Cloud and IoT-based emerging services systems
Added on 2023-06-09
21 Pages17502 Words55 Views
Cloud and IoT-based emerging services systems
Sugam Sharma 1,2 • Victor Chang 3 • U. Sunday Tim 4 • Johnny Wong5 • Shashi Gadia 5
Received: 14 October 2017 / Revised: 27 March 2018 / Accepted: 17 July 2018
Ó Springer Science+Business Media, LLC, part of Springer Nature 2018
Abstract
The emerging services and analytics advocate the service delivery in a polymorphic view that successfully serves a variety
of audience. The amalgamation of numerous modern technologies such as cloud computing, Internet of Things (IoT) and
Big Data is the potential support behind the emerging services Systems. Today, IoT, also dubbed as ubiquitous sensing is
taking the center stage over the traditional paradigm. The evolution of IoT necessitates the expansion of cloud horizon to
deal with emerging challenges. In this paper, we study the cloud-based emerging services, useful in IoT paradigm, that
support the effective data analytics. Also, we conceive a new classification called CNNC {Clouda, NNClouda} for cloud
data models; further, some important case studies are also discussed to further strengthen the classification. An emerging
service, data analytics in autonomous vehicles, is then described in details. Challenges and recommendations related to
privacy, security and ethical concerns have been discussed.
Keywords Emerging services As-a-Service Analytics Cloud computing IoT Big Data CNNC
1 Introduction
Emerging services and analytics promisingly hide the
complex details of the comprehensive data processing from
the end users and deliver a suave and plain results. The
resulted outcomes are easily interpretable even by the
technically naı ̈ve stakeholders [1, 2], managers, and even a
common user. The success of emerging services and ana-
lytics is largely depending upon the blending of various
modern technologies and cloud computing, and IoT (In-
ternet of Things) are just to name a few and are the core
ingredients. The information age is shifting from traditional
human-intervened Internet to Internet of Things (IoT)
(Fig. 1a), where the sensor embedded commodity, also
called things, communicate among themselves [3] on the
existing network resources and help facilitate the infor-
mation collection and analytics with a very high degree of
automation. The pool of IoT-enabled devices is growing
rapidly and IoT footprint in increasing almost all in
domains that includes trivial to complex applications such
as smart grid [4] to advanced smart cities [5]. The IoT
‘‘Things’’ include a diverse range of embedded objects or
devices such heart monitoring implants, sensor equipped
automobiles, thermostat systems and so on [6]. The
increasing IoT maturity and the advancing cloud-based
services are revolutionizing the information generation,
collection, management and analytics. In IoT realm, the
devices collect useful data and then share the data between
other devices [7]; there are 9 billion such devices that exist
today and the number is inflating. Evans [8] believes this
number will reach nearly 50 billion by year 2020 (Fig. 1b),
which substantiates the fact that the permeation of IoT is
expanding in human life. This indicates that the number of
new applications under IoT umbrella is increasing and
consequently a massive amount of data is being generated
at swift rate, called Big Data [9], where the social media
and sensor embedded systems play central role in data
inundation. As, the IoT-enabled devices access and process
the data from several peer devices to constitute immediate
& Sugam Sharma
sugamsha@iastate.edu; ceo@efeed-hungers.com
1 Center for Survey Statistics and Methodology, Iowa State
University, Ames, Iowa, USA
2 Founder & CEO, eFeed-Hungers.com TM , Ames, USA
3 International Business School Suzhou, Xi’an Jiaotong-
Liverpool University, Xi’an, China
4 Department of Agricultural and Biosystems Engineering,
Iowa State University, Ames, Iowa, USA
5 Department of Computer Science, Iowa State University,
Ames, Iowa, USA
123
Cluster Computing
https://doi.org/10.1007/s10586-018-2821-8 (0123456789().,-volV)(0123456789().,-volV)
Sugam Sharma 1,2 • Victor Chang 3 • U. Sunday Tim 4 • Johnny Wong5 • Shashi Gadia 5
Received: 14 October 2017 / Revised: 27 March 2018 / Accepted: 17 July 2018
Ó Springer Science+Business Media, LLC, part of Springer Nature 2018
Abstract
The emerging services and analytics advocate the service delivery in a polymorphic view that successfully serves a variety
of audience. The amalgamation of numerous modern technologies such as cloud computing, Internet of Things (IoT) and
Big Data is the potential support behind the emerging services Systems. Today, IoT, also dubbed as ubiquitous sensing is
taking the center stage over the traditional paradigm. The evolution of IoT necessitates the expansion of cloud horizon to
deal with emerging challenges. In this paper, we study the cloud-based emerging services, useful in IoT paradigm, that
support the effective data analytics. Also, we conceive a new classification called CNNC {Clouda, NNClouda} for cloud
data models; further, some important case studies are also discussed to further strengthen the classification. An emerging
service, data analytics in autonomous vehicles, is then described in details. Challenges and recommendations related to
privacy, security and ethical concerns have been discussed.
Keywords Emerging services As-a-Service Analytics Cloud computing IoT Big Data CNNC
1 Introduction
Emerging services and analytics promisingly hide the
complex details of the comprehensive data processing from
the end users and deliver a suave and plain results. The
resulted outcomes are easily interpretable even by the
technically naı ̈ve stakeholders [1, 2], managers, and even a
common user. The success of emerging services and ana-
lytics is largely depending upon the blending of various
modern technologies and cloud computing, and IoT (In-
ternet of Things) are just to name a few and are the core
ingredients. The information age is shifting from traditional
human-intervened Internet to Internet of Things (IoT)
(Fig. 1a), where the sensor embedded commodity, also
called things, communicate among themselves [3] on the
existing network resources and help facilitate the infor-
mation collection and analytics with a very high degree of
automation. The pool of IoT-enabled devices is growing
rapidly and IoT footprint in increasing almost all in
domains that includes trivial to complex applications such
as smart grid [4] to advanced smart cities [5]. The IoT
‘‘Things’’ include a diverse range of embedded objects or
devices such heart monitoring implants, sensor equipped
automobiles, thermostat systems and so on [6]. The
increasing IoT maturity and the advancing cloud-based
services are revolutionizing the information generation,
collection, management and analytics. In IoT realm, the
devices collect useful data and then share the data between
other devices [7]; there are 9 billion such devices that exist
today and the number is inflating. Evans [8] believes this
number will reach nearly 50 billion by year 2020 (Fig. 1b),
which substantiates the fact that the permeation of IoT is
expanding in human life. This indicates that the number of
new applications under IoT umbrella is increasing and
consequently a massive amount of data is being generated
at swift rate, called Big Data [9], where the social media
and sensor embedded systems play central role in data
inundation. As, the IoT-enabled devices access and process
the data from several peer devices to constitute immediate
& Sugam Sharma
sugamsha@iastate.edu; ceo@efeed-hungers.com
1 Center for Survey Statistics and Methodology, Iowa State
University, Ames, Iowa, USA
2 Founder & CEO, eFeed-Hungers.com TM , Ames, USA
3 International Business School Suzhou, Xi’an Jiaotong-
Liverpool University, Xi’an, China
4 Department of Agricultural and Biosystems Engineering,
Iowa State University, Ames, Iowa, USA
5 Department of Computer Science, Iowa State University,
Ames, Iowa, USA
123
Cluster Computing
https://doi.org/10.1007/s10586-018-2821-8 (0123456789().,-volV)(0123456789().,-volV)
decisions and swift actions, there is a continuous need for
equally strong data model, equipped with robust data
engineering and comprehensive analytical capabilities as
the traditional data models are becoming inadequate in
meeting the growing challenges that information-rich IoT
evolution possess.
To appropriately accommodate and adequately address
the growing IoT data challenges, the concept of cloud
computing framework has emerged as a globally accept-
able solution for secure and efficient data engineering and
functionally-rich analytics for technically naı ̈ve stake-
holders, director, managers to tech-savvy users. The
backbone of the widely successful cloud adaptation is the
secure and robust existing or evolving services that are
delivered in the form of *-as-a-Service and the state of the
art as-a-Service cloud modality has gone beyond the core
Infrastructure-as-a-Service (IaaS), Platform-as-a-Service
(PaaS), Software-as-a-Service (SaaS), and Database-as-a-
Service (DBaaS) [10–12]) and this paper highlights some
of them that are highly relevant in IoT context.
Today, a trend of migration, from on-premises to cloud
environment, of applications from almost all possible areas
is increasingly being observed to utilize the best offered
services at substantially reduced cost of cloud computing;
the progressive maturity of IoT is accelerating this migra-
tion. The need for the efficient and successful management
and computation of the scalable Big Data motivates the
scientific communities to devise and develop the new
scalable systems. Consequently, the appeal introduces and
inducts a high abundance of cloud and non-cloud data
models; it becomes nearly impractical for a user to thor-
oughly access such a large set of data models to understand
individual’s technology in search for the most suit-
able cloud model. In this work, we harness the scientific
name of the data models and introduce a novel CNNC
{Clouda, NNClouda| a- assisted, NN- No Name}
classification. The CNNC classification is based upon the
intuitiveness of the scientific name of the cloud data model
and considers two types of cloud data models—(1) the
model contains the term ‘‘cloud’’ in their name (Clouda),
and (2) the model doesn’t have the word ‘‘cloud’’ in its
name at all (NNClouda). In addition, some interesting case
studies are also explored to further strengthening the core
of CNNC classification amid the broadening spectrum of
emerging services and analytics in IoT age.
Rest of the paper is structured as follows. Section 2
provides a brief overview of the emerging Services and
analytics. This section expands on the concept of cloud
computing and IoT. Also, explored are some of the
emerging analytics techniques. Section 3 elaborates on the
novel CNNC classification and investigates some key data
models of Clouda and NNClouda type to further under-
stand and validate the proposed classification. This section
also discusses some interesting case studies. Section 4
explores the emerging cloud services and analytics models,
with the focus on data analytics in vehicular automation
with various examples illustrated. Section 5 further extends
the comparative discussions on the emerging services and
analytics, mentioned in Sect. 4. Section 6 explains the
impacts and benefits of this work to businesses. Section 7
presents the ethical issues on data analytics in vehicular
automation. Finally, the paper is concluded in Sect. 8.
2 Emerging services and analytics: brief
overview
This section provides the brief overview of some core
components of emerging services and analytics. The dis-
cussion on the emerging cloud services is elaborated a bit
comparatively.
Fig. 1 a Traditional versus IoT paradigm (Source http://www.db.in.tum.de/teaching/ws1314/industrialIoT/). b IoT growth over years (Source
www.ncta.com/broadband-by-the-numbers/)
Cluster Computing
123
equally strong data model, equipped with robust data
engineering and comprehensive analytical capabilities as
the traditional data models are becoming inadequate in
meeting the growing challenges that information-rich IoT
evolution possess.
To appropriately accommodate and adequately address
the growing IoT data challenges, the concept of cloud
computing framework has emerged as a globally accept-
able solution for secure and efficient data engineering and
functionally-rich analytics for technically naı ̈ve stake-
holders, director, managers to tech-savvy users. The
backbone of the widely successful cloud adaptation is the
secure and robust existing or evolving services that are
delivered in the form of *-as-a-Service and the state of the
art as-a-Service cloud modality has gone beyond the core
Infrastructure-as-a-Service (IaaS), Platform-as-a-Service
(PaaS), Software-as-a-Service (SaaS), and Database-as-a-
Service (DBaaS) [10–12]) and this paper highlights some
of them that are highly relevant in IoT context.
Today, a trend of migration, from on-premises to cloud
environment, of applications from almost all possible areas
is increasingly being observed to utilize the best offered
services at substantially reduced cost of cloud computing;
the progressive maturity of IoT is accelerating this migra-
tion. The need for the efficient and successful management
and computation of the scalable Big Data motivates the
scientific communities to devise and develop the new
scalable systems. Consequently, the appeal introduces and
inducts a high abundance of cloud and non-cloud data
models; it becomes nearly impractical for a user to thor-
oughly access such a large set of data models to understand
individual’s technology in search for the most suit-
able cloud model. In this work, we harness the scientific
name of the data models and introduce a novel CNNC
{Clouda, NNClouda| a- assisted, NN- No Name}
classification. The CNNC classification is based upon the
intuitiveness of the scientific name of the cloud data model
and considers two types of cloud data models—(1) the
model contains the term ‘‘cloud’’ in their name (Clouda),
and (2) the model doesn’t have the word ‘‘cloud’’ in its
name at all (NNClouda). In addition, some interesting case
studies are also explored to further strengthening the core
of CNNC classification amid the broadening spectrum of
emerging services and analytics in IoT age.
Rest of the paper is structured as follows. Section 2
provides a brief overview of the emerging Services and
analytics. This section expands on the concept of cloud
computing and IoT. Also, explored are some of the
emerging analytics techniques. Section 3 elaborates on the
novel CNNC classification and investigates some key data
models of Clouda and NNClouda type to further under-
stand and validate the proposed classification. This section
also discusses some interesting case studies. Section 4
explores the emerging cloud services and analytics models,
with the focus on data analytics in vehicular automation
with various examples illustrated. Section 5 further extends
the comparative discussions on the emerging services and
analytics, mentioned in Sect. 4. Section 6 explains the
impacts and benefits of this work to businesses. Section 7
presents the ethical issues on data analytics in vehicular
automation. Finally, the paper is concluded in Sect. 8.
2 Emerging services and analytics: brief
overview
This section provides the brief overview of some core
components of emerging services and analytics. The dis-
cussion on the emerging cloud services is elaborated a bit
comparatively.
Fig. 1 a Traditional versus IoT paradigm (Source http://www.db.in.tum.de/teaching/ws1314/industrialIoT/). b IoT growth over years (Source
www.ncta.com/broadband-by-the-numbers/)
Cluster Computing
123
2.1 IoT
IoT is a network constituted by uniquely identifiable
commodity objects or devices equipped with some sensing
system [13]. IoT paradigm promotes a seamless amalga-
mation between the smart devices, scatter around us, and
the physical world to ensure full automation that eventually
ameliorates human life. Some of the examples of IoT-en-
abled commodity devices or things include heart monitor-
ing implants, automobiles with embedded sensors,
firefighter’ devices, smart thermostat systems, and Wi-Fi
enabled washer/dryers [14]. As the arena of IoT is
expanding, the number of IoT-enabled applications is also
rapidly growing, which results in massive growth of smart
devices in multiple order comparatively (Fig. 1b). This
swift increase in the number of sensing things is respon-
sible for generating and storage of a plethora amount of Big
Data at much faster rate. The data needs to be engineered
and analyzed, which require robust services and analytical
systems as the traditional techniques unable to successfully
and efficiently accost such data stress.
2.2 Big Data
IoT paradigm is increasingly encouraging the ubiquitous
connectivity of the intelligent objects within internal or
external world. The continuous rapid growth of large
number of IoT-enabled objects and storage technology
have resulted into the massive amount of heterogeneous
digital footprints and sizeable traces. A vast amount of data
[15] is being generated by various sensing sources every
day. The actual pattern and nature of such data is indistinct,
but is certainly large, complex, heterogonous, structure and
unstructured [16]. Palermo [17] demonstrates some
important attributes of Big Data such as volume, variety,
and velocity and some core constituents of IoT like sensor-
embedded devices, intelligence for quick decision making,
and connectivity for data sharing. Also, the rapid growth of
sensing devices under IoT purview is generating such a
large scale complex and heterogeneous data that the
available computing capacity of the existing systems
(Fig. 2) unable to successfully match up the data chal-
lenges and today, this has emerged as one of the core issues
for the data science community [18, 19]. The storage
capacity and also the processing power of the existing data
computing systems are failed in handling the stress of Big
Data.
As IoT and its applications are going to majorly
affecting the human life, the scientific communities con-
template a broader outreach from the processing and
sharing of Big Data across the variety of the several
commodity devices around us. Consequently, the
development of new capable services and analytics is
encouraged to cater the current data processing and pre-
sentation need. The exploratory analysis of the various
aspects of Big Data in cloud can help in understanding its
important characteristics [20] that may be very valuable to
stakeholders and managers. Sharma et al. [21–23] provide
the elaborated discussions on the characteristics and com-
plexities of Big Data.
2.3 Based cloud services
Based Cloud services (BCS) combine all different Cloud
computing services altogether to produce greater impacts
to the users as any services in place. BCS may include
infrastructure as a service (IaaS), platform as a service
(PaaS), software as a service (SaaS) and database as a
service (DBaaS) [24]. IaaS provide the infrastructure for
users to gain access to use Cloud services, store and save
data, transfer and backup data between their computers and
Clouds. PaaS offers the platform for the developers, users
and service providers to access their services, data and
requests better. SaaS allows users to use any Cloud services
without using any programming but only interfaces to get
their requests completed. Analytics can be used to get
services requests completed in a few clicks. Apps on the
smart phone can allow users to connect and request ser-
vices, check status and make payment. All the data and
information processed and stored in all these services will
be handed by DBaaS, which can be either managed by the
service providers or users depending on the requirements,
chosen service packages and the purpose of doing DBaaS.
The service often includes replication, mirroring, disaster
recovery, backup and data retrieval. Figure 3 shows the
architecture of the Cloud/IoT services, which connects to
all of IaaS, PaaS, SaaS and DBaaS. Users do not need to
know the complexity and can seamlessly use the services to
get their requests completed.
2.4 Integrated cloud computing and IoT
ecosystem
The growing smart communication among the things,
especially the sensor equipped, under the purview of IoT is
resulting into production of incredibly large amount of Big
Data. Supported by the context-aware computing, Big Data
is sufficient enough to address the nontrivial and compre-
hensive tasks with great degree of automation regardless
the disciplines—financial, health, automobile, etc. Soon,
the concept of IoT is going to deeply pervade through the
human life intending to automate the routine chores. The
ubiquitous sensing of all the devices around us emits
enormous quantity of Big Data that needs to be stored,
computed and visualized and analyzed in efficient manner.
Cluster Computing
123
IoT is a network constituted by uniquely identifiable
commodity objects or devices equipped with some sensing
system [13]. IoT paradigm promotes a seamless amalga-
mation between the smart devices, scatter around us, and
the physical world to ensure full automation that eventually
ameliorates human life. Some of the examples of IoT-en-
abled commodity devices or things include heart monitor-
ing implants, automobiles with embedded sensors,
firefighter’ devices, smart thermostat systems, and Wi-Fi
enabled washer/dryers [14]. As the arena of IoT is
expanding, the number of IoT-enabled applications is also
rapidly growing, which results in massive growth of smart
devices in multiple order comparatively (Fig. 1b). This
swift increase in the number of sensing things is respon-
sible for generating and storage of a plethora amount of Big
Data at much faster rate. The data needs to be engineered
and analyzed, which require robust services and analytical
systems as the traditional techniques unable to successfully
and efficiently accost such data stress.
2.2 Big Data
IoT paradigm is increasingly encouraging the ubiquitous
connectivity of the intelligent objects within internal or
external world. The continuous rapid growth of large
number of IoT-enabled objects and storage technology
have resulted into the massive amount of heterogeneous
digital footprints and sizeable traces. A vast amount of data
[15] is being generated by various sensing sources every
day. The actual pattern and nature of such data is indistinct,
but is certainly large, complex, heterogonous, structure and
unstructured [16]. Palermo [17] demonstrates some
important attributes of Big Data such as volume, variety,
and velocity and some core constituents of IoT like sensor-
embedded devices, intelligence for quick decision making,
and connectivity for data sharing. Also, the rapid growth of
sensing devices under IoT purview is generating such a
large scale complex and heterogeneous data that the
available computing capacity of the existing systems
(Fig. 2) unable to successfully match up the data chal-
lenges and today, this has emerged as one of the core issues
for the data science community [18, 19]. The storage
capacity and also the processing power of the existing data
computing systems are failed in handling the stress of Big
Data.
As IoT and its applications are going to majorly
affecting the human life, the scientific communities con-
template a broader outreach from the processing and
sharing of Big Data across the variety of the several
commodity devices around us. Consequently, the
development of new capable services and analytics is
encouraged to cater the current data processing and pre-
sentation need. The exploratory analysis of the various
aspects of Big Data in cloud can help in understanding its
important characteristics [20] that may be very valuable to
stakeholders and managers. Sharma et al. [21–23] provide
the elaborated discussions on the characteristics and com-
plexities of Big Data.
2.3 Based cloud services
Based Cloud services (BCS) combine all different Cloud
computing services altogether to produce greater impacts
to the users as any services in place. BCS may include
infrastructure as a service (IaaS), platform as a service
(PaaS), software as a service (SaaS) and database as a
service (DBaaS) [24]. IaaS provide the infrastructure for
users to gain access to use Cloud services, store and save
data, transfer and backup data between their computers and
Clouds. PaaS offers the platform for the developers, users
and service providers to access their services, data and
requests better. SaaS allows users to use any Cloud services
without using any programming but only interfaces to get
their requests completed. Analytics can be used to get
services requests completed in a few clicks. Apps on the
smart phone can allow users to connect and request ser-
vices, check status and make payment. All the data and
information processed and stored in all these services will
be handed by DBaaS, which can be either managed by the
service providers or users depending on the requirements,
chosen service packages and the purpose of doing DBaaS.
The service often includes replication, mirroring, disaster
recovery, backup and data retrieval. Figure 3 shows the
architecture of the Cloud/IoT services, which connects to
all of IaaS, PaaS, SaaS and DBaaS. Users do not need to
know the complexity and can seamlessly use the services to
get their requests completed.
2.4 Integrated cloud computing and IoT
ecosystem
The growing smart communication among the things,
especially the sensor equipped, under the purview of IoT is
resulting into production of incredibly large amount of Big
Data. Supported by the context-aware computing, Big Data
is sufficient enough to address the nontrivial and compre-
hensive tasks with great degree of automation regardless
the disciplines—financial, health, automobile, etc. Soon,
the concept of IoT is going to deeply pervade through the
human life intending to automate the routine chores. The
ubiquitous sensing of all the devices around us emits
enormous quantity of Big Data that needs to be stored,
computed and visualized and analyzed in efficient manner.
Cluster Computing
123
Today, cloud computing is the most recommended solution
for above mentioned issues of Big Data that promises to
deliver the efficient services as traditional commodities.
The emerging cloud services are being appreciated for
supporting the analyses of Big Data for naı ̈ve to elite
experts such as stakeholder, mangers and data scientists for
their routine tasks. The needs for the highly robust com-
puting and analytical services to address and analyze the
data-related challenges with the exceedingly reduced cost
have further brought the service-rich cloud computing to
forefront. The rising interest in IoT and cloud provisioning
system has triggered the surge in cloud-housed compre-
hensive analytics. In this paper also, we have illustrated
some important related cloud-based analytic services that
are highly efficient for the dedicated tasks and data inten-
sive applications, they have been developed for; gene
structure prediction, image processing, predicting the
relationship of customers and their purchasing patterns and
web searches are some good examples of such data-rich
applications and for the required appropriate analytical
services to analyze their highly complex and heterogeneous
data cloud is the pertinent choice from the security, effi-
ciency and affordability aspects. Academia, industry and
research communities are rapidly developing robust ana-
lytic frameworks and housing them in cloud environment,
and are delivered and utilized as analytics-as-a-Service. In
2008, to deal with the collaborative complex computations,
instead of data in the cloud, Cerri et al. [25] devised an
analytical framework in cloud, called knowledge in the
cloud that delivered multifarious knowledge. Figure 4
shows the cloud computing and IoT ecosystem. It can be
noticed that a sensing system, may differ in attributes, is
embedded into the objects/devices around us. The device
communication with cloud is facilitated through the robust
Fig. 2 Storage vs. Computation capacity (cc) (Source [19]. The auxiliary cc graphs are to clearly show the cc values as 1.43 and 0.018 MIPS in
year 2000 and 1993 respectively
Fig. 3 Base cloud services
Cluster Computing
123
for above mentioned issues of Big Data that promises to
deliver the efficient services as traditional commodities.
The emerging cloud services are being appreciated for
supporting the analyses of Big Data for naı ̈ve to elite
experts such as stakeholder, mangers and data scientists for
their routine tasks. The needs for the highly robust com-
puting and analytical services to address and analyze the
data-related challenges with the exceedingly reduced cost
have further brought the service-rich cloud computing to
forefront. The rising interest in IoT and cloud provisioning
system has triggered the surge in cloud-housed compre-
hensive analytics. In this paper also, we have illustrated
some important related cloud-based analytic services that
are highly efficient for the dedicated tasks and data inten-
sive applications, they have been developed for; gene
structure prediction, image processing, predicting the
relationship of customers and their purchasing patterns and
web searches are some good examples of such data-rich
applications and for the required appropriate analytical
services to analyze their highly complex and heterogeneous
data cloud is the pertinent choice from the security, effi-
ciency and affordability aspects. Academia, industry and
research communities are rapidly developing robust ana-
lytic frameworks and housing them in cloud environment,
and are delivered and utilized as analytics-as-a-Service. In
2008, to deal with the collaborative complex computations,
instead of data in the cloud, Cerri et al. [25] devised an
analytical framework in cloud, called knowledge in the
cloud that delivered multifarious knowledge. Figure 4
shows the cloud computing and IoT ecosystem. It can be
noticed that a sensing system, may differ in attributes, is
embedded into the objects/devices around us. The device
communication with cloud is facilitated through the robust
Fig. 2 Storage vs. Computation capacity (cc) (Source [19]. The auxiliary cc graphs are to clearly show the cc values as 1.43 and 0.018 MIPS in
year 2000 and 1993 respectively
Fig. 3 Base cloud services
Cluster Computing
123
cloud services. However, in IoT infrastructure a smart
connectivity gateway, may be with the existing networks,
is required to further strengthen the device communication.
The devices proactively generate the heterogeneous Big
Data, which is housed in the cloud in various formats and is
available for analysis through robust cloud services.
Therefore, the communication between the cloud and the
external word is feasible only through robust and reliable
cloud services.
2.5 Machine learning and predictive analytics
Machine learning and its supported predictive analytics
provide the intelligent and efficient processing of data and,
the wide usage of these analytical technologies has brought
them to fore in data science, especially for the new incar-
nation of data into Big Data. The machine learning algo-
rithms help recognize the subtle association of the variables
in the dataset, and satisfying all the input criterions, the
analytical systems quickly deliver the output, presented in
the desired form that is easily interpretable by the end users
or stakeholders [26]. The predictive analytics are highly
sought in business domain in the view to elevate the pro-
duct sale. These techniques are useful in predicting the
items types, a customer may be interested to buy and the
associated approximate spending [27]. As the customer
behaviors is closely predictable using the predictive mod-
eling and this enhances the expansion of the business
opportunities and also the customer may be recommended
and served with special offers and discounts.
2.6 Agile methodology
The stakeholders always desire to have the even the com-
prehensive software products ready in a shorter time. Also,
the functional or visual requirements related to that specific
product from the stakeholders change several times over
the span of that product development. To accommodate all
such limitations, restrictions, and complexity and to
promisingly deliver the robust software systems in smaller
time span, standard agile development is highly efficient
[28].
2.7 Statistical analytics
The evolution of data science in the form of Big Data has
brought the statistical analytics to forefront regardless the
domain. The stakeholders are applying the statistical
approaches on their customer databases to derive useful
and accurate recommendations to improve their business.
Also, some statistical frameworks such as R programing
have gained unexpected popularity due to its rich set of
statistical libraries; which are capable to produce the quick
results and their ability for swift, effective and on-the-spot
variation in presentation as desired by the end users or
stakeholders. Also, regression testing is one of the highly
sought mechanism in statistical analysis along with the
easily interpretable presentation of the crucial outcomes
[29, 30].
2.8 Visual analytics
With the recent advancements of data science, where data
has taken a more complex shape in the form of Big Data,
visual analytics have become extremely relevant to deliver
the elegant and informative presentations of the highly
complex data processed for the end users. Visual analytics
also come equipped with some sophisticated statistical
models to be used by the end users or stakeholders [29, 31]
to obtain the desired interpretation. In business domain, the
Fig. 4 Cloud computing and IoT ecosystem
Cluster Computing
123
connectivity gateway, may be with the existing networks,
is required to further strengthen the device communication.
The devices proactively generate the heterogeneous Big
Data, which is housed in the cloud in various formats and is
available for analysis through robust cloud services.
Therefore, the communication between the cloud and the
external word is feasible only through robust and reliable
cloud services.
2.5 Machine learning and predictive analytics
Machine learning and its supported predictive analytics
provide the intelligent and efficient processing of data and,
the wide usage of these analytical technologies has brought
them to fore in data science, especially for the new incar-
nation of data into Big Data. The machine learning algo-
rithms help recognize the subtle association of the variables
in the dataset, and satisfying all the input criterions, the
analytical systems quickly deliver the output, presented in
the desired form that is easily interpretable by the end users
or stakeholders [26]. The predictive analytics are highly
sought in business domain in the view to elevate the pro-
duct sale. These techniques are useful in predicting the
items types, a customer may be interested to buy and the
associated approximate spending [27]. As the customer
behaviors is closely predictable using the predictive mod-
eling and this enhances the expansion of the business
opportunities and also the customer may be recommended
and served with special offers and discounts.
2.6 Agile methodology
The stakeholders always desire to have the even the com-
prehensive software products ready in a shorter time. Also,
the functional or visual requirements related to that specific
product from the stakeholders change several times over
the span of that product development. To accommodate all
such limitations, restrictions, and complexity and to
promisingly deliver the robust software systems in smaller
time span, standard agile development is highly efficient
[28].
2.7 Statistical analytics
The evolution of data science in the form of Big Data has
brought the statistical analytics to forefront regardless the
domain. The stakeholders are applying the statistical
approaches on their customer databases to derive useful
and accurate recommendations to improve their business.
Also, some statistical frameworks such as R programing
have gained unexpected popularity due to its rich set of
statistical libraries; which are capable to produce the quick
results and their ability for swift, effective and on-the-spot
variation in presentation as desired by the end users or
stakeholders. Also, regression testing is one of the highly
sought mechanism in statistical analysis along with the
easily interpretable presentation of the crucial outcomes
[29, 30].
2.8 Visual analytics
With the recent advancements of data science, where data
has taken a more complex shape in the form of Big Data,
visual analytics have become extremely relevant to deliver
the elegant and informative presentations of the highly
complex data processed for the end users. Visual analytics
also come equipped with some sophisticated statistical
models to be used by the end users or stakeholders [29, 31]
to obtain the desired interpretation. In business domain, the
Fig. 4 Cloud computing and IoT ecosystem
Cluster Computing
123
visual analytics are extremely used to envisage the market
and business trends and customer shopping behavior in
order to improve the future decision making to elevate the
business.
3 CNNC classification for emerging services
models
With the growing projection of IoT smart, embedded, and
sensing devices [8], the number of available cloud-powered
models for Big Data management is increasing; every
month, the inclusion of several new models is being
obviously observed and claim to deliver the effective
solutions when deal with Big Data. As the number of the
cloud providers is rising, it becomes nearly impossible for a
cloud-interested user to surf through that giant pool of
models to comprehend their technology and suitability for
his/her purpose, especially from cloud aspect of IoT. In this
case, it is very likely that a user fails to hit the most per-
tinent model due to several reasons such as gradually
eroded user patience, restricted time availability, etc.
In such scenario, to assists the user, we focus on the
name of the cloud models and propose a novel first-level
classification, called CNNC. The objective of this classi-
fication is to gauge the intuitiveness of the name of the
cloud models.
The CNNC classification is believed to reduce the
search time and consequent pain of the user to a greater
degree. Only the cloud-powered data models are eligible
for this classification. The scientific names of the cloud-
assisted models are considered as the basis for classifica-
tion, and are divided into two categories—Clouda and
NNClouda.
CNNC = {Clouda, NNClouda}, where NN is the No
Name, a is the assisted
We further mature and test the proposed CNNC classi-
fication, by applying and explaining it with some promi-
nent cloud-powered IoT- Big Data models. We believe, the
discussion of this classification will invigorate the com-
munities for more intuitive naming convention, where only
the scientific name of a model itself is merely sufficient to
gesticulate about underlying cloud technologies.
3.1 Clouda (Cloud-assisted) models
Definition 1 Clouda represents that class of data models
that have the term ‘‘cloud’’ in their name.
Hypothesis 1 All cloud-powered Big Data models have
the word ‘‘cloud’’ in their name.
To validate this hypothesis, we review some cloud-
supported Big Data models of Clouda nature.
3.1.1 CloudKit (Cloud)
The CloudKit framework [32] is a new technology that is
meant to be used as service for data transportation with (to/
from) cloud. CloudKit is not a local storage rather is a
transport mechanism that is used as the Rack middleware.
It provides an auto-versioned RESTful JSON storage that
has no schema restriction. Although, it is a middleware
component for Rack-assisted applications but, can be used
on its own.
3.1.2 Cloud Datastore (Cloud)
Cloud Datastore (Google Inc. [33] is a NoSQL datastore to
store the schema-free, non-relational data in cloud. Each
instance of the Cloud Datastore is entirely managed by
Google Inc. Cloud Datastore experiences no downtime and
the data is replicated across multiple datacenters, hence
high data availability. Also, the Cloud Datastore scales up
automatically for growing data traffic, endorses ACID
transactions, and offers eventual consistency for queries.
3.1.3 LightCloud (non-cloud)
LightCloud [34], built on Tokyo Tyrant (FAL Lab [35], is
an open source, distributed key-value database that is
always compared with memcached from performance point
of view. It is tested for storing millions of keys on a very
small group of few servers. However, in contrast to
memcachedb [36], the LightCloud is horizontally scalable
with master–master replication architecture and commod-
ity hardware nodes are easily added to achieve it. Although
Salihefendic [37] in his blog quotes about the LightCloud
as ‘‘Plurk’s open-source cloud database LightCloud...,’’
but, the available documentation on LightCloud does not
confirm the LightCloud reliance on cloud.
3.1.4 Cloudera (cloud)
Cloudera lately has drawn widespread attention for its
completely open source, Apache Hadoop distribution,
popularly known as CDH (Cloudera’s Distribution for
Hadoop) [38]. CDH is a regressively tested distribution for
Apache Hadoop, which has been widely deployed with
extensive outreach. CDH is one of the most complete
solution for Hadoop, which provides to the end users the
batch processing of the massive data, support for MapRe-
duce inclusion, interactive SQL and search, etc.
3.1.5 PiCloud (cloud)
PiCloud [39] provides a cloud computing environment that
helps the users to avail the massive computational power
Cluster Computing
123
and business trends and customer shopping behavior in
order to improve the future decision making to elevate the
business.
3 CNNC classification for emerging services
models
With the growing projection of IoT smart, embedded, and
sensing devices [8], the number of available cloud-powered
models for Big Data management is increasing; every
month, the inclusion of several new models is being
obviously observed and claim to deliver the effective
solutions when deal with Big Data. As the number of the
cloud providers is rising, it becomes nearly impossible for a
cloud-interested user to surf through that giant pool of
models to comprehend their technology and suitability for
his/her purpose, especially from cloud aspect of IoT. In this
case, it is very likely that a user fails to hit the most per-
tinent model due to several reasons such as gradually
eroded user patience, restricted time availability, etc.
In such scenario, to assists the user, we focus on the
name of the cloud models and propose a novel first-level
classification, called CNNC. The objective of this classi-
fication is to gauge the intuitiveness of the name of the
cloud models.
The CNNC classification is believed to reduce the
search time and consequent pain of the user to a greater
degree. Only the cloud-powered data models are eligible
for this classification. The scientific names of the cloud-
assisted models are considered as the basis for classifica-
tion, and are divided into two categories—Clouda and
NNClouda.
CNNC = {Clouda, NNClouda}, where NN is the No
Name, a is the assisted
We further mature and test the proposed CNNC classi-
fication, by applying and explaining it with some promi-
nent cloud-powered IoT- Big Data models. We believe, the
discussion of this classification will invigorate the com-
munities for more intuitive naming convention, where only
the scientific name of a model itself is merely sufficient to
gesticulate about underlying cloud technologies.
3.1 Clouda (Cloud-assisted) models
Definition 1 Clouda represents that class of data models
that have the term ‘‘cloud’’ in their name.
Hypothesis 1 All cloud-powered Big Data models have
the word ‘‘cloud’’ in their name.
To validate this hypothesis, we review some cloud-
supported Big Data models of Clouda nature.
3.1.1 CloudKit (Cloud)
The CloudKit framework [32] is a new technology that is
meant to be used as service for data transportation with (to/
from) cloud. CloudKit is not a local storage rather is a
transport mechanism that is used as the Rack middleware.
It provides an auto-versioned RESTful JSON storage that
has no schema restriction. Although, it is a middleware
component for Rack-assisted applications but, can be used
on its own.
3.1.2 Cloud Datastore (Cloud)
Cloud Datastore (Google Inc. [33] is a NoSQL datastore to
store the schema-free, non-relational data in cloud. Each
instance of the Cloud Datastore is entirely managed by
Google Inc. Cloud Datastore experiences no downtime and
the data is replicated across multiple datacenters, hence
high data availability. Also, the Cloud Datastore scales up
automatically for growing data traffic, endorses ACID
transactions, and offers eventual consistency for queries.
3.1.3 LightCloud (non-cloud)
LightCloud [34], built on Tokyo Tyrant (FAL Lab [35], is
an open source, distributed key-value database that is
always compared with memcached from performance point
of view. It is tested for storing millions of keys on a very
small group of few servers. However, in contrast to
memcachedb [36], the LightCloud is horizontally scalable
with master–master replication architecture and commod-
ity hardware nodes are easily added to achieve it. Although
Salihefendic [37] in his blog quotes about the LightCloud
as ‘‘Plurk’s open-source cloud database LightCloud...,’’
but, the available documentation on LightCloud does not
confirm the LightCloud reliance on cloud.
3.1.4 Cloudera (cloud)
Cloudera lately has drawn widespread attention for its
completely open source, Apache Hadoop distribution,
popularly known as CDH (Cloudera’s Distribution for
Hadoop) [38]. CDH is a regressively tested distribution for
Apache Hadoop, which has been widely deployed with
extensive outreach. CDH is one of the most complete
solution for Hadoop, which provides to the end users the
batch processing of the massive data, support for MapRe-
duce inclusion, interactive SQL and search, etc.
3.1.5 PiCloud (cloud)
PiCloud [39] provides a cloud computing environment that
helps the users to avail the massive computational power
Cluster Computing
123
End of preview
Want to access all the pages? Upload your documents or become a member.
Related Documents
Cloud-based Emerging Services Systemslg...
|19
|15351
|209
Internet of Things(IoT) Assignmentlg...
|16
|4425
|66
Analysis of Lightweight Encryption Scheme for Fog-to-Things Communicationlg...
|11
|10056
|476
Internet of Things: Privacy Issues and Contentslg...
|9
|1402
|116
Security threats in cloud computing and preventive methodslg...
|11
|2278
|63
Big Data Challenges in IoT and Cloudlg...
|6
|6502
|62