkproperties such as latency, reliability, and throughput.

1.6.6 Resource Management

In FDA, resource management can be application-aware management and detec-tion and sharing management. Challenges in the first case are end-device mobility,network latency, storage capacity, and network bandwidth. It plays a vital role inmobile sensing applications. The latter case of resource management is for main-taining the performance of the fog network. The main challenge in detection andsharing management is energy harvesting.

1.6.7 Security and Privacy

Presently, very few researchers have focused on the security and privacy issues ofthe FDA. It is quite difficult to achieve the authentication, authorization, and accesscontrol mechanism in the fog environment as the data is distributed among the dif-ferent fog nodes of the heterogeneous network [34]. Each fog node has differentcomputing capabilities, so above all, security and privacy mechanisms are difficultto achieve. One of the possible and efficient solutions is to create a trusted executionenvironment.

1.6.8 Processing Capability

Fog nodes are battery constrained devices with a limited lifetime. So, the processingof big data over the fog nodes is impossible. Limited processing capability can giveincorrect results to the end nodes, which may be harmful to them. This can be solvedusing distributed fog computing.

Introduction 15

2 Conclusion

The explosion of sensors, smart devices, and IoT generate BD drastically in recentyears. The generation of a huge amount of data from these IoT devices demands theefficient storage, processing, and analysis of data with high frequency in real time. Inthis chapter, we have targeted the embryonic intersects of FC and CC technologieswith a distinctive focus on analytics on fog-based BD. The FDA has extensive usagein a variety of IoT applications such as smart cities, satellite imaging, health care,ITS, and smart grid. FDA is important to unlock the full advantages of these IoTapplications in the present society. This chapter provides an introduction to BD com-munication and analytics in FC architecture related to the concepts of conventionalCC. This chapter also discusses and addresses plentiful challenges in the FDA dataprocessing and analysis and develops a broad classification. Further, an advancedlevel of a process model is abstracted through the FDA life cycle for BD computing.This unique process model targets research challenges and practices regarding FDAcomputing in this field of research.

References

1. Tang, B., Chen, Z., Hefferman, G., Wei, T., He, H., Yang, Q.: A hierarchical distributed fogcomputing architecture for big data analysis in smart cities. In: Proceedings of theASEBigData& SocialInformatics 2015, ASE BDSI ’15. Association for Computing Machinery, New York,NY, USA (2015)

2. Taneja, M., Jalodia, N., Davy, A.: Distributed decomposed data analytics in fog enabled iotdeployments. IEEE Access 7, 40969–40981 (2019)

3. Stackpath: What is fog computing? (2018). https://www.stackpath.com/edge-academy/fog-computing/

4. Buyya, R., Srirama, S.N.: Fog Computing Realization for Big Data Analytics, pp. 259–290(2019)

5. Bonomi, F., Milito, R., Zhu, J., Addepalli, S.: Fog computing and its role in the internet ofthings. In: Proceedings of the First Edition of theMCCWorkshop onMobile CloudComputing,MCC ’12, pp. 13–16 (2012). Association for Computing Machinery, New York, NY, USA

6. Roman, R., Lopez, J., Mambo, M.: Mobile edge computing, Fog et al.: a survey and analysisof security threats and challenges. Future Gener. Comput. Syst. 78, 680–698 (2018)

7. What is fog data analytics? how to leverage it for business purposes? (2018). https://datafloq.com/read/fog-data-analytics-leverage-business-purposes/7310

8. He, J.,Wei, J., Chen, K., Tang, Z., Zhou, Y., Zhang, Y.:Multitier fog computingwith large-scaleiot data analytics for smart cities. IEEE Internet of Things J. 5(2), 677–686 (2018)

9. Kumari,A., Tanwar, S., Tyagi, S.,Kumar,N.,Obaidat,M.S.,Rodrigues, J.J.P.C.: Fog computingfor smart grid systems in the 5g environment: challenges and solutions. IEEEWirel. Commun.26(3), 47–53 (2019). June

10. Kumari, A., Tanwar, S., Tyagi, S., Kumar, N., Parizi, R.M., Choo, K.-K.R.: Fog data analytics:a taxonomy and process model. J. Netw. Comput. Appl. 128, 90–104 (2019)

11. Kumari, A., Tanwar, S., Tyagi, S., Kumar, N., Maasberg, M., Choo, K.-K.R.: Multimedia bigdata computing and internet of things applications: a taxonomy and process model. J. Netw.Comput. Appl. 124, 169–195 (2018)

12. Kumari, A., Tanwar, S., Tyagi, S., Kumar, N., Opportunities and challenges: Fog computingfor healthcare 4.0 environment. Comput. Electr. Eng. 72, 1–13 (2018)

https://www.stackpath.com/edge-academy/fog-computing/https://www.stackpath.com/edge-academy/fog-computing/https://datafloq.com/read/fog-data-analytics-leverage-business-purposes/7310https://datafloq.com/read/fog-data-analytics-leverage-business-purposes/7310

16 A. Kumari et al.

13. Gupta, R., Tanwar, S., Tyagi, S., Kumar, N.: Tactile-internet-based telesurgery system forhealthcare 4.0: an architecture, research challenges, and future directions. IEEE Netw. 33(6),22–29 (2019)

14. Badidi, E., Moumane, K.: Enhancing the processing of healthcare data streams using fogcomputing. In: 2019 IEEE Symposium on Computers and Communications (ISCC), pp. 1113–1118 (2019)

15. Mehta, P., Gupta, R., Tanwar, S.: Blockchain envisioned uav networks: challenges, solutions,and comparisons. Comput. Commun. 151, 518–538 (2020)

16. Nelson, P.: Just one autonomous car will use 4,000 gb of data/day (2016). https://www.networkworld.com/article/3147892/one-autonomous-car-will-use-4000-gb-of-dataday.html

17. Prasad, V., Bhavsar, M., Tanwar, S.: Influence of monitoring: fog and edge computing. ScalableComput. 20, 365–376 (2019)

18. Kumar, R., Kalra, M., Tanwar, S., Tyagi, S., Kumar, N.: Min-parent: an effective approachto enhance resource utilization in cloud environment. In: 2016 International Conference onAdvances in Computing, Communication, Automation (ICACCA) (Spring), April 2016, pp.1–6 (2016)

19. Tanwar, S., Bhatia, Q., Patel, P., Kumari, A., Singh, P.K., Hong,W.: Machine learning adoptionin blockchain-based smart applications: the challenges, and a way forward. IEEE Access 8,474–488 (2020)

20. Tanwar, S., Vora, J., Kanriya, S., Tyagi, S., Kumar, N., Sharma, V., You, I.: Human arthritisanalysis in fog computing environment using Bayesian network classifier and thread protocol.IEEE Consum. Electron. Mag. (2018)

21. Ma,B.B., Fong, S.,Millham,R.:Data streammining in fog computing environmentwith featureselection using ensemble of swarm search algorithms. In: 2018 Conference on InformationCommunications Technology and Society (ICTAS), pp. 1–6 (2018)

22. Gupta, R., Tanwar, S., Al-Turjman, F., Italiya, P., Nauman, A., Kim, S.W.: Smart contractprivacy protection using ai in cyber-physical systems: tools, techniques and challenges. IEEEAccess, p. 1 (2020)

23. Jayaraman, P.P., Bártolo Gomes, J., Nguyen, H., Abdallah, Z.S., Krishnaswamy, S., Zaslavsky,A.: Scalable energy-efficient distributed data analytics for crowdsensing applications in mobileenvironments. IEEE Trans. Comput. Soc. Syst. 2(3), 109–123 (2015)

24. Chiang, M., Zhang, T.: Fog and IoT: an overview of research opportunities. IEEE Internet ofThings J. 3(6), 854–864 (2016)

25. Hathaliya, J., Sharma, P., Tanwar, S., Gupta, R.: Blockchain-based remote patientmonitoring inhealthcare 4.0. In: 2019 IEEE 9th International Conference on Advanced Computing (IACC),December 2019, pp. 87–91 (2019)

26. More, P.: Review of implementing fog computing. Int. J. Res. Eng. Technol. 4(06), 335–338(2015)

27. Darwish, T.S.J., Abu Bakar, K.: Fog based intelligent transportation big data analytics in theinternet of vehicles environment:motivations, architecture, challenges, and critical issues. IEEEAccess 6, 15679–15701 (2018)

28. Gupta, R., Tanwar, S., Tyagi, S., Kumar, N.: Tactile internet and its applications in 5g era: acomprehensive review. Int. J. Commun. Syst. 32(14), e3981 (2019). e3981 dac.3981

29. Borresen, J.L., Jensen, C.S., Torp, K.: Fogbat: combining bluetooth and gps data for bettertraffic analytics. In: 2016 17th IEEE International Conference on Mobile Data Management(MDM), vol. 1, pp. 325–328 (2016)

30. Gupta, R., Tanwar, S., Tyagi, S., Kumar, N.:Machine learningmodels for secure data analytics:a taxonomy and threat model. Comput. Commun. 153, 406–440 (2020)

31. Gupta, R., Tanwar, S., Tyagi, S., Kumar, N., Obaidat, M.S., Sadoun, B.: Habits: blockchain-based telesurgery framework for healthcare 4.0. In: 2019 International Conference on Com-puter, Information and Telecommunication Systems (CITS), August 2019, pp. 1–5 (2019)

32. Aljumah, A., Ahanger, T.A.: Fog computing and security issues: a review. In: 2018 7th Interna-tional Conference on Computers Communications and Control (ICCCC), pp. 237–239 (2018)

https://www.networkworld.com/article/3147892/one-autonomous-car-will-use-4000-gb-of-dataday.htmlhttps://www.networkworld.com/article/3147892/one-autonomous-car-will-use-4000-gb-of-dataday.html