Downloaded from http://iranpaper.

ir
http://www.itrans24.com/landing1.html

Accepted Manuscript

The role of big data analytics in Internet of Things

Ejaz Ahmed, Ibrar Yaqoob, Ibrahim Abaker Targio Hashem,

Imran Khan, Abdelmuttlib Ibrahim Abdalla Ahmed,
Muhammad Imran, Athanasios V. Vasilakos

PII: S1389-1286(17)30259-1
DOI: 10.1016/j.comnet.2017.06.013
Reference: COMPNW 6240

To appear in: Computer Networks

Received date: 19 December 2016

Revised date: 3 June 2017
Accepted date: 15 June 2017

Please cite this article as: Ejaz Ahmed, Ibrar Yaqoob, Ibrahim Abaker Targio Hashem, Imran Khan,
Abdelmuttlib Ibrahim Abdalla Ahmed, Muhammad Imran, Athanasios V. Vasilakos, The role of big data
analytics in Internet of Things, Computer Networks (2017), doi: 10.1016/j.comnet.2017.06.013

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service
to our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and
all legal disclaimers that apply to the journal pertain.
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
1

The role of big data analytics in Internet of

Things
Ejaz Ahmed, Ibrar Yaqoob, Ibrahim Abaker Targio Hashem, Imran Khan, Abdel-
muttlib Ibrahim Abdalla Ahmed, Muhammad Imran, and Athanasios V. Vasilakos

T
Abstract—The explosive growth in the number of devices connected to the Internet of Things (IoT) and the exponential

IP
increase in data consumption only reflect how the growth of big data perfectly overlaps with that of IoT. The management
of big data in a continuously expanding network gives rise to non-trivial concerns regarding data collection efficiency, data
processing, analytics, and security. To address these concerns, researchers have examined the challenges associated

CR
with the successful deployment of IoT. Despite the large number of studies on big data, analytics, and IoT, the
convergence of these areas creates several opportunities for flourishing big data and analytics for IoT systems. In this
paper, we explore the recent advances in big data analytics for IoT systems as well as the key requirements for managing
big data and for enabling analytics in an IoT environment. We taxonomized the literature based on important parameters.

US
We identify the opportunities resulting from the convergence of big data, analytics, and IoT as well as discuss the role of
big data analytics in IoT applications. Finally, several open challenges are presented as future research directions.

Index Terms—Internet of things, big data, analytics, distributed computing, smart city.
AN
✦
M

1 I NTRODUCTION number of human beings in the world. These

Internet-connected objects, which include PCs,
The technological advancements and rapid con- smartphones, tablets, WiFi-enabled sensors,
ED

vergence of wireless communication, digital wearable devices, and household appliances,

electronics, and micro-electro-mechanical sys- form the IoT as shown in Figure 1. Reports
tems (MEMS) technologies have resulted in the show that the number of Internet-connected
PT

emergence of Internet of Things (IoT). Accord- devices is expected to increase twofold from
ing to the Cisco report1 , the number of ob- 22.9 billion in 2016 to 50 billion by 2020 as
jects connected to the Internet has exceeded the shown in Figure 2.
CE

Most IoT applications do not only focus on

• E. Ahmed, I. Yaqoob, I.A.T. Hashem, and A.I.A. Ahmed monitoring discrete events but also on mining
are with the Centre for Mobile Cloud Computing Re-
search, Faculty of Computer Science and Information Tech-
the information collected by IoT objects. Most
AC

nology, University of Malaya. (E-mail: ejazahmed@ieee.org, data collection tools in the IoT environment
ibraryaqoob@siswa.um.edu.my, targio@siswa.um.edu.my, and are sensor-fitted devices that require custom
abdelmuttlib@siswa.um.edu.my)
• I. Khan is working in Schneider Electric Industries, 38TEC, protocols, such as message queue telemetry
Grenoble, France. (Email: imran@ieee.org) transport (MQTT) and data distribution service
• M. Imran is with the College of Computer and Informa- (DDS). Given that sensors are used in nearly
tion Sciences, King Saud University, Saudi Arabia. (E-mail:
dr.m.imran@ieee.org) all industries, the IoT is expected to produce a
• Athanasios V. Vasilakos is working with the Department of Com- huge amount of data. The data generated from
puter Science, Electrical and Space Engineering, Lulea University
of Technology, Sweden (e-mail: athanasios.vasilakos@ltu.se)
IoT devices can be used in finding potential
research trends and investigating the impact
1. http:www.cisco.comcdamen usaboutac79docsinnovIoT IBSGof certain events or decisions. These data are
0411FINAL.pdf processed using various analytic tools [1]. Fig-
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
2

form that can assist in consuming and reading

diverse data sources as well as in accelerating
the data integration process becomes vital. Data
integration and analytics allow organizations
to revolutionize their business process. Specif-
ically, these enterprises can use data analytics
tools to transform a huge volume of sensor-
collected data into valuable insights. Given the
overlapping research trends in these areas, this
paper focuses on the recent advances in man-
agement of big data and analytics in the IoT

T
paradigm.

IP
The contributions of this paper are as fol-
lows:

CR
• We critically review the recent literature.
• We discuss big data processing and ana-
lytics platforms in the IoT environment.
We discuss the key requirements for big
Fig. 1: Big Data Sources in IoT US •

•
data processing and analytics in an IoT
environment.
We taxonomized the literature based on
AN
important parameters.
• We discuss the potential opportunities in
big data processing and analytics in the
IoT paradigm and highlight the role of
M

data analytics in IoT applications.

• We discuss the open research challenges
and highlight the vision of big data ana-
ED

lytics in IoT as future research directions.

These contributions are given in separate
sections from 2-8. We provide concluding re-
PT

marks in section 9.
Fig. 2: Number of Internet-Connected Devices1
CE

ure 3 illustrates the process of data collection, 2 R ECENT A DVANCES IN I OT- BASED
monitoring, and data analytics2 . B IG DATA AND A NALYTICS
Although IoT has created unprecedented
AC

Bashir and Gill [3] propose an IoT big data an-

opportunities that can help increase revenue,
alytics framework to overcome the challenges
reduce costs, and ameliorate efficiencies, col-
of storing and analyzing large amount of data
lecting a huge amount of data alone is insuffi-
originating from smart buildings. The proposed
cient. To generate benefits from IoT, enterprises
framework is composed of three components
must create a platform where they can collect,
which are big data management, IoT sensors,
manage, and analyze a massive volume of sen-
and data analytics. The analytics are performed
sor data in a scalable and cost-effective manner
in real-time in order to be used in different parts
[2]. In this context, leveraging a big data plat-
of the smart building to manage the oxygen
2. http://www.businessinsider.com/how-the-internet-of- level, smoke/hazardous gases, and luminosity.
things-market-will-grow-2014-10 The framework is implemented in Cloudera
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
3

· Capture
· Integrate Explore
· Store
· Preprocess
Analytical Clean
Data Store

· Map Analyze
· Transform
· Cleanse

T
IoT Infrastructure Share

IP
Big Data Platform Big Data Analytics

CR
Fig. 3: Big Data Flow in IoT

Hadoop distribution where the analytics is per-

formed using PySpark. The results show that
the framework can be utilized for IoT-enabled
big data analytics. The proposed framework is
US are then sent to analytics tools to analyze traffic
density and to provide solutions via predictive
analytics. Compared with the existing systems,
the proposed system provides a better alterna-
AN
specifically designed for smart buildings that tive method for managing traffic.
should be extended to make it generalize so Q. Zhang et al. [6] propose Firework, a new
that it can deal with other IoT applications computing paradigm that allows distributed
including smart cities and smart airplanes.
M

data processing and sharing in an IoT-based,

C. Lee et al. [4] propose an IoT-based cy- collaborative edge environment. Firework com-
ber physical system that supports information bines physically distributed data by providing
ED

analysis and knowledge acquisition methods virtual data views to end users using prede-
to improve productivity in various industries. fined interfaces. These interfaces come in the
This system, which focuses on industrial big form of a set of functions and a set of datasets.
data analytics, integrates various data analyt- Firework aims to minimize data access latency
PT

ics components in the form of reconfigurable by moving the processing closer to the data pro-
and interchangeable modules to meet different ducers in the edge network. Firework instance
business needs. The authors also provide a new has multiple stakeholders who must register
CE

context intelligence framework that can help their datasets and corresponding functions that
handle industrial informatics based on the sen- are abstracted as data views. These data views
sors, locations, and unstructured data for big are available to all participants of the same
AC

data mining. A case study is also performed framework instance such that they can merge
to illustrate the design of the proposed cyber multiple data views into a single job to perform
physical system. detailed data analytics. They illustrate such
P. Rizwan et al. [5] study the strengths and concept by performing case studies of connected
weaknesses of various traffic management sys- health and find the lost.
tems. They propose a low cost, real-time traffic M. M. Rathore et al. [7] propose a smart
management system that deploys IoT devices city management system based on IoT that
and sensors to capture real-time traffic infor- exploits big data and analytics. The data are
mation. Specifically, low-cost traffic detection collected by deploying different sensors, in-
sensors are embedded in the middle of the road cluding weather and water sensors, vehicular
for every 500 or 1000 meters. The collected data networking sensors, surveillance objects, smart
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
4

home sensors, and smart parking sensors. An tion framework (RDF) data received from the
architecture for the system and its model for ETL layer. The learning layer extracts several
implementation have also been designed. The features from the data and forms machine-
proposed system is implemented using the learning-based models. The action layer pro-
MapReduce Hadoop ecosystem in a real envi- vides predetermined actions for the output of
ronment. The implementation process involves the learning layer.
several steps, including data generation, data B. Cheng et al. [10] design GeeLytics, an
gathering, data combining, data categorization, edge analytics platform that performs real-time
data preprocessing, and decision making. Spark data processing at the network edges and in
over Hadoop is used for the efficient process- the cloud. This platform addresses the geo-
ing of big data. Smart systems are utilized as distributed and low-latency analytics resulting

T
sources of city data to develop a smart city as an from the large amounts of IoT data. GeeLyt-
implemented system. However, the developed ics is designed to support dynamic stream

IP
smart system is yet to be deployed and its processing topologies by taking into account
accuracy remains untested. the system characteristics of heterogeneous

CR
B. Ahlgren et al. [8] discuss the significance edge/cloud nodes, and the current system
of using IoT to deliver services for improv- workload.
ing the lives of citizens, including transporta- H. Wang et al. [11] discuss the challenges
tion, air quality, and energy efficiency. The au-
thors emphasize that IoT-based systems must
be based on open data and standards, includ-
ing interfaces and protocols, to enable third-
US and opportunities resulting from IoT and big
data for the maritime cluster. They also develop
a new framework for integrating industrial IoT
with big data and analytics technologies. Im-
AN
party innovations by mitigating manufacturer plementing such framework can help increase
lock-ins. Based on this idea, the authors design output and productivity as well as allow whole
and develop a GreenIoT platform in Sweden clusters to continue acting as leaders in the
M

to determine the advantage of open platforms global maritime industry.

and open data for the development of smart Prez and Carrera [12] conduct a comprehen-
cities. However, some guidelines regarding the sive study of the performance characterization
ED

procurement of an open IoT infrastructure, in- of the servIoTicy API. They specifically focus
cluding common data formats and open appli- on the state-of-the-art infrastructure for hosting
cation programming interfaces (APIs), must be IoT workloads in the cloud with an aim to pro-
devised. vide multi-tenant data stream processing capa-
PT

O. B. Sezer et al. [9] propose an augmented bilities, advanced querying mechanisms, multi-
framework that integrates semantic web tech- protocol support, and software solutions by
nologies, big data, and IoT. The key require- combining advanced data-centric technologies.
CE

ments for the proposed framework are ana- Another study [13] partially solves the big data
lyzed, and the conceptual design of the en- storage problem by proposing AllJoyn Lambda,
visioned IoT system is proposed based on a software solution that integrates AllJoyn in
AC

the analysis results. The conceptual framework the Lambda architecture that is used for big
comprises five layers, namely, data acquisi- data storage and analytics.
tion, extract-transform-load (ETL), semantic- A. J. Jara et al. [14] conduct a survey to
rule reasoning, learning, and action. The data highlight the existing solutions and challenges
acquisition layer, which collects data from dif- to big data that are posed by cyber-physical
ferent sources, can be considered as an input systems. Their study focuses on cloud security
layer to the framework. The ETL layer provides and the heterogeneous integration of data from
sensor drivers to transform the data received multiple sources. They highlight the need for
from different types of sensors. The semantic- developing sophisticated data discovery mech-
rule reasoning supports a reasoning engine anisms and for performing real-time stream
to make inferences from the resource descrip- data processing.
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
5

Z. Ding et al. [15] propose a general sta- amine how various technologies, such as data
tistical database cluster mechanism for big analytics and artificial intelligence, can be used
data analysis in the IoT paradigm (IOT- in the smart world to derive situational facts
StatisticDB). They input statistical functions on and to take actions accordingly. They propose a
IOT-StatisticDB via statistical operators inside gaming-based crowdsourcing platform to make
the database management systems (DBMS) ker- use of human intelligence for the successful
nel. The statistical analysis is performed in a completion of certain control tasks. In the fu-
distributed and parallel fashion using multiple ture, proactive monitoring and diagnosis mech-
servers. anisms with a combination of big data mining
C. Vuppalapati et al. [16] examine the role must be devised to ensure safety in the smart
of big data in healthcare and find that body physical world.

T
sensors generate massive amounts of health- R. P. Minch et al. [21] perform an ex-
related data. Two challenges are analyzed in ploratory research about location privacy in the

IP
this context, namely, integrating these mas- era of IoT, big data, and analytics. They identify,
sive data points with electronic health records classify, and describe privacy issues and reveal

CR
(EHR) and presenting these data to doctors in the possible pain points in the context of big
real time. Based on these observations, they data and analytics. They suggest that a reliable
propose a sensor integration framework that framework for securing privacy in a context-
suggests a scalable cloud architecture that can
provide a holistic approach to the EHR sen-
sor system. Apache Kafka and Spark are used
to process large amounts of data in a real-
US aware environment must be developed in the
future.
A. Mukherjee et al. [22] propose an IoT
framework for the effective execution of data
AN
time manner. Although visualizing the health parallel analytic jobs. They aim to identify a
of patients in real time can help detect urgent suitable analytical algorithm that can cope up
situations, this model lacks a security solution. with the requirements of processing and ana-
M

A. Ahmad et al. [17] analyze human behav- lyzing large amounts of data. Their qualitative
ior by using big data and analytics in the social analysis generates promising results because of
IoT paradigm [18]. They propose an architec- the high effectiveness of the parallel analytic
ED

ture that comprises three operational domains. algorithms in an IoT environment. Future stud-
They also analyze an ecosystem that is created ies must address those issues that hinder the
by smart cities and big data. Collaborative fil- implementation of this model in the presence
tering techniques can be used in the future to of fog computing.
PT

accurately analyze human behavior. A. Mukherjee et al. [23] publish a report

D. Arora et al. [19] utilize big data and an- regarding Condor, a grid framework for data-
alytics techniques to classify network-enabled parallel execution in the IoT paradigm. Their
CE

devices. They also analyze the performance of experimental results reveal that Condor has a
four machine learning algorithms, such as k- better scalability and CPU utilization for data-
nearest neighbor (KNN), NaveBayes (NB), sup- parallel jobs compared with a traditional three-
AC

port vector machines (SVM), and random for- tier, server-based architecture.
est. The experimental results show that the NB H. R. Arkian et al. [24] propose MIST, a
algorithm yields the lowest accuracy among all fog-based data analytics scheme with a cost-
classifier models, while the random forest algo- efficient resource provisioning optimization ap-
rithm yields the highest accuracy. Meanwhile, proach that can be used for IoT crowd sens-
the accuracy of KNN and SVM are closely ing applications. This scheme aims to reduce
related to that of the random forest algorithm. the latency of service provisioning in tradi-
I. L. Yen et al. [20] investigate the poten- tional cloud computing frameworks. The exper-
tial of service discovery and composition tech- imental results show that the MIST fog-based
niques in solving real-world problems based scheme outperforms traditional cloud comput-
on the data generated through IoT. They ex- ing as the number of applications that demand
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
6

real-time services increases. Some possible ex- amounts of data that are generated through the
tensions of this work are as follows: (a) adding Internet of Underwater Things (IoUT). They
a selective sensing module to the fog layer, utilize MapReduce to process these data, and
(b) enriching the architecture with privacy- find that MapReduce greatly shortens the query
preserving data analytics capabilities, and (c) execution time compared with SQL. Despite the
considering the mobility of data generators and many advantages of this framework, testing the
data consumers in the resource provisioning applicability of the scalable trust management
part. protocol with IoUT applications and develop-
M. M. Rathore et al. [25] propose a system ing trust-based admission control for IoUT sys-
that deals with several problems in a smart city tems still need to be addressed in the future.
environment, such as enabling objects to react D. Mourtzis et al. [28] reveal that the adop-

T
with respect to context, minimizing the cost of tion of IoT in the manufacturing industry
collecting data generated by smart devices, and can transform traditional systems into modern

IP
obtaining insights into the data if these data are ones. Moreover, such transformation leads to a
collected and processed in real time. The pro- data production process that turns industrial

CR
posed system has a four-tier architecture, where data into industrial big data, which are ren-
the bottom tier is responsible for data gen- dered useless without analytics power. Adopt-
eration and collection, the intermediate tier 1 ing data analytics can empower enterprises
enables communication among sensors, relays,
base stations, and the Internet, the intermediate
tier 2 is responsible for data management and
processing using the Hadoop framework, and
US to derive new data-driven strategies that can
easily manage competitive pressure. They also
demonstrate how the IoT paradigm can be im-
plemented in a simple case of a company with
AN
the top tier is responsible for applying data almost 100 machines.
analysis techniques and generating results. The R. Ramakrishnan et al. [29] analyze the cur-
implementation results show that the proposed rent energy development in India and deter-
M

system is more scalable and efficient in terms mine the benefits that can be obtained through
of throughput and processing time than the cloud computing and analytics. They also ad-
current systems. However, this system lacks an vocate that the usage of analytics can improve
ED

intelligent decision-making technique that can energy conservation, reduce operation costs,
cope with big data in an IoT environment. and empower customers.
F. Alam et al. [26] examine the applicabil-
ity of eight data mining algorithms, including 3 B IG DATA P ROCESSING AND A NA -
PT

SVM, KNN, NB, C4.5, C5.0, linear discrimi- LYTICS P LATFORMS

nant analysis (LDA), artificial neural network
(ANN), and deep learning ANN (DLANN), This section investigates the big data processing
CE

for IoT-generated data. These algorithms are and analytics platforms that can be used for
also compared in terms of their confusion ma- large amounts of IoT-generated data. In IoT,
trix, classification accuracy, and execution time. the big data processing and anlytics can be per-
formed closer to data source using the services
AC

With regard to classification accuracy, C4.5,

C5.0, ANN, and DLANN outperform SVM, of mobile edge computing [30], [31], cloudlets
KNN, NB, and LDA. However, C4.5, C5.0, and [32] and fog computing [33].
ANN are very similar in terms of classifica-
tion accuracy. Meanwhile, NB and LDA have 3.1 Apache Hadoop
the fastest execution time, with LDA having a First used by Yahoo! and Facebook, Hadoop
slightly better processing time than NB. The au- [34] is an open source data processing plat-
thors are planning to conduct a detailed study form that stores and processes large amounts
on larger and diverse IoT datasets in the future. of data on a cluster of commodity hardware.
M. H. Berlian et al. [27] introduce a The Hadoop architecture contains several com-
framework for monitoring and analyzing large ponents, of which the most important are the
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
7

Hadoop Distributed File System (HDFS) and Hive, while SAP uses Sybase IQ to provide a
the MapReduce programming model. HDFS is columnar DBMS. Hana also has a built-in ana-
used to store the data, while MapReduce is lytics library for containing, spatial processing,
used to process these data in a distributed man- and supporting R language and text analytics
ner [35]. Despite its many advantages, Hadoop libraries. Apart from its low latency, SAP-Hana
lacks encryption at the storage and network can also analyze both text and unstructured
levels, has a limited flexibility, is considered data. However, in this tool, all data in a row
unsuitable for small data sets, and has a high must be read even though only the data from
I/O overhead. a few columns are required to be accessed.
Moreover, the capabilities of SAP-Hana are not
3.2 1010data strong enough compared with those of other

T
solutions.
1010data [36] consists of a columnar database

IP
and mostly deals with semi-structured data,
such as IoT data. Aside from its data visual- 3.5 HP-HAVEn

CR
ization, reporting, and integration capabilities,
this tool provides advanced analytic services, HP introduced the Hadoop Autonomy Vertica
including optimization and statistical analysis. Enterprise (HAVEn) [39] security, a new big IoT
1010data is also very supportive for large-scale data platform architecture for a large number of
infrastructure. This tool also works in a cen-
tralized fashion and applies access controls to
interact with back-end systems. 1010data can
US HP systems that can be used with any number
of applications. HP provides reference hard-
ware configurations for the major distributors
AN
satisfy customer demand through its advanced of the Hadoop software. Autonomy’s IDOL
analytic capabilities. However, 1010data is con- software provides search and exploration ser-
sidered ineffective in terms of data extraction, vices for unstructured data. Vertica is an analyt-
transformation, and loading. ical DBMS for a massively parallel processing
M

columnar database that aims to accelerate the

analysis of big structured datasets. HP HAVEn
3.3 Cloudera Data Hub
is currently collaborating with several com-
ED

Cloudera introduced the Enterprise Data Hub panies to complement legacy enterprise data
[37], a Hadoop-based framework for big IoT warehouses. HP also introduced a “Flex-Zone”
data processing and analytics that can be uti- to facilitate the exploration of large datasets
lized as a central point in managing mas-
PT

before defining the database scheme. The only

sive amounts of IoT data from enterprises. To drawback of HP-HAVEn is that an increment
achieve reliability, high performance, security, in the number of tenants generates a large
and data-access control, the Cloudera Data Hub
CE

database catalog where the lock holding and

combines Cloudera Manager, Navigator, and release processes in all operations are deceler-
its backup and recovery components. However, ated.
this tool does not have its own hardware and
AC

software systems and merely relies on third

parties when identifying serious privacy and 3.6 Hortonworks
security concerns.
Hortonworks [40] focuses on building a big IoT
data analytics and management platform based
3.4 SAP-Hana on Hadoop. The Hortonworks Data Platform
SAP-Hana [38] is an in-memory platform for (HDP) has a free open source software distribu-
performing big IoT data analytics and address- tion and focuses on the improvement of Hive.
ing transactional needs. SAP supports various However, with its HDP plugin, Hortonworks
distributed solutions to accommodate big un- cannot reduce the number of node-groups or
structured data. Hana accesses big data through hosts per node group in the generated cluster.
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
8

3.7 Pivotal big data suite Recently, MapR added LucidWorks Search and
The Pivotal big data suite (Pivotal BDS) [41], stream processing options into Hadoop to en-
which is usually deployed in a public cloud, hance its predictive capabilities and enable fast
comprises three solutions, namely, Pivotal processing. However, MapR has a higher com-
HDB, Pivotal Greenplum, and Pivotal GemFire, plexity compared with Hadoop.
all of which are delivered under a single license.
Pivotal is an analytical database that combines
massively parallel processing (MPP)-based ana- 4 R EQUIREMENTS
lytics performance with robust ANSI SQL com-
pliance and helps in performing predictive an- The requirements of big data and analytics
alytics on data that are stored in HDFS using in IoT have exponentially increased over the

T
SQL syntax and other related tools. Pivotal years and promise dramatic improvements in

IP
Greenplum is an open source MPP analytical decision-making processes. As a result, the de-
database that is used for performing rapid an- mands of adapting data analytics to big data
alytics on voluminous amounts of data and in IoT have increased as well, thereby changing

CR
provides high query performance on petabyte- the way that data are collected, stored, and ana-
scale data volume. Pivotal GemFire is an in- lyzed. Big data and analytics have great poten-
memory data grid that is designed to support tial for extracting meaningful information from
high volumes of operational and transactional
applications. Despite its many benefits, Pivotal
BDS is still in its infancy and its wide adoption
US the data produced by sensor devices. The gen-
eral requirements for big data and IoT define
the functional and nonfunctional specifications
AN
is mired by many unresolved issues. for data analytics. This section presents the
key requirements for big data and analytics in
the IoT environment. These requirements play
3.8 Infobright
an important role in improving IoT services
M

A tool specifically designed for solving data through analytics.

management and analytic problems [42], In-
fobright can analyze up to 50 terabytes of
ED

data. With its high compression and data skip- 4.1 Connectivity
ping ratio, Infobright is considered suitable for
machine-generated data, such as IoT data. In- The IoT paradigm is gradually leading to the
fobright mostly works with Hadoop or high- ubiquitous connectivity of intelligent sensor-
PT

scale data warehouses. The data skipping tech-equipped objects in a smart environment. One
nology and columnar design of this tool ensureof the key requirements of IoT is to provide a
that only the concerned data will be used in reliable connectivity for big data and analytics
CE

each query. These data are also indexed auto- to facilitate the combination and integration
matically without the need of any partitioningof huge volumes of machine-generated sen-
and tuning. However, all queries cannot be sor data. Thus, numerous objects around us
AC

answered optimally using the Infobright opti- have a great potential to be connected to high-
mizer. performance computing infrastructures to en-
hance IoT services. Moreover, with the growing
presence of WiFi and 4G-LTE wireless Internet
3.9 MapR access, the evolution toward ubiquitous infor-
MapR [43] supports big data and analytics as mation and communication networks is already
well as adopts several components of Hadoop evident [44]. However, a seamless connection
to improve its performance (e.g., replacing among different objects in smart cities [45], such
HDFS with an NFS-like network file system to as IoT, cloud computing, big data, and ana-
achieve security and high availability). MapR lytics, must be established before embedding
also has its own system recovery approach. intelligence into our environment.
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
9

4.2 Storage things”. Big data and analytics in IoT require

The continuous rapid growth of a large number streaming events on the fly and storing stream-
of IoT-enabled objects has resulted in the stor- ing data in an operational database. Given that
age of massive amounts of heterogeneous data much of these unstructured data are streamed
in low-cost commodity hardware on a real-time directly from web-enabled “things”, big data
basis. The key requirements of big data storage implementations must perform analytics with
in IoT include handling very large amounts of real-time queries to help organizations obtain
unstructured data and providing low latency insights quickly, rapidly make decisions, and
for analytics. Moreover, the applications of big interact with people and other devices in real
data technologies for IoT can enable efficient time [15].
data storage and processing in order to produce

T
information that can enhance different smart 4.5 Benchmark

IP
city services. The spectrum of IoT data sources Big data and analytics have attracted much
includes sensor data, smartphones, and social attention from the academia and various orga-
media that are modeled in different ways and

CR
nizations, and many organizations have started
use various communication protocols and inter- pursuing IoT businesses as well. However,
faces. Most IoT services are based on M2M com- these organizations face some challenges in
munication protocols, which require handling storing and analyzing vast amounts of data
a large number of streams and directly benefit
from the widely distributed storage capacities
of cloud computing infrastructure [46].
US
that are collected through sensors in an IoT
environment. Solving these problems requires
a deep understanding that can be achieved by
AN
using a big data and analytics platform. Bench-
4.3 Quality of services mark plays an important role in this context by
The resource management of IoT sensors and providing organizations with a way to judge
mobile devices is the primary requirement for the quality of big data and analytics solutions.
M

quality of service (QoS) to effectively analyze a An excellent system benchmark can also pro-
huge amount of data. Although many studies vide simple and straightforward comparisons
have attempted to meet the QoS requirement, of various solutions.
ED

how to unify and integrate the QoS architecture

into IoT to support big data and analytics war-
5 TAXONOMY
rants further research [47]. The QoS provided
PT

by an IoT network must be reliable and must Figure 4 shows the thematic taxonomy of big
guarantee a mobile and efficient transfer of data and analytics solutions that are designed
data from those sources where big data are for IoT systems. These solutions are categorized
generated. The QoS support in this network based on the following attributes: a) big data
CE

is extremely important to big data and ana- sources, b) system components, c) big data en-
lytics. However, to create a reliable network, abling technologies, d) functional elements, and
many emerging networking technologies must e) analytics type.
AC

be introduced into IoT to enable real-time event

transfer and improve big data processing capa- 5.1 Big data sources
bilities. Big data are generated by an infrastructure that
is deployed to run various IoT applications,
4.4 Real-time analytics including city management, manufacturing, in-
Streaming analytics has rapidly emerged as a telligent transport systems (ITS), smart build-
key IoT initiative for timely decision-making ing, and monitoring sensors.
processes [48]. One of the most prominent fea- The city management uses connected cam-
tures of IoT is its real-time or near-time commu- eras, sensors, and actuators to make the lives of
nication of information regarding “connected citizens secure and convenient. However, these
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
10

T
IP
CR
US
AN

Fig. 4: Taxonomy of Big Data and Analytics Solutions for IoT Systems
M

devices generate a bulk of data that must be ing has the same goal as that in the other appli-
managed and analyzed in real time to obtain cation domains. The relevant information is ex-
ED

relevant insights. Similarly, the manufacturing tracted from a wide range of existing data and
industry has deployed IoT devices that con- then provided to decision makers for service
tinuously generate a huge amount of data to management and to the users of the building.
PT

maximize the productivity and efficiency of Big data in the IoT environment are com-
its operations. To obtain insights from these monly used for the collection and storage of
data, big data and analytics solutions have been monitoring sensor data, performing data ana-
CE

used in designing and testing new products, lytics, making forecasts, and generating alerts
optimizing services and marketing, minimizing if abnormal deviations are detected.
defects, and improving yields.
5.2 System Components
AC

Along with big data and analytics, the pro-

liferation of sensors [49], connected vehicle Big data and analytics solutions usually com-
technologies [50], [51], and IoT [52] have re- prise five system components, namely, data
sulted in the creation of intelligent transporta- acquisition, data retention, data transport, data
tion systems, thereby significantly increasing processing, and data leverage.
the amount of real-time big data that must Big data acquisition involves collecting, fil-
be communicated, aggregated, analyzed, and tering, and cleaning the data before they are
managed. The ITS can take advantage of big transferred into the data warehouse. This com-
data and analytics to enhance the decision- ponent is commonly governed by four at-
making capabilities of its users. tributes, namely, volume, variety, velocity, and
The use of big data solutions in smart build- value. Big data retention deals with the extant
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
11

policies and requires the management to meet rules or policy engines, edge computing de-
big data archival requirements. Various big data vices, and data output.
retention policies involve privacy and legal The raw data are collected from different
concerns against economics to identify archival resources and transferred to edge analytics sys-
rules, retention time, data formats, and encryp- tems. These systems are based on a rules/policy
tion methods. The big data must be transported engine that defines and applies rules to the
across different data sites to guarantee load input data in order to obtain insights. The edge
balancing, business continuity, and replication. computing device is another key player in the
Big data is a term used for large and com- operation of the analytics system. Processing
plex datasets that cannot be processed by tra- the data that are generated by IoT devices on
ditional software. The key challenges involved the edge devices can bring several advantages,

T
in big data processing are related to capturing, such as low latency, minimal bandwidth con-
storage, analysis, search, updating, visualiza-

IP
sumption, data integrity, security, and low cost
tion, and privacy. Big data leverage involves [30], [31], [53]. These data are also made avail-
ensuring how a business can reap benefits from able to the consumer in real time.

CR
their data to increase their revenue.

5.5 Analytics Type

5.3 Big Data Enabling Technologies
The big data enabling technologies in the IoT
context are related to ubiquitous wireless com-
munication, real-time analytics, machine learn-
US
Analytics can be divided into three types,
namely, desrciptive analytics, predictive ana-
lytics, and perspective analytics. Descriptive
AN
ing, and data capturing elements, such as com- analytics, which defines “what has happened
modity sensors and embedded systems. or what is happening,” helps find new busi-
The key ubiquitous wireless communication ness opportunities and challenges. Predictive
technologies that are used for transporting big analytics, which defines “what will happen
M

data in IoT include IEEE 802.15.4, IEEE 802.11, and why it will happen,”is enabled by using
IEEE 802.15.1, and IEEE 802.16. various technologies, such as text/web/data
Real-time analytics make the big data gen- mining, to accurately predict future conditions
ED

erated by IoT devices ready to use as they and states. Prescriptive analytics, which defines
enter the system. Real time can be defined as a “what should I do and why should I do it,”
level of computer responsiveness that is either utilizes simulation, expertise, and decision sup-
port systems to investigate various choices and
PT

instantaneous or nearly instantaneous.

Unlike traditional analytic tools, machine provide suggestions to decision makers.
learning can exploit the hidden insights in
CE

big data and extract values from big data

sources with minimal human interaction. Ma- 6 T HE ROLE OF DATA ANALYTICS IN
chine learning is well suited in the IoT context I OT APPLICATIONS
because of the different data sources and the
AC

Big data technologies can offer data storage

huge amount and variety of data involved.
and processing services in an IoT environment,
The big data in IoT are collected by us-
while data analytics allow business people to
ing several sensors and actuators. These sensor
make better decisions. IoT applications are the
technologies have key roles in collecting and
major sources of big data. This section explains
transmitting data to the nearby edge resources
the role of big data and analytics in different
for further processing.
IoT applications, including smart grids, smart
healthcare, smart transportation, and smart in-
5.4 Key Elements ventory systems [44], [54], [55]. Table 1 summa-
The big data and analytics solutions for IoT rizes the benefits of big data and analytics in
comprise four key elements, namely, input, IoT applications.
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
12

TABLE 1: Benefits of Data Analytics for IoT Applications

IoT Application Benefits of Data Analytics
(a) Reduce the number of accidents by looking into the history of the mishaps
Smart (b) Minimize traffic congestion
Transportation (c) Optimize shipment movements
(d) Ensure road safety
(a) Predict epidemics, cures, and disease
Smart Healthcare (b) Help insurance companies make better policies
(c) Pick up the warning signs of any serious illnesses during their early stages
(a) Help design an optimal pricing plan according to the current power consumption
Smart Grid (b) Predict future supply needs
(c) Ensure an appropriate level of electricity supply
(a) Detect fraudulent cases
Smart Inventory (b) Strategically place an advertisement

T
System (c) Understand customer needs
(d) Identify potential risks

IP
CR
6.1 Smart Transportation 6.3 Smart Grid
Finding valuable information has become a key Smart grids rapidly generate data, and find-
concern in this modern age of technologies ing useful information from these data has
where vehicles are connected to the Internet
and generate large amounts of data. Data an-
alytics can help transport management authori-
ties to find out the history of road mishaps (e.g.,
US become imperative. In a smart grid environ-
ment, large amounts of data are collected from
various sources, such as the power utilization
habits of users, phasor measurement data for
AN
under what circumstances did the accident oc- situational awareness, and energy consumption
cur and at what speed were the drivers driving data measured by widespread smart meters,
during the mishap), minimize the number of to name a few [55]. Proper analytics can help
road accidents, determine the time when the decision makers measure the appropriate level
M

traffic load reaches its peak, and prepare an op- of electricity supply that they must provide to
timal route plan that can help minimize traffic their customers. Analytics may also help busi-
congestion. ness people predict electricity demands in the
ED

The analytics of smart transport data can near future. The strategic objectives of specific
indirectly optimize shipment movements, im- organizations can also be met through proper
prove road safety, and enhance end-to-end user analytics (e.g., pricing plans that are consistent
PT

experience in terms of delivery time. with supply, demand, and production models).

6.2 Smart Healthcare

CE

6.4 Smart Inventory System

Over the past few years, voluminous amounts
of data have been created in the healthcare Finding useful information from large amounts
sector. However, such rapid increase in data of inventory systems data can help business
AC

production has created challenges in extract- owners generate more profit. The analytics of
ing valuable information from big healthcare inventory-systems-generated datasets can help
datasets that can help predict epidemics and one acquire knowledge about market trends.
find cures for various diseases. Data analytics Product recommendations can be generated af-
can help healthcare specialists analyze a large ter analyzing seasonal variations. The analytics
amount of patient data and learn the history of inventory data can also help detect fraud-
of a disease (in the case of family doctors). ulent cases. Analytics may aid advertisers in
Insurance companies may also use data analyt- strategically placing their advertisements. Pre-
ics when making policies. Healthcare specialists dictive analytics can help people make valu-
may also detect serious illnesses at their early able decisions and understand further their cus-
stages and subsequently prevent the loss of life. tomers and products. Data analytics can also
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
13

help companies identify their potential risks only for its domain, but cross-domain data
and opportunities. have emerged as efficient solutions to different
problems [56]. Different types of data, such
7 O PPORTUNITIES as runtime data, device metadata, commercial
The current IoT environment provides the fol- data, retail data, and enterprise data, can now
lowing opportunities for effective big data and be used because of the various enabling tech-
analytics. nologies that complement IoT, including big
data, cloud, semantic web, and data storage
7.1 Decision making technologies.
The proliferation of IoT devices, smart phones,
7.4 Value Added Applications

T
and social media offers decision makers with
an opportunity to extract valuable information Deep learning [57], machine learning [58], and

IP
about their users, to predict future trends, and artificial intelligence [59] are key technologies
detect fraud. Big data can generate significant that provide value added applications using

CR
value by making information transparent and IoT and big data. Before the emergence of IoT
usable to organizations, thereby helping them and cloud computing, massive amounts of data
expose variability and boost their performance. and computation power are unavailable for
Much of the data generated through IoT and certain applications, thereby preventing them
various analytics tools create a large num-
ber of opportunities for organizations. These
tools leverage predictive modeling, classifica-
USfrom using such technologies. Different data
analytics platforms [60], business intelligence
platforms [61], visualization applications [62],
AN
tion, and clustering techniques to offer various and analytics applications [63] have recently
data mining solutions. Mining IoT can also im- emerged and helped industries and organiza-
prove the decision-making habits of individuals tions transform their operations, improve their
using big data. productivity and diagnostics, and increase their
M

agility. Such level of detail was not possible

7.2 Improved Efficiency before the emergence of IoT.
The processing and data storage demands of
ED

advanced analytics applications have limited 8 O PEN R ESEARCH C HALLENGES

their adoption in many domains. However,
such barriers are beginning to fall because of IoT systems have the potential to solve many
PT

IoT. Big data technologies, such as Hadoop and problems, but numerous challenges remain un-
cloud-based mining tools, offer substantial ad- addressed. The solutions to some of these chal-
vantages in terms of cost reduction compared lenges are yet to be provided by big data
and analytics solutions themselves, while oth-
CE

with traditional mining techniques. Moreover,

traditional analytics techniques require data to ers require concentrated efforts from the IoT
be in a certain format, which is difficult to community, hardware and platform vendors,
governments, and policy makers.
AC

achieve when using IoT data. However, using

current big data technologies that build around
low-cost community hardware can help im- 8.1 Exploiting the Temporal Usefulness of
prove analytics capability and reduce process- IoT
ing costs. IoT data have a profound impact on the digi-
tized world. However, these data have a tem-
7.3 Independence from Data Silos poral aspect that can be useful in making real-
The advent of IoT and enabling technologies time decisions, improving quality, and provid-
such as cloud computing has allowed the re- ing an excellent user experience. For example,
moval of data silos in different domains. Typ- a consumer-oriented organization can combine
ically, each type of data is considered useful available consumer data with daily parking lot
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
14

occupation data to offer incentives to their cus- different domains and businesses. Another op-
tomers or manage their inventory in a proactive tion is to use non-ontology vocabularies, such
manner on a daily or seasonal basis. In typical as the Haystack project [67], which focuses on
IoT solutions, the insights from the IoT data are defining metadata tags for annotation in the
often either time consuming or not put into use building automation domain.
immediately. This trend changes into a proac- However, this option lacks integration with
tive one to make correlations, derive insights, other similar vocabularies. One may also en-
and find seasonal, emerging, and diminishing counter several disjoint vocabularies unless
patterns using IoT data [64]. In many critical they evolve into ontologies that can be linked
industrial applications, these correlations, in- and shared across domains. Another option is
sights, and patterns can help increase opera- to use open standards, such as the one from

T
tional efficiency and achieve effective control Hypercat consortium [68] that uses a standard
in real time. Therefore, we must implement

IP
catalogue format to encode metadata as RDF
solutions that can handle data at the device triples and link them together by using URLs.
or gateway level where the IoT data from de- However, such efforts are yet to be extended to

CR
vices, sensors, and processes are initially re- the global level.
ceived. Exploiting semantically annotated data
[65] or using a rules engine to locally process
8.3 Diversity Issues
information are potential avenues to explore in
future research. Applying semantics is particu-
larly useful because of its capability to provide
the required abstractions, whereas annotated
US The IoT paradigm has heterogeneous proto-
cols, standards, and platforms. The industrial
world also faces IT and OT integration issues
AN
data still retain their semantics when pushed that demonstrate much technological fragmen-
to IoT/cloud platforms for analytics. tation. The current protocols have several ini-
tiatives, including CoAP, MQTT, XMPP, DDS,
M

STOMP, HTTP, and AMQP. Although the IoT

8.2 Adding Semantics to IoT Data paradigm does not have a universal proto-
The usefulness of any type of data can be en- col, multiple protocols may co-exist because of
ED

hanced by adding metadata to their contexts the different requirements and their intended
and meanings. This practice is particularly im- uses. Therefore, IoT systems may be unable
portant in IoT by helping users process and to support multiple protocols in an extensi-
utilize heterogeneous IoT data at the device, ble way. Intelligent gateway solutions, such as
PT

gateway, and cloud levels with different scopes that proposed in [69], must provide seamless
and granularities. One option is to base the integration and interoperability between vari-
solutions on their ontology, which is a formal ous protocols. In terms of standards, several
CE

representation of concepts and the relationships organizations, such as ITU-T, IETF, ISO/IEC,
among these concepts. Therefore, ontology can IEEE, ETSI, oneM2M, and 3GPP, have shown
be used to create vocabularies of metadata for some efforts. While we may assume that all
AC

annotating IoT data at the source or near the these standardization activities will provide in-
source. Given that ontologies are very easy teroperability (or some form of it), they may
to share and link, they can provide the right lead to a higher ambiguity because instead of
context and meanings of IoT data in an open having a broad scope, they all provide specific
manner. Ontologies are also useful for inte- and isolated solutions that only cover their own
grating IoT data from multiple domains [66]. domains [70].
Although several efforts have been made to In terms of IoT platforms, several initiatives
create general and domain-specific ontologies, have been launched to generate profit from
more efforts are required in some areas, espe- IoT by providing connectivity, data storage, big
cially in the industrial world, to create specific data analysis, predictions, and machine learn-
ontologies for linking and sharing data across ing. The big industry players have achieved
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
15

much progress in offering diverse IoT plat- approach does not have apply scheme map-
forms with a rich feature set. IBM Watson, ping or query languages and can store any
Microsoft Azure, GE Predix, Cisco Jasper, and data without restrictions. However, Data Lakes
PTC ThingWorx are examples of enterprise- introduces few problems. First, given that any
grade platforms that face a vendor lockdown. data can be inserted, data swaps may occur
Open source IoT platform initiatives, such as in the future [76]. To avoid such problem, we
thingsboard.io, Kaa, and DeviceHive, are few must have oversights for data quality, impose
good examples in this regard. metadata inclusion, and ensure data prove-
nance. Second, using Data Lakes may lead to
a loss of agility, which is especially true for
8.4 Security Challenges large organizations that intend to use a large

T
A major hindrance in the broad integration of pool of data for quick analysis and decision
IoT in industries lies in its security. Several making yet are unable to do so efficiently be-

IP
challenges, such as the recent Dyn attack [71], cause they must go through several steps be-
underscore the importance of having secure IoT fore extracting something meaningful from the

CR
devices, platforms, and applications which oth- data. These organizations must instead make a
erwise can lead to major catastrophes, such as clear distinction between those data that can be
the successful execution of a massive DDOS at- used for decision making in near real time and

US
tack. These attacks can have devastating effects those data that can be used to derive business
on the businesses of many critical industries, strategies. The latter data type is more suitable
threaten national security, and even directly or for storage in Data Lake because these data will
indirectly affect human lives. The IT profession- not be used immediately.
AN
als in these industries have their hands full with
the security issues of BYOD [72], [73] and the
implementation of on-site cloud infrastructures
M

in their organizations. Therefore, IoT security 8.6 Data Provenance

issues only add to their worries. Security is also
not the first topic in the current IoT discussions Data provenance is linked to the authentic-
ED

and is still largely treated as a compulsory ity and integrity of the data as well to their
yet secondary subject. Such disregard can be traceability to determine the owners and mod-
attributed to the lack of organizational policies ifiers of the data at each step [77]. However,
and the ambiguities in government laws [74]. given that big data provides deep insights and
PT

To guarantee a successful implementation of analytics that may lead to some form of au-
IoT, solving these security issues must be given tonomous actuation in the real world, we must
priority in the IoT realm. These issues not only ensure that the data used for making such
CE

require technical solutions but also the appro- actuation are coming from a legitimate source.
priate enforcement of policies and guidelines. Several large-scale initiatives, including smart
The views of all stakeholders in IoT must also cities and smart health, plan to make use of
AC

be considered. big data and analytics, thereby making this

issue even more critical. Although much of the
current studies on IoT have focused on data
8.5 Data Management Issues management, only few have tried to address
IoT data are valuable assets. With the expo- the data provenance issue, such as [78]. Having
nential increase in the number of IoT devices, the ability to trace data ownership in IoT can
systems, and processes, new approaches, such be beneficial for monetization purposes when
as Data Lakes [75], have emerged to handle big different actors share their data [79]. Existing
data. Data Lakes stores structured and unstruc- studies in the IoT domain, such as [80], can be
tured data without any pre-conceived notion of used as basis for devising technical solutions to
how these data will be used afterward. This this issue.
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
16

8.7 Data Governance and Regulation ACKNOWLEDGMENT

One critical aspect of IoT data is related to data This work is supported by the Deanship of
governance and regulating its use by differ- Scientific Research at King Saud University
ent entities. Providing unsupervised or uncon- through Imran’s Research Group No. (RG #
trolled access to data introduces privacy con- 1435-051)
cerns and hampers the participation of private
owners, such as citizens who share their data R EFERENCES
from the sensors installed in their homes or in
[1] I. Yaqoob, I. A. T. Hashem, A. Gani, S. Mokhtar, E. Ahmed,
public places for monitoring purposes [81]. We N. B. Anuar, and A. V. Vasilakos, “Big data: From begin-
must provide IoT device owners with options ning to future,” International Journal of Information Manage-
ment, vol. 36, no. 6, pp. 1231–1247, 2016.

T
and tools to specify their preferences and prior- [2] F. J. Riggins and S. F. Wamba, “Research directions on
itize/limit the use of data from their devices the adoption, usage, and impact of the internet of things

IP
[82]–[84]. Future studies must also focus on through the use of big data analytics,” in Proceedings
of 48th Hawaii International Conference on System Sciences
developing policy frameworks to identify the (HICSS’15). IEEE, 2015, pp. 1531–1540.

CR
stakes and concerns of data owners, data con- [3] M. R. Bashir and A. Q. Gill, “Towards an iot big data an-
sumers, and all the other actors between these alytics framework: Smart buildings systems,” in High Per-
formance Computing and Communications; IEEE 14th Interna-
two. The input from regulatory authorities or tional Conference on Smart City; IEEE 2nd International Con-
governments will be necessary, but care must ference on Data Science and Systems (HPCC/SmartCity/DSS),
be taken to not have centralized control over
the data. Data owners must be given more
power to allow them to make decisions within
US 2016 IEEE 18th International Conference on. IEEE, 2016, pp.
1325–1332.
[4] C. Lee, C. Yeung, and M. Cheng, “Research on iot based
cyber physical system for industrial big data analytics,” in
AN
Industrial Engineering and Engineering Management (IEEM),
the scope of the overall policy framework. The 2015 IEEE International Conference on. IEEE, 2015, pp.
general public must be made aware of their role 1855–1859.
and must be given easy-to-use tools for sharing [5] P. Rizwan, K. Suresh, and M. R. Babu, “Real-time smart
traffic management system for smart cities by using in-
their data with other parties. ternet of things and big data,” in Emerging Technological
M

Trends (ICETT), International Conference on. IEEE, 2016,

pp. 1–7.
[6] Q. Zhang, X. Zhang, Q. Zhang, W. Shi, and H. Zhong,
9 C ONCLUSION “Firework: Big data sharing and processing in collabo-
ED

rative edge environment,” in Hot Topics in Web Systems

IoT is one of the biggest sources of big data, and Technologies (HotWeb), 2016 Fourth IEEE Workshop on.
which are rendered useless without analytics IEEE, 2016, pp. 20–25.
[7] M. M. Rathore, A. Ahmad, and A. Paul, “Iot-based smart
power. IoT interacts with big data when vo-
PT

city development using big data analytical approach,” in

luminous amounts of data are needed to be Automatica (ICA-ACCA), IEEE International Conference on.
IEEE, 2016, pp. 1–8.
processed, transformed, and analyzed in high [8] B. Ahlgren, M. Hidell, and E. C.-H. Ngai, “Internet of
frequency. This work specifically focuses on the things for smart cities: Interoperability and open data,”
CE

big data context. First, we investigate the recent IEEE Internet Computing, vol. 20, no. 6, pp. 52–56, 2016.
[9] O. B. Sezer, E. Dogdu, M. Ozbayoglu, and A. Onal, “An
literature on big data processing and analytics extended iot framework with semantics, big data, and
solutions for IoT. Second, we identify the nu- analytics,” in Big Data (Big Data), 2016 IEEE International
AC

merous requirements for big data and analytics Conference on. IEEE, 2016, pp. 1849–1856.
[10] B. Cheng, A. Papageorgiou, F. Cirillo, and E. Kovacs,
in IoT. Third, we taxonomized the literature. “Geelytics: Geo-distributed edge analytics for large scale
Fourth, we determine the various opportunities iot systems based on dynamic topology,” in Internet of
that are brought about by big data. Fifth, we Things (WF-IoT), 2015 IEEE 2nd World Forum on. IEEE,
2015, pp. 565–570.
highlight the role of data analytics in IoT ap- [11] H. Wang, O. L. Osen, G. Li, W. Li, H.-N. Dai, and W. Zeng,
plications. Sixth, we present the open research “Big data and industrial internet of things for the maritime
industry in northwestern norway,” in TENCON 2015-2015
challenges that must be addressed in the fu- IEEE Region 10 Conference. IEEE, 2015, pp. 1–5.
ture. Seventh, we conclude that the existing big [12] J. L. Pérez and D. Carrera, “Performance characterization
data solutions in the IoT paradigm are still in of the servioticy api: an iot-as-a-service data management
platform,” in Big Data Computing Service and Applications
their infancy and the challenges associated with (BigDataService), 2015 IEEE First International Conference on.
them must be solved in the future. IEEE, 2015, pp. 62–71.
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
17

[13] M. Villari, A. Celesti, M. Fazio, and A. Puliafito, “Alljoyn things (iot),” Procedia Computer Science, vol. 98, pp. 437–
lambda: An architecture for the management of smart 442, 2016.
environments in iot,” in Smart Computing Workshops [27] M. H. Berlian, T. E. R. Sahputra, B. J. W. Ardi, L. W.
(SMARTCOMP Workshops), 2014 International Conference Dzatmika, A. R. A. Besari, R. W. Sudibyo, and S. Sukarid-
on. IEEE, 2014, pp. 9–14. hoto, “Design and implementation of smart environment
[14] A. J. Jara, D. Genoud, and Y. Bocchi, “Big data for cyber monitoring and analytics in real-time system framework
physical systems: an analysis of challenges, solutions and based on internet of underwater things and big data,” in
opportunities,” in Innovative Mobile and Internet Services Electronics Symposium (IES), 2016 International. IEEE, 2016,
in Ubiquitous Computing (IMIS), 2014 Eighth International pp. 403–408.
Conference on. IEEE, 2014, pp. 376–380. [28] D. Mourtzis, E. Vlachou, and N. Milas, “Industrial big
[15] Z. Ding, X. Gao, J. Xu, and H. Wu, “Iot-statisticdb: a data as a result of iot adoption in manufacturing,” Procedia
general statistical database cluster mechanism for big data CIRP, vol. 55, pp. 290–295, 2016.
analysis in the internet of things,” in Green Computing [29] R. Ramakrishnan and L. Gaur, “Smart electricity distri-
and Communications (GreenCom), 2013 IEEE and Internet of bution in residential areas: Internet of things (iot) based

T
Things (iThings/CPSCom), IEEE International Conference on advanced metering infrastructure and cloud analytics,”
and IEEE Cyber, Physical and Social Computing. IEEE, 2013, in Internet of Things and Applications (IOTA), International
Conference on. IEEE, 2016, pp. 46–51.

IP
pp. 535–543.
[16] C. Vuppalapati, A. Ilapakurti, and S. Kedari, “The role [30] E. Ahmed and M. H. Rehmani, “Mobile edge computing:
of big data in creating sense ehr, an integrated approach Opportunities, solutions, and challenges,” pp. 59–63.
[31] A. Ahmed and E. Ahmed, “A survey on mobile edge

CR
to create next generation mobile sensor and wearable
data driven electronic health record (ehr),” in Big Data computing,” in Intelligent Systems and Control (ISCO), 2016
Computing Service and Applications (BigDataService), 2016 10th International Conference on. IEEE, 2016, pp. 1–8.
IEEE Second International Conference on. IEEE, 2016, pp. [32] U. Shaukat, E. Ahmed, Z. Anwar, and F. Xia, “Cloudlet
293–296. deployment in local wireless networks: Motivation, ar-
[17] A. Ahmad, M. M. Rathore, A. Paul, and S. Rho, “Defin-
ing human behaviors using big data analytics in social
internet of things,” in Advanced Information Networking andUS
Applications (AINA), 2016 IEEE 30th International Conference
on. IEEE, 2016, pp. 1101–1107.
[33]
chitectures, applications, and open challenges,” Journal of
Network and Computer Applications, vol. 62, pp. 18–40, 2016.
F. Bonomi, R. Milito, J. Zhu, and S. Addepalli, “Fog com-
puting and its role in the internet of things,” in Proceedings
of the first edition of the MCC workshop on Mobile cloud
AN
[18] E. Ahmed and M. H. Rehmani, “Introduction to the spe- computing. ACM, 2012, pp. 13–16.
cial section on social collaborative internet of things,” p. [34] J. Nandimath, E. Banerjee, A. Patil, P. Kakade, S. Vaidya,
382384, 2017. and D. Chaturvedi, “Big data analysis using apache
hadoop,” in Information Reuse and Integration (IRI), 2013
[19] D. Arora, K. F. Li, and A. Loffler, “Big data analytics for
IEEE 14th International Conference on. IEEE, 2013, pp. 700–
classification of network enabled devices,” in Advanced In-
M

703.
formation Networking and Applications Workshops (WAINA),
2016 30th International Conference on. IEEE, 2016, pp. 708– [35] I. A. T. Hashem, N. B. Anuar, A. Gani, I. Yaqoob, F. Xia,
713. and S. U. Khan, “Mapreduce: Review and open chal-
lenges,” Scientometrics, pp. 1–34, 2016.
[20] I.-L. Yen, G. Zhou, W. Zhu, F. Bastani, and S.-Y. Hwang,
ED

[36] V. Morabito, “Managing change for big data driven inno-

“A smart physical world based on service technologies,
vation,” in Big Data and Analytics. Springer, 2015, pp.
big data, and game-based crowd sourcing,” in Web Services
125–153.
(ICWS), 2015 IEEE International Conference on. IEEE, 2015,
[37] A. Bhardwaj, S. Bhattacherjee, A. Chavan, A. Deshpande,
pp. 765–772.
A. J. Elmore, S. Madden, and A. G. Parameswaran,
PT

[21] R. P. Minch, “Location privacy in the era of the internet of “Datahub: Collaborative data science & dataset version
things and big data analytics,” in System Sciences (HICSS), management at scale,” arXiv preprint arXiv:1409.0798,
2015 48th Hawaii International Conference on. IEEE, 2015, 2014.
pp. 1521–1530.
[38] F. Färber, S. K. Cha, J. Primsch, C. Bornhövd, S. Sigg,
CE

[22] A. Mukherjee, H. S. Paul, S. Dey, and A. Banerjee, “Angels and W. Lehner, “Sap hana database: data management
for distributed analytics in iot,” in Internet of Things (WF- for modern business applications,” ACM Sigmod Record,
IoT), 2014 IEEE World Forum On. IEEE, 2014, pp. 565–570. vol. 40, no. 4, pp. 45–51, 2012.
[23] A. Mukherjee, S. Dey, H. S. Paul, and B. Das, “Utilis- [39] S. Burke, “Hp haven big data platform is gain-
ing condor for data parallel analytics in an iot contex-
AC

ing partner momentum,” CRN [online] http://www. crn.

tan experience report,” in Wireless and Mobile Computing, com/news/applications-os/240161649, 2013.
Networking and Communications (WiMob), 2013 IEEE 9th [40] (2017, Accessed on 3rd June) Hortonworks. [Online].
International Conference on. IEEE, 2013, pp. 325–331. Available: https://hortonworks.com/
[24] H. R. Arkian, A. Diyanat, and A. Pourkhalili, “Mist: Fog- [41] Y. Zhuang, Y. Wang, J. Shao, L. Chen, W. Lu, J. Sun, B. Wei,
based data analytics scheme with cost-efficient resource and J. Wu, “D-ocean: an unstructured data management
provisioning for iot crowdsensing applications,” Journal of system for data ocean environment,” Frontiers of Computer
Network and Computer Applications, vol. 82, pp. 152–165, Science, vol. 10, no. 2, pp. 353–369, 2016. [Online].
2017. Available: http://dx.doi.org/10.1007/s11704-015-5045-6
[25] M. M. Rathore, A. Ahmad, A. Paul, and S. Rho, “Urban [42] D. Slezak, P. Synak, J. Wróblewski, and G. Toppin, “In-
planning and building smart cities based on the internet fobright analytic database engine using rough sets and
of things using big data analytics,” Computer Networks, vol. granular computing,” in Granular Computing (GrC), 2010
101, pp. 63–80, 2016. IEEE International Conference on. IEEE, 2010, pp. 432–437.
[26] F. Alam, R. Mehmood, I. Katib, and A. Albeshri, “Analysis [43] (2017, Accessed on 3rd June) Mapr. [Online]. Available:
of eight data mining algorithms for smarter internet of https://mapr.com/
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
18

[44] E. Al Nuaimi, H. Al Neyadi, N. Mohamed, and J. Al- [60] V. O. Safonov, “Example of a trustworthy cloud comput-
Jaroodi, “Applications of big data to smart cities,” Journal ing platform in detail: Microsoft azure,” Trustworthy Cloud
of Internet Services and Applications, vol. 6, no. 1, p. 1, 2015. Computing, pp. 147–270, 2016.
[45] E. Ahmed, M. Imran, M. Guizani, A. Rayes, J. Lloret, [61] J. Vidal-Garcı́a, M. Vidal, and R. H. Barros, “Computa-
G. Han, and W. Guibene, “Enabling mobile and wireless tional business intelligence, big data, and their role in
technologies for smart cities: Part 2,” IEEE Communications business decisions in the age of the internet of things,” in
Magazine, vol. 55, no. 3, pp. 12–13, 2017. The Internet of Things in the Modern Business Environment.
[46] G. Suciu, V. Suciu, A. Martian, R. Craciunescu, A. Vulpe, IGI Global, 2017, pp. 249–268.
I. Marcu, S. Halunga, and O. Fratu, “Big data, internet of [62] Y. Jeong, H. Joo, G. Hong, D. Shin, and S. Lee, “Aviot:
things and cloud convergence–an architecture for secure Web-based interactive authoring and visualization of in-
e-health applications,” Journal of medical systems, vol. 39, door internet of things,” IEEE Transactions on Consumer
no. 11, pp. 1–8, 2015. Electronics, vol. 61, no. 3, pp. 295–301, 2015.
[47] J. Jin, J. Gubbi, T. Luo, and M. Palaniswami, “Network [63] M. Strohbach, H. Ziekow, V. Gazis, and N. Akiva, “To-
architecture and qos issues in the internet of things for a wards a big data analytics framework for iot and smart

T
smart city,” in Communications and Information Technologies city applications,” in Modeling and processing for next-
(ISCIT), 2012 International Symposium on. IEEE, 2012, pp. generation big-data technologies. Springer, 2015, pp. 257–
956–961.

IP
282.
[48] R. Tönjes, P. Barnaghi, M. Ali, A. Mileo, M. Hauswirth, [64] S. Aljawarneh, V. Radhakrishna, P. V. Kumar, and
F. Ganz, S. Ganea, B. Kjærgaard, D. Kuemper, S. Nechifor V. Janaki, “A similarity measure for temporal pattern dis-
et al., “Real time iot stream processing and large-scale

CR
covery in time series data generated by iot,” in Engineering
data analytics for smart city applications,” in poster session, & MIS (ICEMIS), International Conference on. IEEE, 2016,
European Conference on Networks and Communications, 2014. pp. 1–4.
[49] E. Ahmed, S. Ali, A. Akheenzada, and I. Yaqoob, “Cogni- [65] C. El Kaed, I. Khan, H. Hossayni, and P. Nappey, “Sqeniot:
tive radio sensor networks: Bridging the gap for network,” Semantic query engine for industrial internet-of-things
Cognitive Radio Sensor Networks: Applications, Architectures,
and Challenges: Applications, Architectures, and Challenges.
IGI Global, p. 160, 2014.
[50] S. A. A. Shah, E. Ahmed, F. Xia, A. Karim, M. A. Qureshi,
I. Ali, and R. M. Noor, “Coverage differentiation based
US gateways,” Submitted IEEE GLOBECOM, 2016.
[66] T. Banerjee and A. Sheth, “Iot quality control for data and
application needs,” IEEE Intelligent Systems, vol. 32, no. 2,
pp. 68–73, 2017.
AN
[67] A. A. Bhattacharya, D. Hong, D. Culler, J. Ortiz, K. White-
adaptive tx-power for congestion and awareness control
house, and E. Wu, “Automated metadata construction to
in vanets,” Mobile Networks and Applications, pp. 1–12.
support portable building applications,” in Proceedings of
[51] I. Yaqoob, I. Ahmad, E. Ahmed, A. Gani, M. Imran, and the 2nd ACM International Conference on Embedded Systems
N. Guizani, “Overcoming the key challenges to establish- for Energy-Efficient Built Environments. ACM, 2015, pp.
ing vehicular communication: Is sdn the answer?” IEEE 3–12.
M

Communications Magazine, 2017.

[68] T. Jaffey, J. Davies, and P. Beart, “Hypercat 3.00 specifica-
[52] I. Yaqoob, E. Ahmed, I. A. T. Hashem, A. Ahmed, A. Gani,
tion,” Hyper-cat Limited, 2016.
M. Imran, and M. Guizani, “Internet of things architec-
ture: Recent advances, taxonomy, requirements, and open [69] P. Desai, A. Sheth, and P. Anantharam, “Semantic gateway
ED

challenges,” IEEE Wireless Communications, 2017. as a service architecture for iot interoperability,” in Mobile
Services (MS), 2015 IEEE International Conference on. IEEE,
[53] Y. Jararweh, A. Doulat, O. AlQudah, E. Ahmed, M. Al-
2015, pp. 313–319.
Ayyoub, and E. Benkhelifa, “The future of mobile cloud
computing: integrating cloudlets and mobile edge com- [70] A. Meddeb, “Internet of things standards: who stands out
puting,” in Telecommunications (ICT), 2016 23rd Interna- from the crowd?” IEEE Communications Magazine, vol. 54,
PT

tional Conference on. IEEE, 2016, pp. 1–5. no. 7, pp. 40–47, 2016.
[54] N. Bessis and C. Dobre, Big data and internet of things: a [71] C. Chris Mishler and C. CIA, “The future of the internet
roadmap for smart environments. Springer, 2014. of things,” Strategic Finance, vol. 97, no. 5, p. 62, 2015.
[55] I. A. T. Hashem, V. Chang, N. B. Anuar, K. Adewole, [72] S. Tanimoto, S. Yamada, M. Iwashita, T. Kobayashi,
CE

I. Yaqoob, A. Gani, E. Ahmed, and H. Chiroma, “The role H. Sato, and A. Kanai, “Risk assessment of byod: Bring
of big data in smart city,” International Journal of Information your own device,” in Consumer Electronics, 2016 IEEE 5th
Management, vol. 36, no. 5, pp. 748–758, 2016. Global Conference on. IEEE, 2016, pp. 1–4.
[56] A. Brring, S. Schmid, C. K. Schindhelm, A. Khelil, [73] K. Hajdarevic, P. Allen, and M. Spremic, “Proactive se-
curity metrics for bring your own device (byod) in iso
AC

S. Kbisch, D. Kramer, D. L. Phuoc, J. Mitic, D. Anicic, and

E. Teniente, “Enabling iot ecosystems through platform 27001 supported environments,” in Telecommunications Fo-
interoperability,” IEEE Software, vol. 34, no. 1, pp. 54–61, rum (TELFOR), 2016 24th. IEEE, 2016, pp. 1–4.
Jan 2017. [74] V. A. Almeida, D. Doneda, and J. de Souza Abreu, “Cyber-
[57] X. W. Chen and X. Lin, “Big data deep learning: Chal- warfare and digital governance,” IEEE Internet Computing,
lenges and perspectives,” IEEE Access, vol. 2, pp. 514–525, vol. 21, no. 2, pp. 68–71, 2017.
2014. [75] H. Fang, “Managing data lakes in big data era: What’s
[58] J. Qiu, Q. Wu, G. Ding, Y. Xu, and S. Feng, “A survey a data lake and why has it became popular in data
of machine learning for big data processing,” EURASIP management ecosystem,” in Cyber Technology in Automa-
Journal on Advances in Signal Processing, vol. 2016, no. 1, tion, Control, and Intelligent Systems (CYBER), 2015 IEEE
pp. 1–16, 2016. International Conference on. IEEE, 2015, pp. 820–824.
[59] O. Etzion, “When artificial intelligence meets the internet [76] R. Hai, S. Geisler, and C. Quix, “Constance: An intelligent
of things,” in Proceedings of the 9th ACM International data lake system,” in Proceedings of the 2016 International
Conference on Distributed Event-Based Systems. ACM, 2015, Conference on Management of Data. ACM, 2016, pp. 2097–
pp. 246–246. 2100.
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
19

[77] W. She, I.-L. Yen, F. Bastani, B. Tran, and B. Thurais- is associate editor of IEEE Communication Mag-
ingham, “Role-based integrated access control and data azine, IEEE Access, and Wiley Wireless Commu-
provenance for soa based net-centric systems,” in Service
Oriented System Engineering (SOSE), 2011 IEEE 6th Interna- nications and Mobile Computing, Elsevier Journal
tional Symposium on. IEEE, 2011, pp. 225–234. of Network and Computer Applications, and KSII
[78] B. Glavic, “Big data provenance: Challenges and implica-
tions for benchmarking,” in Specifying big data benchmarks.
TIIS. He has also served as a Lead Guest Edi-
Springer, 2014, pp. 72–80. tor/Guest Editor and Chair/Co-chair in inter-
[79] Y. Zhang and J. Wen, “An iot electric business model based national journals and international conferences,
on the protocol of bitcoin,” in Intelligence in Next Genera-
tion Networks (ICIN), 2015 18th International Conference on. respectively. His areas of interest include Mo-
IEEE, 2015, pp. 184–191. bile Cloud Computing, Mobile Edge Comput-
[80] Q. H. Cao, I. Khan, R. Farahbakhsh, G. Madhusudan, ing, Internet of Things, Cognitive Radio Net-
G. M. Lee, and N. Crespi, “A trust model for data shar-
ing in smart cities,” in Communications (ICC), 2016 IEEE works, and Smart Cities. He has successfully

T
International Conference on. IEEE, 2016, pp. 1–7. published his research work in more than fifty
[81] I. Khan, F. Belqasmi, R. Glitho, N. Crespi, M. Morrow, and
international journals and conferences.

IP
P. Polakos, “Wireless sensor network virtualization: Early
architecture and research perspectives,” IEEE Network,
vol. 29, no. 3, pp. 104–112, 2015.

CR
[82] E. D. Ragan, A. Endert, J. Sanyal, and J. Chen, “Charac-
terizing provenance in visualization and data analysis:
an organizational framework of provenance types and
purposes,” IEEE transactions on visualization and computer
graphics, vol. 22, no. 1, pp. 31–40, 2016.

US
Ibrar Yaqoob received his Ph.D. degree
[83] C. H. Suen, R. K. Ko, Y. S. Tan, P. Jagadpramana, and B. S.
Lee, “S2logger: End-to-end data tracking mechanism for
in Computer Science from the University of
cloud data provenance,” in Trust, Security and Privacy in
Malaya, Malaysia, in 2017. He earned 550 plus
Computing and Communications (TrustCom), 2013 12th IEEE
AN
citations, and 50 plus impact factor during his
International Conference on. IEEE, 2013, pp. 594–602.
[84] M. B. Jones, B. Ludäscher, T. McPhillips, P. Missier,
Ph.D. candidature. He worked as a researcher
C. Schwalm, P. Slaughter, D. Vieglais, L. Walker, and
at Centre for Mobile Cloud Computing Re-
Y. Wei, “Dataone: A data federation with provenance sup-
search (C4MCCR), University of Malaya. His
port,” in Provenance and Annotation of Data and Processes:
M

6th International Provenance and Annotation Workshop, IPAW

research experience spans over more than three
2016, McLean, VA, USA, June 7-8, 2016, Proceedings, vol.
9672. Springer, 2016, p. 230. and half years. He has published a number of
research articles in refereed international jour-
ED

nals and magazines. His numerous research

articles are very famous and among the most
downloaded in top journals. His research inter-
ests include big data, mobile cloud, the Internet
PT

of Things, cloud computing, and wireless net-

works.
Ibrahim Abaker Targio Hashem received
CE

his Ph.D. degree in Computer Science from

the University of Malaya, Malaysia, in 2017.
He received his M.S. degree in computing in
AC

2012, Malaysia, and the B.E. degree in com-

Ejaz Ahmed worked at Centre for Mobile
puter science in 2007, Sudan. Hashem obtained
Cloud Computing Research (C4MCCR), Uni-
professional certificates from CISCO (CCNP,
versity of Malaya, Malaysia. Before that, he has
CCNA, and CCNA Security) and APMG Group
worked as Research Associate in CogNet (Cog-
(PRINCE2 Foundation, ITIL v3 Foundation,
nitive Radio Network) Research Lab SEECS,
and OBASHI Foundation). He worked as a Tu-
NUST Pakistan from December 2009 to Septem-
tor at CISCO Academy, University of Malaya.
ber 2012, and in CoReNet (Center of Research in
His main research interests include big data,
Networks and Telecom), CUST, Pakistan, from
cloud computing, distributed computing, and
January 2008 to December 2009. His research
network.
experience spans over more than ten years. He
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
20

Abdelmuttlib Ibrahim Abdalla Ahmed

(abdelmuttlib@siswa.um.edu.my) received his
B.Sc. degree in computer science from OIU,
Sudan, and his M.S. degree in computer science
Imran Khan is working as innovation from IIUI, Pakistan. He is currently pursuing a
project leader in Schneider Electric. He is lead- Ph.D. degree at the University of Malaya. His
ing the design and specification of data and research Interest areas include trust and reputa-
information management architectures for sus- tion systems, security and digital forensics, In-
tainable energy management in various indus- ternet of Things, mobile and cloud computing,
trial domains. He received Ph.D. degree in and vehicular networks.
Computing and Networks from Institut Mines-

T
Tlcom, Tlcom SudParis jointly with UPMC
Paris VI, France, M.S. degree in Multimedia

IP
and Communication from M.A. Jinnah Univer-
sity, Pakistan and B.S. degree in Computer Sci-

CR
ence from COMSATS Institute of IT, Pakistan.
During his Ph.D. he worked as collaborating
researcher at Concordia University, Montreal,
Canada to lead a 3 year Cisco funded project.
He was also involved in several European re-
US Muhammad Imran (cimran@ksu.edu.sa) is
search projects funded by ITEA2 and H2020. an assistant professor in the College of Com-
During M.S. Imran was member of Center of puter and Information Science, King Saud Uni-
AN
Research in Networks and Telecom (CoReNeT) versity. His research interests include mobile ad
and worked on projects funded by the French hoc and sensor networks, WBANs, IoT, M2M,
Ministry of Foreign Affairs and the Internet multihop wireless networks, and fault-tolerant
Society (ISOC). He has number of publications computing. He has published a number of re-
M

in peer reviewed conferences and journals, and search papers in peer reviewed international
has also contributed to the IETF standardiza- journals and conferences. His research is fi-
ED

tion activities. His current research interests are nancially supported by several grants. He is
Internet of Things (IoT), data and information serving as a Co-Editor-in-Chief for EAI Transac-
management using semantic web technologies, tions on Pervasive Health and Technology. He
cloud and edge computing, software defined also serves as an Associate Editor for the Wire-
PT

automation and wireless sensor networks. less Communication and Mobile Computing
Journal (Wiley), the Inderscience International
Journal of Autonomous and Adaptive Com-
CE

munications Systems, Wireless Sensor Systems

(IET), and the International Journal of Infor-
mation Technology and Electrical Engineering.
AC

He has served/serves as a Guest Editor for

IEEE Communications Magazine, IJAACS, and
the International Journal of Distributed Sensor
Networks. He has been involved in a num-
ber of conferences and workshops in various
capacities such as a Program Co-Chair, Track
Chair/Co-Chair, and Technical Program Com-
mittee member. These include IEEE GLOBE-
COM, ICC, AINA, LCN, IWCMC, IFIP WWIC,
and BWCCA. He has received a number of
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

ACCEPTED MANUSCRIPT
21

awards such as an Asia Pacific Advanced Net- bile/Wireless Networks, IoT, Sensor Networks.
work fellowship. He has authored or coauthored over 250 tech-
nical papers in major international journals and
conferences. Moreover, he is author/co-author
of five books and more than 20 book chapters.
He served or is serving as an Editor or/and
Guest Editor for many technical journals, such
as the IEEE Transactions on Network and Ser-
vice Management, IEEE Transactions on Cloud
Athanasios V. Vasilakos currently Profes- Computing, IEEE Transactions on Cybernet-
sor at Lulea University of Technology, Sweden. ics, IEEE Transactions on Information Forensics

T
He is also General Chair of the European Al- and Security. Moreover, he has served as Gen-
liances for Innovation. His research interests eral Chair, Technical Program Committee Chair

IP
include Cloud Computing, Smart Grid, Energy for many international conferences.
Security and Harvesting, Social Networks, Mo-

CR
US
AN
M
ED
PT
CE
AC