Decision Analytics Journal 7 (2023) 100200

Contents lists available at ScienceDirect

Decision Analytics Journal

journal homepage: www.elsevier.com/locate/dajour

A robust resource allocation model for optimizing data skew and

consumption rate in cloud-based IoT environments
K. Raghavendar ∗, Isha Batra, Arun Malik
Department of Computer Science and Engineering, Lovely Professional University, Punjab, India

ARTICLE INFO ABSTRACT

Keywords: The Internet of Things (IoT) is a network of connected objects designed to collect and exchange data using
Resource allocation smart equipment and technologies. A significant challenge in guaranteeing a high level of end-user experience
IoT is the administration of IoT services. IoT networks are constructed using a variety of smart technologies such
Cloud computing
as detectors, controllers, Radio-frequency identification (RFID), Universal Mobile Telecommunications Systems
Data skew
(UMTS), Third Generation Cellular Networks (3G), and Global Systems for Mobile communications (GSM).
Average consumption rate
Cloud technology significantly impacts how these networks grow by providing processing capabilities, network
bandwidth, virtualized systems, and system software in an integrated environment. Capacity management,
which assures effective resource use and load-balancing, avoids service level agreement (SLA) infractions, and
enhances machine efficiency by minimizing operational expenses and power utilization, represents one of the
fundamental problems in cloud-based ecosystems. To address these concerns, IoT-based robust decision-making
resource management is often used. In this study, we investigate resource provisioning methods and identify
the factors that must be considered for better utilization of resources in distributed systems. Specifically, we
aim to improve the minimization rate, data skew rate, and approximate amount rate. We also highlight the
challenges and complexities of hybrid optimization for efficient cloud-based capital allocation in the IoT.

1. Introduction being monitored, such as changes in temperature, movement patterns,

or even the presence of people or objects. However, the sheer amount
Smart gadgets and sensor systems that are digitally make up the of data generated by these detectors can be overwhelming, making
Web of Things (IoT). To interact and transport data from one node it challenging to process and analyse the data effectively. Therefore,
to another, these gadgets are connected together utilizing both wired sophisticated data management and analysis tools are required to make
and wireless network innovations [1–4]. Software agencies, sensors, sense of this vast amount of data and extract actionable insights from
smart devices, sensor data, moreover individuals make up the IoT it. Real-time information is kept using cloud services on various local
infrastructure network [5]. To actualize also that Stuff link all gadgets connected computer servers, which then transfer the information ac-
and are incorporated further into physical environment, these inter-
cessibility to the worldwide interconnected data centres for everyone
linked separate devices create local networks and are linked to the
connected phones with worldwide locations [9]. The author of this
worldwide organization [6–8]. Every edge-node inside a cyber–physical
research investigated several resource distribution strategies enabling
environment is conceived as an IoT device which may collaborate
IoT on the cloud systems. Those tactics have just been divided into
with different networking components to fulfil one or more consumer
activities. IoT networks generally have limited access to components groups depending on the QoS, context, cost, energy usage, as well as
such as computing energy, RAM, network connectivity, and memory, SLA. Additionally, its creator covered a range of IoT system allocating
the fact that these infrastructure and the service suppliers supply the resources criteria.
computer capabilities. The detectors in IoT equipment generates a The Internet of Things (IoT) has become an integral part of our
tonne of actual information. IoT devices are equipped with various sen- daily lives with its numerous smart gadgets and sensor systems. The
sors and detectors that generate vast amounts of real-time data. These vast amount of data generated by these devices can provide valuable
detectors can include temperature sensors, humidity sensors, motion insights into various aspects of the environment or object being mon-
sensors, light sensors, and many others, depending on the specific use itored. However, the management and analysis of this data can be
case of the device. The data generated by these detectors can provide challenging due to the limited resources available in IoT networks. To
valuable insights into various aspects of the environment or object address this issue, various resource distribution strategies have been

∗ Corresponding author.
E-mail addresses: [email protected] (K. Raghavendar), [email protected] (I. Batra), [email protected] (A. Malik).

https://doi.org/10.1016/j.dajour.2023.100200
Received 4 February 2023; Received in revised form 3 March 2023; Accepted 10 March 2023
Available online 14 March 2023
2772-6622/© 2023 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
K. Raghavendar, I. Batra and A. Malik Decision Analytics Journal 7 (2023) 100200

Fig. 1. IoT ecosystem basics components.

developed to enable IoT on cloud systems. As a researcher, my personal Identifiers

contribution is to investigate these strategies and categorize them based The recognition process is an essential component of the IoT ar-
on quality of service (QoS), context, cost, energy usage, and service chitecture that provides each item with a clear identification within
level agreement (SLA) criteria. This research aims to provide a better the network. There are two stages involved in recognition: naming
understanding of the challenges and opportunities in IoT resource and addressing. Naming refers to the title given to the object, while
allocation and to develop efficient resource distribution strategies to addressing describes the object’s address. Naming and addressing are
improve the overall performance of IoT networks. two distinct methods of recognition, although they are closely related.
Various methods can be used to accelerate the naming of objects in
2. IoT, internet, and resources top management foundational as- the network, including ubiquitous identifiers such as electronic product
pects codes (EPCs) and uCode. These methods help to provide unique titles
for objects in the network, which enables efficient communication and
2.1. IoT environment data transmission. Addressing is also an important component of the
recognition process [10]. Every object in the network has a unique IP
Numerous advantages and features are available to the consumer address, which helps to identify it within the network. Initially, IPv4
offered by the web of things. Therefore, some components are required was used to assign addresses, but as the number of devices connected
in order to use them properly. This section will cover the IoT compo- to the internet increased, it became impossible to meet the demand.
nents. The components needed to deliver IoT functionality are shown As a result, IPv6 is now employed, which uses a 128-bit identification
in Fig. 1. and location system. This system enables a vast number of unique
The consumer benefits and features provided by the web of things addresses, ensuring that every object in the network can be identified
include and located efficiently. In summary, the recognition process is critical
• Increased convenience: The Web of Things allows consumers to in the IoT architecture as it provides each item with a unique identi-
remotely control and monitor their devices from anywhere, at any fication within the network. The two stages involved in recognition,
time, using a smartphone or other connected device. naming and addressing, are closely related, but they serve different
• Improved energy efficiency: Smart devices connected through purposes. The use of ubiquitous identifiers and IP addresses enables
the Web of Things can automatically adjust their energy usage efficient communication and data transmission in the network, which
based on real-time data, resulting in reduced energy waste and is essential for the success of IoT applications.
lower utility bills. Sensing Devices
• Enhanced safety and security: Smart home security systems, This entails gathering environmental information and delivering it
such as connected cameras and door locks, can help keep homes to nearby, Instances of IoT sensors include remotely or cloud-based
and families safe by providing real-time alerts and remote moni- analytics. We may recognize smart gadgets, mobile sensors, or actu-
toring. ators. The data that has been gathered is sent to the storage media.
• Personalized experiences: The Web of Things enables personal- information about things.
ized experiences through the use of sensors and data analytics, Many detection tools, including those used to acquire data, include:
allowing devices to adapt to individual user preferences and
needs. • Portable devices: Portable devices such as smartphones, tablets,
• Simplified maintenance and repairs: Connected devices can and laptops are commonly used as data gathering tools due to
send alerts when maintenance is needed or repairs are required, their ubiquity and ease of use.
making it easier for consumers to keep their devices in good • Actuators: Actuators are used to convert electrical signals into
working order. physical actions, such as opening or closing a valve or turning

2
K. Raghavendar, I. Batra and A. Malik Decision Analytics Journal 7 (2023) 100200

Table 1
Important technologies components for IoT.
Elements of the Principal systems
internet of things
Classification IPv4 and IPv6 as well as embedded device codes (eCodes) and ubiquitous coding (uCode),
Sensing Actuators, detectors, and monitoring devices include RFID tags.
Communications Wireless Sensor Network (WSN), Bluetooth, Long Term Evolution (LTE), Radio Frequency Identification (RFID), Near Field
Communications (NFC), and LTE. Wireless Sensor Network (WSN) is a group of spatially distributed sensors that cooperate to
monitor physical or environmental conditions. Bluetooth is a wireless technology used to transfer data between devices over short
distances. Long Term Evolution (LTE) is a standard for wireless broadband communication for mobile devices and data terminals.
Radio Frequency Identification (RFID) is a technology that uses electromagnetic fields to automatically identify and track tags
attached to objects. Near Field Communications (NFC) is a wireless communication technology used for short-range transactions
such as contactless payments. Finally, LTE (Long Term Evolution) is a standard for wireless broadband communication for mobile
devices and data terminals that uses advanced antenna technology to improve coverage and data speeds.
Computation Raspberry Pi, Arduino, and Operating system Galil from Intel.
Services A jointly aware information aggregate that connects to omnipresent identity
Semantics RDF, OWL, and EXI

on a motor. They can be used to gather data by monitoring the enabling communication between devices and applications, as it allows
movements or actions of machines or devices. the network to identify which objects or sensors sent the request and
• Smart sensors: Smart sensors are used to detect and measure determine their location and status. This information can be used to
physical phenomena such as temperature, humidity, pressure, and trigger various actions or responses, such as sending alerts, initiating
motion. They are commonly used in industrial and IoT applica- data transfers, or activating specific functions within the network.
tions for data gathering and monitoring. Another service that gathers all the data about the items is called the
• RFID tags: Radio frequency identification (RFID) tags are small aggregate of knowledge. The aggregate services also do the analysis.
electronic devices that can be attached to objects to provide real- The third service is a collaborative effort that uses the information
time tracking and identification. They are commonly used in gathered to take actions and communicates the proper replies to the
logistics and supply chain management for data gathering and devices. The information gathered could include environmental data
inventory tracking. such as temperature, humidity, and air quality, machine data such as
• Wearable sensors: Wearable sensors are small electronic devices performance metrics and error logs, user data such as location and
that can be worn on the body to gather data on physiological behaviour, traffic data such as vehicle speed and volume, and security
parameters such as heart rate, blood pressure, and activity levels. data such as video feeds and access logs. Based on this information,
They are commonly used in healthcare and fitness applications actions can be taken to optimize systems or address potential issues,
for data gathering and monitoring. such as adjusting heating or cooling systems or detecting security
threats. Proper replies are then communicated back to devices to ensure
Instruments for Collaboration
they operate as intended. The service relies on advanced data analyt-
IoT communication systems exchange heterogeneous artefacts to
ics and communication technologies to enable seamless and efficient
create smart services. Communication, which involves connecting and
collaboration between different components in the IoT ecosystem.
enabling interaction between various devices, is among those funda-
The last resource is a widely used services that is used to quickly
mental Objects of IoT. Devices may transport and send letters, papers,
replace technology without restrictions on time or location. This service
and more types of data at the network layers. These devices can
is typically known as ‘‘cloud computing’’ and refers to the delivery
transport and send various types of data at the network layers, in-
of computing resources, such as servers, storage, applications, and
cluding text-based data such as emails, documents, and messages.
services, over the internet on a pay-per-use basis. Cloud computing
In addition, they can also transmit multimedia data such as images,
allows organizations to quickly and easily provision resources without
videos, and audio. With the increasing use of IoT devices, the amount
the need for expensive and complex IT infrastructure. It also enables
of data being transmitted through them is growing rapidly, making
access to resources from anywhere with an internet connection, making
it essential to have efficient and effective communication protocols in
it a flexible and scalable solution for businesses of all sizes. The use of
place to manage this data traffic. Numerous technologies make commu-
cloud computing can provide benefits such as cost savings, increased
nication easier, including Bluetooth [11], Near-field communications
agility, and improved security and reliability. Overall, cloud computing
(NFC) [12], Wi-Fi [13], RFID [14], and long-term evolution (LTE).
has become a widely used and essential resource for organizations
Compute Devices looking to quickly replace technology without being limited by time
Devices are used to compute the information that the devices have or location.
acquired. It is employed to create functionality for the Internet of
Semantics
Things. Hardware platforms include Raspberry Pi, Arduino, and Gad-
The Internet of Things is concerned with making users’ chores easier
geteer, but the computer system is crucial to the execution of software
to do. It is the most crucial element of IoT since it is responsible
applications. Different working platforms, utilized operating systems
for carrying out its duties. It serves as the mind of the IoT. It deter-
include Android, Mini OS [15], Riot OS [16], Lite OS [17], etc.
mines after considering all of the details how to respond by providing
IoT applications communications to the equipment.
4 main operations are offered by programs [18,19]. The initial Table 1 lists the main methods applied every IoT module.
resource connected to an identity. It is employed to find out which
objects sent the request and identify them. The initial resource con- 2.2. IoT architecture frameworks
nected to an identity in IoT applications refers to the first point of
contact or interaction between a device or sensor and the network. This Three perspectives may be used to analyse an IoT architecture:
resource is typically assigned a unique identity or identifier, such as an things-oriented, internet-oriented, and semantic-oriented [20]. Accord-
IP address or MAC address, which is used to identify and locate the ing to a things-oriented perspective, intelligent, automated guided
device within the network. The initial resource plays a crucial role in gadgets are linked together for a variety of everyday uses through NFC

3
K. Raghavendar, I. Batra and A. Malik Decision Analytics Journal 7 (2023) 100200

Fig. 2. IoT’s multi-tier design.

and RFID technologies. The Things-oriented perspective focuses on the tangible objects. An essentially coated nature of the IoT is seen in Fig. 3,
physical objects or devices that make up the IoT system. This perspec- together with the elements that make up each layer.
tive considers the devices’ capabilities, constraints, and connectivity
Layer 1: Physical Equipment and Controls
options and how they interact with each other and the broader network.
A high layer receives data from the external world from this layer of
The goal of this perspective is to design an efficient and scalable IoT
devices or cognition. Physical sensors and components are part of this
system that can handle a large number of devices and data streams. An layer. In essence, this layer’s functions include object identification and
internet-centric point of view concentrates on the interactions between atmospheric data collection, including temperature, pressure, water
these smart devices and the web utilizing distinctive identifiers IP quality, air quality, motion detection, and humidity are a few examples.
addresses, established network technologies in order to cross-border Multiple gadgets may be controlled using controls and physical objects.
interactions among these application-based connected phones. These These are the ‘‘things’’ utilizing the Iot devices, which also includes
interactions typically utilize distinctive identifiers such as IP addresses, a variety of edge communication transmission and receiving devices.
which allow devices to communicate with each other and access the There are now a lot of devices on the market. As more gadgets are
internet. Established network technologies such as Wi-Fi, Bluetooth, being joined to the IoT, it will be essentially limitless.
and cellular networks are used to enable cross-border interactions
among these application-based connected devices. For example, a smart Layer 2: Connectivity
thermostat can be controlled remotely using a smartphone app that The Connectivity layer is a critical component of the IoT architec-
ture that enables communication between the physical sensors and the
communicates with the thermostat over Wi-Fi or cellular networks. In
upper layers of the system. This layer provides the necessary intercon-
this way, the internet-centric approach enables seamless communica-
necting systems like switches, gateways, and routers that allow various
tion and control between smart devices and the web, creating a more
IoT components to interact with each other. It plays a significant role
interconnected and intelligent environment. This approach has signifi-
in transmitting the data obtained from sensors to the upper layers
cant implications for a wide range of industries, including healthcare,
for processing. Data transmission from physical channels to the server
transportation, and energy, among others. According to a semantic
or other machines is a crucial part of the Connectivity layer, and it
perspective on Via IoT architectures, useful information is provided
is done through various communication protocols such as ‘‘ZWave’’,
using data generated by IoT devices and solve architectural modelling
‘‘SigFox’’, Zigbee, or Bluetooth. These protocols ensure that data is
issues effectively [21]. The Semantic-oriented perspective focuses on
transmitted efficiently, securely, and reliably. For instance, Zigbee is
the data generated by the IoT system and how it can be organized,
a popular wireless communication protocol used for IoT applications
interpreted, and used by applications and services. This perspective
in homes and buildings due to its low power and low bandwidth
considers the semantic meaning of the data and how it can be mapped requirements. The Edge Node device, located in the middle of the IoT
to different domains and ontologies. The goal of this perspective is to system, is an essential component of the Connectivity layer. It acts as a
enable data interoperability, data sharing, and data analytics across gateway between the physical sensors and the Cloud transport services,
different applications and domains. providing local data processing, filtering, and aggregation. This helps to
The made a distinction, these three tiers are regarded as network reduce the amount of data that needs to be transmitted to the Cloud,
layer, application layer, and presentation layer. of standard IoT archi- minimizing latency and bandwidth requirements, making the system
tecture, according to the majority of academics. Additionally, some more efficient and cost-effective. Overall, the Connectivity layer plays a
scientists have looked into an intermediate supporting layer between crucial role in enabling communication and data transmission in an IoT
the connection and application layers and is also a part of the most architecture. It is essential to design and implement this layer correctly
recent IoT design. Fig. 2 shows the IoT’s multi-tier design. to ensure the reliability, efficiency, and security of the system. Proper
Technology in the fog and cloud make up the support layer. The selection of communication protocols and interconnecting systems is
seven-tier architecture, which incorporates procedures for collabora- critical for ensuring smooth communication between IoT components,
tion, is described in this section, apps for data extraction and accu- enabling efficient data transmission, and enabling the Edge Node device
mulation, network virtualization, networking, a controller, as well as to act as an efficient gateway.

4
K. Raghavendar, I. Batra and A. Malik Decision Analytics Journal 7 (2023) 100200

Fig. 3. IoT’s basic structure.

Layer 3: Edge Connecting the required application format [22]. These sources could include
Cloud boundary, commonly referred to as cloud technology, is the sensors, RFID tags, smart devices, and other connected objects that
next phase. The connection phase’s data is needed for Layer 3 to archive generate data in various formats and at different rates. The data ag-
and analyses information at a higher level. The processing certain gregation layer is responsible for collecting and processing this data,
facts could be reduced to minimal quantities of data since this layer’s removing duplicates and errors, and transforming it into a standard
modules deal with a lot of data. edge connecting refers to the next format that can be easily analysed and interpreted by other layers.
phase in IoT architecture, where data is processed and analysed at the By combining information from multiple sources, this layer enables
boundary between the cloud and edge devices. This phase is crucial more accurate and comprehensive analysis of data, leading to bet-
for optimizing the performance and efficiency of IoT systems, as it ter insights and decision-making. Overall, the data aggregation layer
enables faster processing and reduces the amount of data that needs plays a critical role in ensuring that IoT systems can efficiently and
to be transmitted to the cloud. At the edge, data can be pre-processed effectively collect, store, and analyse data from multiple sources. A
and filtered to remove unnecessary information, allowing only the most publishing computer infrastructure or information dispersion service is
relevant data to be sent to the cloud for further analysis. This approach a crucial part of the huge high implementation infrastructure (DDS),
not only reduces the amount of data that needs to be transmitted, which makes it simpler to transfer data comparing cloud technology,
but also reduces latency and improves system response times. Edge data aggregation, mechanisms, user layers. This infrastructure makes
computing modules typically include sensors, actuators, gateways, and deployment easier and boosts speed for all applications – aside from
other devices that can collect and process data in real-time. Overall, the most straightforward ones – regardless of whether they use a
edge connecting plays a vital role in the success of IoT systems, as it quick messaging bus or a high-performance solution. The term ‘‘most
enables more efficient and effective data analysis and decision-making. straightforward’’ refers to applications that have less complex data
processing and transfer requirements. These applications may use a
Layer 4: Data Accumulation simple messaging bus to transfer data between nodes in the network.
The data should be kept once it has been gathered. The location On the other hand, high-performance applications that require fast
of information storage is crucial. While some data may be retained and efficient data processing and transfer may use more sophisticated
indefinitely, most must be uploaded to the server, big data technologies solutions, such as high-performance messaging systems or distributed
would be able to examine the information using their computational data aggregation mechanisms. These solutions enable the applications
capacity. The primary goal of this layer is to preserve Phase 3 informa- to handle large volumes of data and process it quickly, making it
tion. Make a many data available to the upper layers by gathering, stor- possible to meet the performance requirements of the application. Re-
ing, and placing it in the warehouses. In order to provide information gardless of the complexity of the application, the publishing computer
for higher layer query-based processing, it merely alters event-based infrastructure or information dispersion service provides a platform
data. In addition to supporting Mongo, Spark, Hadoop file systems, Cas- that simplifies deployment and boosts the speed and efficiency of data
sandra, and other NoSQL databases with greater sophistication, this transfer, making it easier to develop and maintain high-performance
layer may also be built in SQL. applications in the context of the Internet of Things.
Layer 5: Extracting data Layer 6: The Application Layer
Information from several sources are efficiently and effectively com- It has to do with deciphering data from different IoT apps. Several
bined by this layer, which also transforms the stored information into IoT applications are included in it, such as connected automobiles,

5
K. Raghavendar, I. Batra and A. Malik Decision Analytics Journal 7 (2023) 100200

Fig. 4. Applications for IoT supported by cloud.

smart cities, smart networks, smart buildings, smart agriculture, and maintain physical hardware. Hybrid, public, private, and community
more [23]. The software application of the data package and the control cloud deployment methods are the four additional deployment meth-
software are carried out in this phase, which is self-explanatory. Just a ods offered by cloud computing in accordance with user preferences
few examples of IoT applications include Process improvement, logisti- and functionalities [29]. IoT one of the many due of its numerous
cal planning, statistical analysis, control signals, watchful observation, features and capacities, cloud computing is being used in creating
alert management, and consumption models. technologies. Several real-world uses, such as smart cities, healthcare,
and the agricultural industry mobility, advanced vehicles, among many
Layer 7: Collaboration and Processes
other things, have heavily utilized the integration of Internet with
Users are now presented with application processing and collab-
Cloud Computing [30]. The use of smart gadgets that are connected to
oration, whereas the lowest layers’ processed data are employed for
the internet globally is one of the current technological trends. These
business goals. This contains groups of people working together, enter-
devices generate enormous amounts of content which, due to a shortage
prises, and methods for making decisions using data from the Internet
of storage capacity, cannot be saved locally. Devices that generate enor-
of Things.
mous amounts of content include smartphones, tablets, smartwatches,
Its layer categorizes individuals who can effectively work together
fitness trackers, and other internet-connected gadgets. These devices
and interact while using IoT information. As provides further capabili- collect data and information from various sensors, such as GPS, ac-
ties like the ability to produce graphics and models for commercial use celerometers, and heart rate monitors. They also allow users to generate
as well as other database-based application layer recoveries. Executives content, such as photos, videos, text messages, and social media posts.
can also use it to make correct data-driven business decisions [24]. The reason why these devices generate enormous amounts of content is
due to their ability to capture and transmit data in real-time. Users can
2.3. IoT infrastructure is supported by cloud technology easily capture and share their experiences, thoughts, and activities with
others through various digital platforms. Additionally, the increasing
Technology’s National Bureau of Standards (NIST) describes Using popularity of social media, online streaming, and cloud storage has
cloud technology as a framework allows end users to access a global, also contributed to the growth of content generated by these devices.
suitable, on-demand pooled pool of services, computational power This vast amount of content presents both opportunities and challenges
(applications, storage, network, and services) using the pay-per-use for businesses and organizations, as they need to find ways to leverage
model least quantity of effort or service provider communication [25]. and manage it effectively. Data is a lifeblood of all industries; without
Due to its low cost and quick elastic capabilities, cloud computing is analytics, data is meaningless for obtaining the information needed for
being quickly adopted by IT sectors and professionals for project de- future plans and strategies. Smart connected devices gather enormous
ployment [26]. The core components of cloud technology, a Web tech- volumes of data, necessitating powerful computer systems that must
nology, are on-demand awareness, widespread internet connectivity, be able to process such enormous amounts of data. The individual
demanding capacity, quick adaptation, and quantifiable services [27]. systems do not have enough capacity to analyse the information for
By employing the software development concepts, frameworks, and analysis. Using the collective resources offered by cloud technology,
utilities made available by the cloud environment, Platform as a Service these restrictions of vast memory capacity, processing capabilities, and
(PaaS) allows cloud clients to build their services on the provider’s connection bandwidth may be solved. IoT can operate smoothly thanks
architecture. The Infrastructures as a Service (IaaS) paradigm enables to cloud technology, as seen in Fig. 4. Cloud storage infrastructures is
users to install different components, including networks, computa- used as a hidden layer between applications and IoT devices, concealing
tional power, memory, and other fundamental computing needs, in all aspects from the creation of Internet-of-things apps [31]. Cloud
order to run any applications or operating system installed on a vir- storage infrastructures are used as a hidden layer between applications
tual machine (VM) by the web resource provider [28]. This model and IoT devices to provide a centralized and scalable storage solution
provides users with the flexibility to customize and configure their for the vast amount of data generated by these devices. Cloud storage
computing infrastructure as needed to support their specific application allows IoT devices to transmit data to the cloud, where it can be stored
requirements. For example, users can deploy virtual machines (VMs) and processed by applications and services. This approach enables IoT
to run any operating system or application, and they can also scale up devices to be lightweight and energy-efficient, as they do not need
or down their computing resources in response to changing workload to store or process large amounts of data locally. Additionally, cloud
demands. Additionally, users can typically control the networking and storage infrastructures can provide advanced features such as data
security settings of their infrastructure, allowing them to configure analytics, data visualization, and data mining, which can be used to
virtual networks, firewalls, and other security measures to protect their extract valuable insights from the IoT data. The use of cloud storage
applications and data. Overall, the IaaS paradigm provides a high infrastructures also enables flexible and scalable deployment models, as
degree of control and flexibility to users, enabling them to build and organizations can easily add or remove devices and applications from
manage their computing infrastructure without having to invest in and the network as needed.

6
K. Raghavendar, I. Batra and A. Malik Decision Analytics Journal 7 (2023) 100200

Fig. 5. Three-tiered structures (IoT, Cloud, and Edge).

2.4. The IoT needs resource management The variety and distance of IoT devices provide various issues for re-
source allocation. Many obstacles and issues connected to the allocation
The Quality the Service (QoS) standards are guaranteed by a highly of IoT resources have not yet been addressed despite extensive research
effective, well-maintained, and cost-effective network. IoT architecture by academic and industry researchers. Many academics have developed
connects a variety of adding assets to the network. Allocating resources novel approaches and algorithms to address these problems. In this
is a key element of QoS requirements since it is done using IoT network work, the author reviews a number of resource allocation methods that
judiciously and successfully. As a result of data in IoT architecture have been put out for the Cloud-based IoT environment and categorizes
is split up into several data streams obtained from various sensors them according to their features and the resource allocation criteria
and that networked devices provide a variety of services, resource that they have improved. The author also covered a number of IoT
allocation is also responsible for maintaining strong security require- allocation of resources characteristics, including some that still need
ments. The networked resources used by the IoT include computational development.
components, storage, and energy. IoT networked devices can better
utilize cloud resources to increase system efficiency and productivity by 3. Associated works
effectively allocating them to them. Because IoT resources and devices
are heterogeneous and widely dispersed, allocation of resources and IoT designs, wireless sensor networks, IoT integration with cloud
management are crucial items that make up the IoT ecosystem. Both computing, and surveys and reviews of IoT technologies have all been
the system and the IoT network app user connect with each other
the subject of several academic articles. On the other hand, there is
through an entity, which can be an object, a person, or a location. These
just one survey article available about resource allocation in an IoT
allotted items are referred to as resources, and they can be divided into
system. Various facets of IoT architecture are the focus of numerous
different groups according to the data they exchange in a network.
researchers. They put forth strategies and innovations that enable IoT
According to Fig. 5, The IoT system’s three levels are separated into
networked systems for effective and efficient operation. For cloud-
IoT objects, edge computing, and cloud infrastructure. The system had
based IoT allocation of resources methods, there is no review paper
to choose the resources it would need from these three levels first,
available. In [32], the author provided a significant survey that covered
then it had to choose the nodes that could do the user’s job, and
different difficulties and problems with allocating IoT resources from
last it had to Plan for the user’s task to be finished on the chosen
an architecture perspective. The author of this review study explored
node. Lastly, interaction between both network servers is carried out to
the architectures and infrastructure that control how resources are
guarantee that the user’s task is effectively fulfilled. Just a few of the
many elements of resource distribution include resource provisioning, distributed in an IoT environment. The strategies for scheduling and
resource monitoring resource identification, and resource scheduling. maximizing resources from a under this research, cloud viewpoint is
The necessary Service Quality (QoS), efficiency gains, reduced power not explored [33]. Is a review study that examines the difficulties
consumption, improved resource utilization, and, most importantly, associated with various IoT devices? The author of this article explores
the support level agreement among cloud-based Iot device software how fog computing and IoT are related and also discusses several
vendors and their consumers, in which customer requirements must resource allocation techniques presently in use in situations where fog-
be accurately paired, are all supported by efficient resource distri- based IoT is prevalent. According to the author, who also identified
bution. The success of cloud-based IoT applications largely depends two additional significant concerns In an IoT architecture built on fog,
on the Service Quality (QoS) provided to the end-users. It is crucial first off, edge computing only considers the least latency and is not
to achieve efficiency gains, reduce power consumption, and improve concerned with the closest network that offers computing power or
resource utilization for better QoS. The support level agreement among computational capabilities. Distributing resources between IoT smart
cloud-based IoT device software vendors is the most crucial aspect, devices and fog computing is the second problem the author identified.
as it determines the level of support and maintenance provided to This is due to the restricted resource capacity of fog computing. The
the end-users. The efficient collaboration of different vendors and the author advocates using cloud computing over fog computing to address
effective utilization of resources can lead to significant cost savings and the issue of limited fog resource availability. This study’s author looked
enhanced performance for IoT applications. Therefore, it is important at a number of IoT resources management via the web strategies and
to consider all these factors while designing and deploying cloud-based split them to distinct categories. These techniques’ further limitations
IoT solutions. and advancements are also covered and presented in tabular style.

7
K. Raghavendar, I. Batra and A. Malik Decision Analytics Journal 7 (2023) 100200

Table 2
IoT resource allocation methods with SLA awareness.
Algorithm Improvement Limitations
Resource allocation based on SLAs [34] Boost system performance and decrease SLA Alternative approaches are not compared with the
violations. suggested strategy.
SLA-aware cloud-based resource allowance control To decrease SLA violations, the user’s work is The job waiting time is not defined in a dynamic
[35] broken down into smaller tasks. manner.
IoT resources and services delegation [36] Boost system effectiveness and performance. There is not any proof of actual implementation.
Allocating resources for Iot [37]. By decreasing the mark up of task arrival, you can The suggested approach does not function in a
increase system performance and decrease SLA multi-tenant setting, such as several data centres.
breach.

cost of the penalty. In article [35], alternative SLA violation-based

strategy is described. In order to decrease SLA violations, the author
recommended a method for effectively scheduling and limiting the
user’s job. The suggested method breaks down user labour into a
number of smaller jobs and increases server capacity in order to reduce
overall task execution measurement time, eliminate SLA violations, and
increase vendor profitability [36]. Presents an allocation of resources
technique for IoT devices that is mindful of SLAs. The author sug-
gested a Delegating services and allocating resources method based
on architecture for an IoT environment based on processing in the
fog and the cloud. A decision tree that has been linearized is used to
handle user activities in an algorithm that focuses on SLA and QoS.
According to [37], a new IoT resource allocation problem focused
on SLA was resolved by buffering scheduling and capping the speed
during which clients can perform tasks in order to achieve SLA. The
suggested solution performs better in IoT networks with high IoT device
Fig. 6. Classification of IoT allocation of resources techniques. traffic and conforms to SLA violations without knowing the user’s task
arrival pattern in advance. Table 2 describes these algorithms, their
advancements, and their drawbacks.
There are several parameters for resource allocation that have been
4.2. Allocating IoT resources with context
described and specified. This article discusses about both the level of
progress made in these factors and the amount in the other criteria are
still required. Numerous studies have used game mechanics to allocate resources
in device-to-device interaction with IoT devices. Effective data trans-
4. Fog IoT resource planning categories mission for device-to-device interaction depends on resources planning
in a wireless network [38]. Introduced a location-aware approach
Approaches the author studied and categorized fog IoT resource for D2D communication in a cell connection using the Evolutionarily
provisioning methods in this area based on the characteristics each Stable strategy. The author suggested a situation-specific context-aware
methodology supplied as part of the research. Fig. 6 describes the algorithm that calculates the network’s bandwidth to make the most of
precise classification of IoT allocation of resources methods. These how much each station can communicate overall [39]. Identifies the
types of IoT resource distribution have been investigated in many communication equipment’ two separate pairing processes as the cell
articles. affiliation issue. Through the correlation among the gadgets in each lo-
cal area connection, the suggested model enhanced the cell association
4.1. IoT resource allocation with SLA awareness approach and assured a common outcome. Another optimal resource
allocation strategy based on the association of one or more devices was
The service level agreement is present in every service-oriented put forth in [40] and improved resource consumption among network
infrastructure here between suppliers and the customers is crucial. The devices. The suggested context-aware resource allocation mechanism
SLA should be violated as little as possible to maximize revenue for between friend IoT network systems has succeeded in meeting dynamic
the service provider and satisfy customer needs. Numerous Various QoS requirements. The advantages and disadvantages of context-aware
research methods have been used to reduce SLA in this area of study allocation of resources approaches are shown in Table 3.
violations and increase customer acceptance of the system. Service
level agreement (SLA) is a contract between a service provider and 4.3. Allocating IoT resources with QoS awareness
a customer that outlines the level of service that will be provided,
as well as the penalties or consequences if the service falls below The service-based application must take Quality of Service into
the agreed-upon level. SLAs typically include metrics such as uptime, account (QoS). There should be a match between the service level
response time, and resolution time, and are designed to ensure that agreement and service quality (SLA). Quality of Service (QoS) refers
the service provider meets the customer’s expectations. Additionally, to the level of performance and reliability that a network or service
some research is being done on SLA-focused resource allocation for IoT- provides to its users. QoS is typically measured by metrics such as
enabled systems. The author took into account the cost of the fine for throughput, latency, and packet loss, and is designed to ensure that
an SLA violation in [34]. the network or service meets the user’s performance requirements.
By computing the customer’s gain and announcing a champion to QoS can be implemented using various techniques, such as traffic
the client who gives the sources the largest profit, a computationally shaping, prioritization, and congestion control, and is essential for
auction approach is offered to reduce the penalty for SLA breach. In ensuring that critical applications and services receive the necessary
order to discourage SLA violations, this approach might reduce the resources and bandwidth to function properly. This topic has been

8
K. Raghavendar, I. Batra and A. Malik Decision Analytics Journal 7 (2023) 100200

Table 3
IoT resource allocation methods that consider context.
Algorithm Improvements Limitations
D2D communication using game theory in an IoT By applying the Evolutionarily stable game model Only the Cloud is the focus of this strategy, and it
network that is centred on the cloud [38]. and maximizing band-width use, D2D is contrasted with the current algorithms.
communication is improved.
Utilizing a correlation-based resource allows for For improved performance in the cell association The technique makes no reference of the empirical
action [39]. approach, Duplicate data produced by the evidence.
numerous sensors has been removed.
Method for Allocating Resources Optimally [40] Performance and resource utilization have both There is no concrete proof for the procedure.
improved.

Table 4
IoT allocation of resources methods with QoS awareness.
Method Developments Drawbacks
Method for allocating downlink resources [41] Systems performance has improved and Practical application is lacking.
interference among the communication channels
has been minimized.
Scheme for allocating radio resources [42] The volume of traffic between M2M sensors has The technique is not used in actual situations.
decreased.
Work distribution among an IoT device group [43] With a 5% inaccuracy, the ideal resource So much focus must be given to the QoS
allocation has been accomplished. component.
IoT work scheduling based on consensus [44] Both chatter and broadcast techniques are Real-world situations and QoS are less taken into
employed, but broadcast produces the best account.
outcomes.

the subject of several studies for various scenarios. Device to device in terms of QoS and resource limitations, [47] derives the generalized
communication was less intrusive because to a QoS-based strategy Nash equilibrium (GNE) approach and its special criteria. The devices
to distributing IoT resources that was published in [41]. The usage are able to reach CH equilibrium (CHE) for rationally corresponds
of resources can be restricted to D2D users by using the PFR ap- to the varied computing capabilities and access information of each
proach and incursion restricted area management strategy. The wireless MTD and HTD thanks to the employment of a cognitively hierarchical
connection route gain controls the Device-to-device users’ access to game theory in this manner. Our suggested strategy cuts MTD energy
services and also balances workload and improves system efficiency. use by 78%. By regulating process admission and resource utilization
In order to comply with QoS requirements, a mathematical formalism for the IoT system, a new ground-breaking to further improve system
for the heterogeneous pedestrians of M2M sensors has been provided performance, ECIoT framework has been presented in [48]. For make
in [42]. An optimization protocol depending on the resource provision- the system more useful, ECIoT employs a cross-layer dynamic network
ing method for heterogeneous IoT systems has been introduced in [43]. optimization approach based on Lyapunov optimizing parameters. The
The programme took into account the communications nodes’ work shortcomings and advancements made in these suggested algorithms
rate and buffering exhaustion. The strategy can adapt to the dynamic are shown in Table 5.
and varied features of networked IoT devices. An interaction allocation
of resources approach using two the Internet of Things were published 4.5. Cost-aware IoT resource allocation
in [44]. To communicate and update research shows among the nodes,
these strategies made use of chat and broadcasting mechanisms. The The Internet of Things (IoT) network’s heterogeneous, extremely
proposed technique was evaluated under three possible scenarios: the resource-intensive smart devices are managed via the cloud, edge com-
whole system, a specific job at a specific rate, and a specific job at puting and fog, infrastructures. To complete their functions and fulfil
a whole frequency. To decrease message transfer and boost system QoS standards, these numerous network infrastructure devices demand
dependability, the system’s output errors has been decreased to 5% resources. The services provided and activating fees for each connection
as compared to conventional approach. Table 4 lists the benefits and interface device were used to compute the total consumption expen-
drawbacks of QoS-based resource allocation approaches. diture. The Service-to-Interface Allocation Cost Problem is the name
given to this cost estimating issue. Two SLA techniques are theoretically
4.4. IoT utilization of resources with energy awareness presented in [49] to reduce the cost of computation. The first technique
satisfied the resource’s demands in a single round, but the second
In order to decrease ecological footprint and create a green compute way satisfied it over the course of several rounds. To lower the cost
cluster, it is crucial to manage the effective use of energy or power of activation, The recommended remedy splits the cost of activation
since the enormous diverse and power-hungry smart gadgets in large and allocates it among multiple access points. Resource release and
numbers. In the past, many scholars have put in a lot of work has allocation must be strategic and effective in order to lower service
gone into lowering power use across a variety of fields, like edge costs for cloud resources. To monitor the resource delivery usage by
computing, cloud computing, and computing in the fog. The decrease in the Internet of Things, a multi-agent-based Fog allocation of resources
energy use in IoT-based contexts has shown some signs of success. The technique has been developed [50]. The audit of resource consumption
‘‘virtual cloud of anything’’ (V-FOE) and the ‘‘digital fog of virtually aids in reducing inefficient resource use and increasing resource uti-
anything’’ allocation of resources technique’s thermal efficiency are lization, which enhances performance and lowers system costs overall.
documented in [45] (FoE). Three techniques of interaction between It has been proposed to reduce the cost of network resources using
the equipment in the system of cloud computing system (FOCAN) [46] an alternative strategy [51] based on the Stackelberg game model.
have been established in order to efficiently achieve the QoS and energy In an effort to cut costs, the approach looks at the bottom and top
consumption. The architecture of smart cities with fog assistance uses layers of the network and confirms the Nash equilibrium position of the
this method to spread infrastructure functions among digital sensors. no-n collaborative game played between them. By creating the Stack-
In order to address the heterogeneity of Internet-of-things resources elberg game for the full network, an iterative approach is employed

9
K. Raghavendar, I. Batra and A. Malik Decision Analytics Journal 7 (2023) 100200

Table 5
Energy-Conscious allocating IoT resources methods.
Algorithm Improvements Limitations
Where the energy-efficient resources are located in Reducing energy usage, delaying and improving Neither real-time monitoring has taken place.
Fog of Everything FOE system efficiency
A smart urban architecture for resource The technique lowers power consumption while There has been no execution.
distribution called FOCAN [46] increasing delay.
For allocation of resources, consider the perceptual The process cuts electricity consumption by 78%. One simulation alone has been run.
hierarchy theory [47].
The allocation of resources under joint check System performance improvement and There is no proof that it has been used in practise.
pointing [48] communication latency reduction

Table 6
IoT allocation of resources methods that consider cost.
Methods Developments Drawbacks
Diverse utilization of resources for adaptable By distributing the goods among the terminals, the The cost of using the multi-round approach has
services [49] cost has decreased. somewhat increased.
Architecture for clouds resource allocation Boost system efficiency and cut costs overall by Reduced resource utilization costs are necessary.
optimization [50] using VMs as little as possible.
Heterogeneous-oriented recognizes for re-sources Cost and resource utilization have decreased There is no proof of practical in the
[51] implementation
Flexible allocative use of heterogeneous assets [52] Costs have decreased, and system performance has The topic is largely theoretical and has not been
improved. put into practise.

to get the Nash equilibrium. IoT devices have made use of a wide □ SLA to avoid cost overruns and lower the chance of user and supplier
range of services and heterogeneous network interfaces [52]. Creates interaction violations, It is necessary to uphold the between the product
a mixed-integer programming-based specification-based resource to customer and the service supplier, a service contract.
heterogeneous service model (MILP). The price of the facilities might □ Time Because of the enormous bulk of data that smart devices
be cut by dynamically allocating them among the different interfaces. frequently create, it is an important part of the IoT ecosystem. It is a
Table 6 discusses the development and constraints of cost-conscious strategy to plan out a task’s completion in an IoT context.
allocation of resources methods for IoT environments. □ Cost IoT environments utilize a number of services from cloud infras-
tructure providers and pay the service provider money in exchange. The
5. IoT resource planning approaches variables system’s productivity and performance are used to calculate the cost of
services. The system should be as inexpensive as possible.
5.1. Parameters of resource allocation □ An extensive network of intelligent devices and network nodes
connected to big data centres make up the energy IoT ecosystem.
The correct operation of these technology and data storage requires
The distribution of resources is essential to the ecosystems based on
a substantial how much power. In order to cut costs and eliminate
the fog Iot technology. The IoT environment’s linked devices generate
carbon footprints from the environment, energy usage should be kept
enormous a lot of information is kept on the web quantities for later
to a minimum.
analysis to provide information that is useful for the system’s architec-
□ Accessibility Usually accessibility of a resource is a measurement
ture and applications. There have been several allocations of resources
of its timeliness and dependability within a certain time frame. To
parameters covered in [53,54]. The author outlines the different alloca-
decrease service delays, there should be a high level of available
tion of resources parameters for an IoT environment in this part. These
resources.
aspects must be considered while developing IoT resource allocation
□ Utilization The percentage of a resource that the system must con-
methods are improved.
sume is called resource utilization. To increase the system’s efficiency
□ Performance for the IoT layer to complete the user’s task, there is a and durability, the resources should be used as effectively as possible.
certain amount of effort involved. The system needs to operate at peak
efficiency. 5.2. Improvement have been done in these parameters
□ Bandwidth the overall number of activities or jobs that the Iot device
has done is used to measure its efficiency. To complete all of the user’s The researcher examined a several cloud-based IoT resources al-
tasks, Iot applications must have high throughput. location approaches Within the research and categorized them into a
□ the quality of services that the cloud system will provide to the client number of classes. This section has examined the allocation of resources
in compliance with the SLA commitment is measured by QoS, or service characteristics that are improved by the methods listed in the category
quality. and the parameters have been marked to match to the method or
□ Delay the length of period require to react to a customer’s pend- algorithm being researched. We can see which parameters have grown
ing job while the device is also carrying out additional task. To im- the most and which still need additional study according to field of
prove system performance, the channel’s latency should be as little as science (see Fig. 7).
possible. As previously saw in the section above that the techniques or ap-
□ Bit rate is the pace at which Internet of Things Information is sent proaches offered have increased the quantity of criteria for allocation of
between locations by (IoT) devices. In a world wide web IoT context, resources. However, a lot of allocation of resources indicators still need
data transfer rates must be high. to be improved and researchers must pay close attention to this. Fig. 8
□ Dependability is the capacity to complete the assigned work on displays several parameters together with the number of algorithms
schedule without being adversely affected by a system breakdown. in which they can be enhanced. The chart clearly shows that several
To guarantee that users complete their tasks on time, the system’s researchers have greatly enhanced the security level. The connectivity
dependability should be excellent. broadcaster’s latency, delay, and resource availability, however, have

10
K. Raghavendar, I. Batra and A. Malik Decision Analytics Journal 7 (2023) 100200

Fig. 7. Level of variable modification for management of resources.

Fig. 8. Showing the overall hybrid cloud set up for optimization rate, skew rate and consumption rate analysis.

not changed considerably. There is also potential for academics to resources better provisioning criteria over time. Researchers can use
suggest additional resource distribution methods it may make these several objective and naturally derived methods.

11
K. Raghavendar, I. Batra and A. Malik Decision Analytics Journal 7 (2023) 100200

6. Problems and obstacles their cost. When compared to cutting-edge blocking-based load balanc-
ing algorithms on a Hadoop cluster in a cloud architecture, our solution
The Internet of Things ecosystem offers possible services that will performed best than earlier methods in terms of performance speed and
increase the effectiveness of the IoT ecosystem. The IoT in the web capacity.
resource distribution confronts a variety of challenges in spite of the In order to increase make span and load balancing, Rohail Gulbaz
potential benefits. This article discusses a number of research on the et al. [59] offer the Balancer Genetic Algorithm (BGA), a revolution-
topic of IoT resource distribution. IoT with edge, cloud, and fog, ary load balancing scheduler. As part of the resources are left idle,
computing cooperation requires more study to be done for a successful insufficient load balancing might result in an overhead of resource use.
integration of the Iot platform. The majority publications are based on BGA incorporates an approach to load balancing that considers the real
one or two elements of allocation of resources. The majority of the weight, measured in million directives issued to VMs. It is also em-
research that has been done thus far has been focused on simulation, phasized that multi-objective optimization should be used to enhance
and there is no proof that the suggested method has been used in a load balancing and task scheduling. In experiments, work distribution
genuine IoT environment. Whenever these techniques are applied in the that are skewed, normal, and uniform as well as various batch sizes
real IoT environment, there may be issues because there has not been are employed. When compared to several cutting-edge techniques for
a comprehensive implementation testing in the real IoT environment. makespan, throughput, and load balancing, BGA has shown appreciable
The actual IoT ecosystem has its own difficulties and quirks. Additional improvement.
issue is that the resource allocation methodologies evaluated in this
study do not take into consideration all elements of resource distribu- 7.2. To determine the data skew load balancing optimization system effi-
tion, including resource monitoring, resource identification, resource ciency for different scenarios such as data skew rate
provisioning, resource modelling, resource estimation, and resource
scheduling. Another problem is that research tends to focus more on re- Since the construct of the IoT was first introduced, which permits
source allocation strategy than optimization of allocation of resources, the orchestration and administration of a significant number of Web
which is more crucial for a successful and productive IoT ecosystem. devices, there has been a trend towards the centralization of IoT
services (data and apps) running on the cloud. IoT Gateway is in charge
7. Research work of minimizing the frequency of network link(s) overloads or failures
connected to an IoT gateway to manage a group of sensors through IoT
7.1. To determine the data skew load balancing optimization system effi- management services efficiently. We use a fog of things load balancing
ciency for different scenarios such as optimization rate strategy to solve these issues, and in this section, we describe our
findings, deployed and implemented achievements, and evaluation of
The Locally Shortest Queuing (LSQ) class of load balancing methods outcomes [60].
is introduced by Shay Vargaftik et al. in [55]. In such approaches, each Simulated annealing-based load-balanced cluster system’s major ob-
controller continues to utilize JSQ on its local view while maintaining jective is to prolong network lifespan while maintaining acceptable
its own, perhaps out-of-date picture of the computer waiting time. This sensing coverage in conditions where data transmission from sensor
local view is periodically updated with a little amount of connection nodes is regular or erratic. We show through simulation experiments
overhead. As longer as all these local estimations of the service waiting comparing the proposed algorithm to the most widely used state-of-
time are accurate is contained in assumptions, as we clearly show, the the-art clustering approaches that it is capable of improving both range
network is quite resilient. Last but not least, we demonstrate in simu- of communications services and by maintaining more network devices
lations how straightforward and reliable LSQ policies work admirably operational for extended periods of time at low computation cost,
and greatly exceed current low-communication strategies when given network coverage may be increased [61].
an identical communications allowance. There are many uses for resource-constrained networks in daily life.
The Blind GB-PANDAS method, which would be fully blind to job To increase the longevity of these networks, a consistent load balanc-
entry patterns and delivery rates, is provided by (Ali Yekkehkhany ing approach is a difficult problem. The method takes into account
and Rakesh Nagi) [56]. Routing in Blind GB-PANDAS follows an elements like distance, remaining energy, and hop count in order to bal-
exploration-exploitation strategy. We demonstrate that, when service ance energy use among network participants and lengthen the network
durations of various task kinds on numerous servers are distributed lifetime. In simulations, In terms of parameters like throughput, node
randomly and in an unknown manner, Blind GB-PANDAS provides the lifespan, packet loss ratio (PLR), communication costs, latency, and
maximum throughput. Blind GB-PANDAS wants to send arriving jobs to computing costs, our suggested system outperforms the existing so-
the servers that has the lowest quantity of balanced tasks; however, this lutions. Additionally, our suggested method protects individual nodes
is not possible since service charges fluctuate. As a consequence, it is from existing schemes in the operational environment and extends the
more difficult to perform the throughput optimization evaluation that it lifespan of WSNs [62].
would be in the scenario when response rates are known. Our thorough A sort of self-system with constrained power and transmission
study demonstrates that, in terms of average time to complete tasks sources is wireless sensor networks (WSN). The route establishment
under heavy workloads, Blind GB-PANDAS greatly surpasses preceding is optimized using a powerful heuristic updating technique built on
techniques. a greedy predicted energy cost measure. Finally, EBAR uses a power
In order to facilitate the subsequent development of enhanced up- opportunistic broadcasting strategy to lessen the power consumption
dates that can scale to levels in society, Hang Nguyen et al. [57] address resulting from the control overhead. According to the findings of this
problems with the creation and management of a Big Data Stream thorough investigation, EBAR offers a considerable improvement over
Publish Subscribe (BAD Pub/Sub) systems. The goal of the proposed the most cutting-edge methods, including EEABR, Sensor Ant, and
BAD Pub/Sub system is to absorb enormous volumes of data from IACO [63].
several authors and organizations and to deliver enhanced, customized Despite the fact that RPL significantly addresses IoT network needs,
alerts to end customers (subscribers) who exhibit interested in such there are still some unresolved issues because it was not primarily
data objects via customized routes. developed for Iot devices. First, a description of the CAOF, which
Doaa Medhat et al. [58] offer adaptable and cost-conscious task calculates the score while considering the node’s context into account.
scheduling system using Map Reduce that can consistently connect By slowly moving away from a highly placed number and approaching
information from several, international sites. They do this by spreading the actual rank amount, CAOF also prevents the thundering herd ef-
the multilingual finding tasks on accessible computers depending on fect. Second, we introduce the context-aware routing metric (CARF), a

12
K. Raghavendar, I. Batra and A. Malik Decision Analytics Journal 7 (2023) 100200

brand-new routing metric that recursively analyses the use of electricity systems. These cloud-based solutions provide centralized platforms for
and queues of parent chains as they travel towards to the root while storing and analysing data, improving data security, and simplifying
minimizing the legacy of parents who came before as they travelled fur- device integration. Cloud-based solutions have several benefits for IoT-
ther along the way. When compared to the RPL standard specification, based home automation systems, including improved data security,
evaluation results reveal a reduction in packet loss and an improvement centralized data management, and the ability to analyse data to identify
in network longevity [64]. performance improvement opportunities. Additionally, these solutions
make it easier to integrate different devices and systems, simplifying
7.3. To determine the data skew load balancing optimization system effi- the process for homeowners. Overall, cloud-based solutions can help
ciency for different scenarios such as consumption rate address the challenges posed by IoT-based home automation systems
and improve their overall performance.
To successfully handle the data skew issue, Xiaoke Zhu et al. [65] Almurisi and Tadisetty [70] stated that owing to the complexity
develop DLB is a load balancing method based on deep learning. The of WSN and the varied requirements of IoT applications, there were
fundamental idea behind DLB is to substitute deep learning models, substantial issues that needed to be resolved. A cloud-based virtual-
which can evenly balance different workload and data distributions ization environment for IoT-based WSN was suggested as a result of
among servers, for the Using hashing algorithm in task scheduling the research, and it offered a potential way to deal with the problems
strategies. Results of experiments utilizing both artificial and actual that IoT-based WSNs have with resource management, scalability, de-
data sets of data demonstrate that our DLB, developed using Cloud pendability, and security. The suggested environment can be a useful
services and deployed in a real-world cloud environment, is capable of resource for academics and professionals involved in the IoT and cloud
offering more balanced mappings than conventional hashing function- computing fields.
based task scheduling approaches, particularly when the workloads is The research by Al Masarweh et al. [71] addressed on the difficulties
significantly skewed. in data management in a complex IoT environment that uses both
Zahra Movahedi et al. [66] address the job scheduling problem cloud and fog computing. The research revealed a sophisticated broker
with the goal of enhancing time and energy usage as two QoS factors management system that made use of a number of algorithms to
in a fog environment. In order to deliver the best solutions, we first enhance data transfer, decrease latency, and boost system performance.
present a fog-based architecture for managing scheduling algorithm The performance of the suggested system was assessed by the authors
using simulation-based tests, and its effectiveness was contrasted with
requirements. Second, research creates an Integer Linear Programming
that of existing methods. The outcomes demonstrated the suggested
(ILP) optimization model that takes into consideration both the energy
system’s potential to enhance the performance of IoT systems in compli-
consumption of the fog and the passage of time for the task schedul-
cated contexts by outperforming existing techniques in terms of latency
ing problem. Last but not least, we propose an enhanced approach
reduction and energy economy.
termed the Opposition-based Chaotic Whale Optimization Algorithm
The research by G Soniya Priyatharsini et al. [72] addressed the
(OppoCWOA) to enhance the performance of the original WOA in
security challenges associated with cloud-based IoT systems. The re-
solving the simulated work scheduling problem. Detailed simulations
search suggested a self-secured paradigm to protect data sent between
and comparisons with the original WOA and many well-known meta-
IoT devices and the cloud using encryption, digital signatures, and
heuristic algorithms, including Particle Swarm Optimization (PSO),
access control techniques. The performance of the suggested model
Artificial Bee Colony (ABC), and Genetic Algorithm, demonstrate the
was assessed by the authors using simulation-based tests, and it was
efficacy of the proposed OppoCWOA (GA).
contrasted with current methods. The findings demonstrated that the
This method of task scheduling has an essential impact on the
suggested architecture offered an extremely high level of security and
computer’s clustering efficiency. This study focuses on the benefits and
decreased the possibility of data breaches in cloud-based IoT sys-
disadvantages of popular load balancing techniques while illustrating
tems. Overall, the research made a significant addition to the field of
them. In addition, a balancing method created using the load forecast is
cloud-based IoT security and showed how the suggested self-secured
provided by (Lijie Yan and Xudong Liu). This method aids in determin- architecture may improve the security of such systems.
ing the appropriate smoothness coefficients using the server nodes’ load The research by Christos L. Stergiou and Kostas E. Psannis [73]
time series for the current phrases and realistic data to help forecasting addressed the challenges of managing and analysing big data generated
utilizing the node’s load value at the following instant as a result of the by industrial IoT systems. In order to manage and analyse the data
dynamic exponential smoothing model [67]. produced by IoT devices, the research presents a digital twin intelligent
To discover food for the building of load balance models, bat system that made use of cloud computing, big data analytics, and
behaviour uses echolocation. Virtual computers (VMs) that are utilized virtual reality. The performance of the suggested system was assessed
to update the migrating process for the essential information by the bats by the authors using simulation-based tests, and its effectiveness was
are categorized using the suggested method’s Naive-Bayes algorithm. contrasted with that of existing methods. The outcomes demonstrated
As long as their priority in the task queue was lowest, the migrated that the suggested approach offered a more effective and efficient
jobs distributed strongly loaded virtual machine instances to lightly method of managing and analysing large amounts of data in industrial
loaded virtual machine instances. It is revealed that activities with the IoT systems. The study proved the potential of the suggested digital
highest priority should be prioritized above reducing services in the VM twin intelligent system to improve the performance of industrial IoT
as workloads migrate from severely loaded virtual machines to lighter systems and made a significant contribution to the field of IoT-based
loaded virtual machine instances and should not be scheduled to the big data management.
same VM. Traditional methods like Dynamic load balance (DLB) and The research paper by B. Raviprasad et al. [74] addressed the accu-
round robin (RR) were compared to the experimental findings of the racy challenges associated with IoT sensor data. In order to assess the
loading balance bat algorithm (LBBA) [68]. accuracy of sensor data sent to the cloud, the research developed a deep
The main problem with IoT-based home automation systems is the learning-based method. The performance of the suggested methodology
management and organization of the vast amount of data generated by was assessed by the authors using simulation-based tests, and it was
these systems. Additionally, data security is a major concern, as the sen- then contrasted with current approaches. The findings demonstrated
sitive information of homeowners may be vulnerable to cyber threats. that the suggested method offered a more precise method of assessing
Furthermore, the integration of different IoT devices and systems can the veracity of sensor data in cloud-based IoT systems. Overall, the
be complex, which may lead to compatibility issues. The solution research made a significant addition to the field of IoT sensor data
proposed by Sangaraju et al. [69] is cloud-based solutions that can accuracy and showed how the suggested deep learning-based technique
store and manage the data generated by IoT-based home automation may improve sensor data quality in IoT settings that are cloud-based.

13
K. Raghavendar, I. Batra and A. Malik Decision Analytics Journal 7 (2023) 100200

Fig. 9. Showing the overall hybrid cloud set up for optimization rate analysis.

Fig. 10. Showing the overall hybrid cloud set up for skew rate analysis.

The research by N. Sai Lohitha and M. Pounambal [75] proposed a second part, which is the hardware analysis, the node-mcu network
novel approach to improve the efficiency of data processing in cloud- cloud is created as the hybrid cloud and data modelling is carried out
based IoT environments. The research integrated publish/subscribe for doing optimization rate, skew rate and consumption rate analysis
and push-pull strategies to address the issue of processing massive as shown in Figs. 9–11.
amounts of real-time data in IoT applications. The authors tested the
performance of the suggested methodology and contrasted it with cur- 8. Conclusion
rent approaches. The findings demonstrated that the suggested method
offered a more effective and efficient means of processing real-time For the innovative paradigm to work effectively and efficiently,
data in IoT systems based on the cloud. Overall, the study made a resource allocation is a crucial feature of the IoT ecosystem that is
significant addition to the field of IoT data processing and showed how cloud-based. The resource allocation approaches for IoT settings that
the integrated strategy that was suggested might potentially increase are SLA-aware, Context-aware, QoS-aware, Power-aware, and Cost-
data processing efficiency in cloud-based IoT systems. aware were examined and categorized into numerous groups in this
article. The author looked closely at each of these strategies to deter-
7.4. Flow diagram mine their flaws and possible improvements. The resource allocation
parameters that have been discovered to be enhanced are also provided
The system has been created so that the node data is used as input and described. The study includes a graphics that illustrates how much
for our suggested hybrid cloud, which will decide the selection based certain metrics have improved. Finally, the study’s problems and future
on the proposed AI model. The system is divided into simulation envi- directions were examined.
ronment and hardware environment as define in Fig. 8. The simulation The system is set up in such a way that the node data is fed into
analysis does the load balancing and after data aggregation it does the our suggested hybrid cloud, which makes the decision based on the
optimization rate, skew rate and consumption rate analysis. On the AI model. The system is divided into two environments: simulation

14
K. Raghavendar, I. Batra and A. Malik Decision Analytics Journal 7 (2023) 100200

Fig. 11. Showing the overall hybrid cloud set up for consumption rate analysis.

and hardware. After data collection, the simulation analysis performs [10] S. Li, N. Zhang, S. Lin, L. Kong, A. Katangur, M.K. Khan, Joint admission control
load balancing, optimization rate, skew rate, and consumption rate and resource allocation in edge computing for internet of things, IEEE Netw. 32
(2018) 72–79.
analysis. The node-mcu network cloud is established as a hybrid cloud
[11] C. Qiu, H. Yao, C. Jiang, S. Guo, F. Xu, Cloud computing assisted blockchainen-
in the second stage, which is the hardware analysis, and data modelling abled internet of things. IEEE Trans. Cloud Comput. (nd), http://dx.doi.org/10.
is performed for optimization rate, skew rate, and consumption rate 1109/TCC.2019.2930259.
analysis. [12] Yi Zhang, Heping Liu, Zhigang Zhang, Yong Luo, Qiang Guo, Shuiping Liao,
Cloud computing-based real-time global optimization of battery aging and
energy consumption for plug-in hybrid electric vehicles, J. Power Sources
Declaration of competing interest (ISSN: 0378-7753) 479 (2020) 229069, http://dx.doi.org/10.1016/j.jpowsour.
2020.229069.
The authors declare that they have no known competing finan- [13] J. Huang, Y. Yin, Q. Duan, H. Yan, A game-theoretic analysis on context-aware
resource allocation for device-to-device communications in cloud-centric internet
cial interests or personal relationships that could have appeared to
of things, in: 2015 3rd International Conference on Future Internet of Things and
influence the work reported in this paper. Cloud (FiCloud), 2015, pp. 80–86.
[14] R. Want, An introduction to RFID technology, IEEE Perv. Comput. 5 (2006)
Data availability 25–33;
S.F. Abedin, M.G.R. Alam, S. Il, C.S.H. Moon, An optimal resource allocation
scheme for fog based P2P IoT network. Efrom ‘T’ ‘Z’ ‘IJN’ ‘I’m ‘ijS’, 2015, pp.
Data will be made available on request. 395–397.
[15] B. Manate, T.F. Fortis, V. Negru, Optimizing cloud resources allocation for an
References internet of things architecture, Scalable Comput. 15 (2015) 345–355.
[16] X. Chen, L. Chen, M. Zeng, X. Zhang, D. Yang, Downlink resource allocation for
deviceto-device communication underlying cellular networks, in: 2012 IEEE 23rd
[1] U.K. Jena, P.K. Das, M.R. Kabat, Hybridization of meta-heuristic algorithm for
International Symposium on Personal Indoor and Mobile Radio Communications,
load balancing in cloud computing environment, J. King Saud Univ. Comput. Inf.
PIMRC, 2012, pp. 232–237.
Sci. (ISSN: 1319-1578) (2020) http://dx.doi.org/10.1016/j.jksuci.2020.01.012,
[17] G. Colistra, V. Pilloni, L. Atzori, Task allocation in group of nodes in the IoT: A
https://www.sciencedirect.com/science/article/pii/S1319157819309267.
consensus approach, in: 2014 IEEE International Conference on Communications,
[2] N. Koshizuka, K. Sakamura, Ubiquitous ID: Standards for ubiquitous computing
ICC, 2014, pp. 3848–3853.
and the internet of things, IEEE Perv. Comput. 9 (2010) 37, Sensors.
[18] https://www.sciencedirect.com/science/article/pii/S0378775320313641.
[3] S.A. Ali, M. Affan, M. Alam, A study of efficient energy management techniques
[19] S. Khan, S.A. Ali, N. Hasan, K.A. Shakil, M. Alam, Big data scientific workflows
for cloud computing environment, in: 2019 9th International Conference on
in the cloud: Challenges and future prospects, in: Studies in Big Data, Springer
Cloud Computing, 2019, pp. 13–18, http://dx.doi.org/10.1109/CONFLUENCE.
International Publishing, 2018, pp. 1–28, http://dx.doi.org/10.1007/978-3-030-
2019.8776977.
03359-0_1.
[4] W. Stallings, The internet of things: network and security architecture, Internet [20] A. Bassi, M. Bauer, M. Fiedler, R.V. Kranenburg, Enabling Things to Talk,
Prot. J. 18 (4) (2015) 381–385. Springer, New York, 2013.
[5] Arabinda Pradhan, Sukant Kishoro Bisoy, A novel load balancing technique [21] P.G.V. Naranjo, Z. Pooranian, M. Shojafar, M. Conti, R. Buyya, FOCAN: a fog-
for cloud computing platform based on PSO, J. King Saud Univ. Comput. Inf. supported smart city network architecture for management of applications in the
Sci. (ISSN: 1319-1578) (2020) http://dx.doi.org/10.1016/j.jksuci.2020.10.016, internet of everything environments, J. Parallel Distrib. Comput. (2017) arXiv
https://www.sciencedirect.com/science/article/pii/S1319157820304961. preprint.
[6] S.A. Ali, M. Alam, A relative study of task scheduling algorithms in cloud [22] E. Baccelli, O. Hahm, M. Günes, M. Wählisch, T.C. Schmidt, RIOT OS: Towards
computing environment, in: 2016 2nd International Conference on Contemporary an OS for the internet of things, in: Proc. IEEE Conf. INFOCOM WKSHPS, 2013,
Computing and Informatics (IC3I), 2016, pp. 105–111, http://dx.doi.org/10. pp. 79–80.
1109/IC3I.2016.7917943. [23] L. Yang, S.H. Yang, L. Plotnick, How the internet of things technology enhances
[7] A. Malik, H. Om, Cloud computing and internet of things integration: archi- emergency response operations, Technol. Forecast. Soc. Change 80 (2013)
tecture, applications, issues, and challenges, in: Sustainable Cloud and Energy 1854–1867.
Services, Springer, 2018, pp. 1–24. [24] A. Singh, Y. Viniotis, An SLA-based resource allocation for IoT applications in
[8] B. Pourghebleh, N.J. Navimipour, Data aggregation mechanisms in the internet cloud environments, in: Cloudification of the Internet of Things (CIoT), 2016,
of things: a systematic review of the literature and recommendations for future pp. 1–6.
research, J. Netw. Comput. Appl. 97 (2017) 23–34. [25] S. 4Shorgin, K.E. Samouylov, Y.V. Gaidamaka, A. Chukarin, I.A. Buturlin, V.
[9] Y. Choi, Y. Lim, Optimization approach for resource allocation on cloud com- Begishev, Modeling radio resource allocation scheme with fixed transmission
puting for IoT, J. Distrib. Sens. Netw. (2016) http://dx.doi.org/10.1155/2016/ zones for multiservice M2 M communications in wireless IoT infrastructure,
3479247. ACIIDS 2 (2015) 473–483.

15
K. Raghavendar, I. Batra and A. Malik Decision Analytics Journal 7 (2023) 100200

[26] Ankit Shah, Saksham Mamta Padole, Resource aware block rearrangement [50] P. Mell, T. Grance, The NIST definition of cloud computing, 2011;
algorithm for load balancing in hadoop, Procedia Comput. Sci. (ISSN: 18770509) M. Muntjir, M. Rahul, H.A. Alhumyani, An analysis of Internet of Things (IoT):
167 (2020) 47–56, http://dx.doi.org/10.1016/j.procs.2020.03.181, https://www. novel architectures, modern applications, security aspects and future scope with
sciencedirect.com/science/article/pii/S1877050920306463. latest case studies, Int. J. Eng. Res. Technol. 6 (6) (2017) 422–447.
[27] Sambit Kumar Mishra, Bibhudatta Sahoo, Priti Paramita Parida, Load balancing [51] J. Lin, W. Yu, N. Zhang, X. Yang, H. Zhang, W. Zhao, A survey on inter-
in cloud computing: A big picture, J. King Saud Univ. Comput. Inf. Sci. (ISSN: net of things: architecture, enabling technologies, security and privacy, and
1319-1578) 32 (2) (2020) 149–158, http://dx.doi.org/10.1016/j.jksuci.2018.01. applications, IEEE Internet Things J. 4 (2017) 1125–1142.
003, https://www.sciencedirect.com/science/article/pii/S1319157817303361. [52] L. Atzori, A. Iera, G. Morabito, The internet of things: A survey, Comput. Netw.
[28] Mohammad Haris, Swaleha Zubair, Mantaray modified multi-objective harris 54 (15) (2010) 2787–2805;
hawk optimization algorithm expedites optimal load balancing in cloud com- E. Baccarelli, P.G.V. Naranjo, M. Scarpiniti, M. Shojafar, J.H. Abawajy, Fog
puting, J. King Saud Univ. Comput. Inf. Sci. (ISSN: 1319-1578) (2021) http:// of everything: energy-efficient networked computing architectures, research
dx.doi.org/10.1016/j.jksuci.2021.12.003, https://www.sciencedirect.com/science challenges, and a case study, IEEE Access 5 (2017) 9882–9910.
/article/pii/S1319157821003372. [53] A.A. Alsaffar, H.P. Pham, C.S. Hong, E.N. Huh, M. Aazam, An architecture of IoT
[29] Q. Zhang, L. Cheng, R. Boutaba, Cloud computing: state-of-the-art and research service delegation and resource allocation based on collaboration between fog
challenges, J. Internet Serv. Appl. 1 (1) (2010) 7–18. and cloud computing, Mob. Inf. Syst. (2016) http://dx.doi.org/10.1155/2016/
[30] P. Mcdermott-Wells, What is bluetooth? IEEE Potentials 23 (2004) 33–35. 6123234.
[31] S. Horrow, A. Sardana, Identity management framework for cloud based internet [54] V. Angelakis, I. Avgouleas, N. Pappas, E. Fitzgerald, D. Yuan, Allocation of
of things, in: Proceedings of the First International Conference on Security of heterogeneous resources of an IoT device to flexible services, IEEE Internet
Internet of Things, 2012, pp. 200–203. Things J. 3 (2016) 691–700.
[32] F. Mattern, C. Floerkemeier, From Active Data Management To Event-Based [55] Shay Vargaftik, Isaac Keslassy, Ariel Orda, LSQ: Load balancing in large-scale
Systems and more, Springer, New York, 2010, From the internet of computers heterogeneous systems with multiple dispatchers, 2020.
to the Internet of Things.
[56] Ali Yekkehkhany, Rakesh Nagi, Blind GB-PANDAS: A blind throughput-optimal
[33] E. Ferro, F. Potorti, Bluetooth and Wi-Fi wireless protocols: A survey and a
load balancing algorithm for affinity scheduling, 2020.
comparison, IEEE Wirel. Commun. 12 (2005) 12–26;
[57] Hang Nguyen, Md Yusuf Sarwar Uddin, Nalini Venkatasubramanian, Multistage
M. Gigli, S. Koo, Internet of things: services and applications categorization, Adv.
adaptive load balancing for big active data publish subscribe systems, in: DEBS
Internet Things (2011).
’19: The 13th ACM International Conference on Distributed and Event-Based
[34] G.V. Crosby, F. Vafa, Wireless sensor networks and LTE-a network convergence,
Systems (DEBS ’19), June 24–28, 2019, Darmstadt, 2019.
in: Proceedings of the IEEE 38th Conference on Local Computer Networks, LCN,
2013, pp. 731–734. [58] Doaa Medhat, Ahmed H. Yousef, Cherif Salama, Cost-aware load balancing
for multilingual record linkage using MapReduce, Ain Shams Eng. J. (ISSN:
[35] L. Wang, G.V. Laszewski, A. Younge, X. He, M. Kunze, J. Tao, C. Fu, Cloud
2090-4479) 11 (2) (2020) 419–433, http://dx.doi.org/10.1016/j.asej.2019.08.
computing: a perspective study, New Gener. Comput. 28 (2) (2010) 137–146;
009.
R. Want, Near field communication, IEEE Perv. Comput. 10 (2011) 4–7.
[36] V. Angelakis, I. Avgouleas, N. Pappas, D. Yuan, Flexible allocation of hetero- [59] R. Gulbaz, A.B. Siddiqui, N. Anjum, A.A. Alotaibi, T. Althobaiti, N. Ramzan,
geneous resources to services on an IoT device, in: 2015 IEEE Conference on Balancer genetic algorithm—A novel task scheduling optimization approach
Computer Communications Workshops (INFOCOM WKSHPS), 2015, pp. 99–100. in cloud computing, Appl. Sci. 11 (2021) 6244, http://dx.doi.org/10.3390/
app11146244.
[37] X.J. Xing, J.L. Wang, M.D. Li, Services and key technologies of the internet of
things, ZTE Commun. 2 (2010) 11; [60] Ernando Batista, Gustavo Figueiredo, Maycon Peixoto, Martin Serrano, Cassio
Z. Yan, P. Zhang, A.V. Vasilakos, A survey on trust management for Internet of Prazeres, [IEEE 2018 IEEE international conference on Internet of Things
Things, J. Netw. Comput. Appl. 42 (2014) 120–134. (iThings) and IEEE green computing and communications (GreenCom) and IEEE
[38] H.Y. Lan, H.T. Song, H.B. Liu, G.Y. Zhang, Heterogeneous oriented resource cyber, physical and social computing (CPSCom) and IEEE smart data (Smart-
allocation method in internet of things, Appl. Mech. Mater. 427 (2013) Data) halifax, NS, Canada (2018.7.30-2018.8.3)], in: 2018 IEEE International
2791–2794. Conference on Internet of Things (IThings) and IEEE Green Computing and
Communications (GreenCom) and IEEE Cyber, Physical and Social Computing
[39] S. Khan, K.A. Shakil, S. Ali, M. Alam, On designing a generic framework for big
(CPSCom) and IEEE Smart Data (SmartData) Load Balancing in the Fog of
data as-a-service, in: 2018 1st International Conference on Advanced Research
Things Platforms Through Software-Defined Networking, 2018, pp. 1785–1791,
in Engineering Sciences, ARES, 2018, pp. 1–5, http://dx.doi.org/10.1109/ARESX.
http://dx.doi.org/10.1109/Cybermatics_2018.2018.00297.
2018.8723269.
[61] Shesha Sreenivasamurthy, Katia Obraczka, [IEEE 2018 IEEE 26th international
[40] N. Abuzainab, W. Saad, C.S. Hong, H.V. Poor, Cognitive hierarchy theory for
symposium on modeling, analysis, and simulation of computer and telecommu-
distributed resource allocation in the internet of things, 2017, arXiv preprint;
nication systems (MASCOTS) milwaukee, WI, USA (2018.9.25-2018.9.28)], in:
A. Al-Fuqaha, M. Guizani, M. Mohammadi, M. Aledhari, M. Ayyash, Internet
2018 IEEE 26th International Symposium on Modeling, Analysis, and Simulation
of things: a survey on enabling technologies, protocols, and applications, IEEE
of Computer and Telecommunication Systems (MASCOTS) Clustering for Load
Commun. Surv. Tutor. 17 (2015) 2347–2376.
Balancing and Energy Efficiency in IoT Applications, 2018, pp. 319–332, http:
[41] G. Colistra, V. Pilloni, L. Atzori, The problem of task allocation in the internet
//dx.doi.org/10.1109/MASCOTS.2018.00038.
of things and the consensus-based approach, Comput. Netw. 73 (2014) 98–111.
[62] Muhammad Adil, Rahim Khan, Mohammed Amin Almaiah, Muhammad Bin-
[42] B. Varghese, R. Buyya, Next generation cloud computing: New trends and
sawad, Jehad Al, An efficient load balancing scheme of energy gauge nodes to
research directions, Future Gener. Comput. Syst. 79 (2018) 849–861.
maximize the lifespan of constraint oriented networks, 2020, Received August
[43] F.C. Delicato, P.F. Pires, T. Batista, The resource management challenge in IoT,
4, 2020, accepted August 8, 2020, date of publication August 11, 2020, date of
in: Resource Management for Internet of Things, Springer, 2017, pp. 7–18.
current version August 24, 2020.
[44] G. Dittmann, J. Jelitto, A blockchain proxy for lightweight IoT devices, in: 2019
[63] Xinlu Li, Brian Keegan, IEEE Senior Member, Fredrick Mtenzi, Thomas Weise,
Crypto Valley Conference on Blockchain Technology, CVCBT, 2019, pp. 82–85,
Ming Tan, Energy-efficient load balancing ant based routing algorithm for
http://dx.doi.org/10.1109/CVCBT.2019.00015.
wireless sensor networks, 2019, Received June 28, 2019, accepted August 6,
[45] Q. Cao, T. Abdelzaher, J. Stankovic, T. He, The liteos operating system:
2019, date of publication August 12, 2019, date of current version August 28,
Towards unixlike abstractions for wireless sensor networks, in: Proceedings of the
2019.
International Conference on Information Processing in Sensor Networks, 2008,
pp. 233–244. [64] Seyedreza Taghizadeh, Hossein Bobarshad, Halima Elbiaze, CLRPL: Context-
aware and load balancing RPL for iot networks under heavy and highly dynamic
[46] A. Singh, Y. Viniotis, Resource allocation for IoT applications in cloud envi-
load, 2018, Received January 29, 2018, accepted February 27, 2018, date of
ronments, in: 2017 International Conference on Computing, Networking and
publication April 9, 2018, date of current version May 16, 2018.
Communications, ICNC, 2017, pp. 719–723.
[47] F.A. Alaba, M. Othman, I.A.T. Hashem, F. Alotaibi, Internet of things security: [65] Xiaoke Zhu, Qi Zhang, Taining Cheng, Ling Liu, WeiZhou, Jing He, DLB: Deep
a survey, J. Netw. Comput. Appl. 88 (2017) 10–28. learning based load balancing, 2021.
[48] G. Marques, N. Garcia, N. Pombo, A survey on IoT: architectures, elements, [66] Z. Movahedi, B. Defude, A.m. Hosseininia, An efficient population-based multi-
applications, qos, platforms and security concepts, in: Advances in Mobile Cloud objective task scheduling approach in fog computing systems, J. Cloud. Comput.
Computing and Big Data in the 5G Era, Springer, 2017, pp. 115–130. 10 (2021) 53, http://dx.doi.org/10.1186/s13677-021-00264-4.
[49] H.F. Atlam, A. Alenezi, R.J. Walters, G.B. Wills, J. Daniel, Developing an [67] Lijie Yan, Xudong Liu, The predicted load balancing algorithm based on the
adaptive riskbased access control model for the internet of things, in: 2017 dynamic exponential smoothing, Open Phys. 18 (1) (2020) 439–447, http://dx.
IEEE International Conference on Internet of Things (IThings) and IEEE Green doi.org/10.1515/phys-2020-0108.
Computing and Communications (GreenCom) and IEEE Cyber, Physical and [68] Laheeb Mohammed Ibrahim1, Ibrahim Ahmed Saleh, A solution of loading
Social Computing (CPSCom) and IEEE Smart Data (SmartData), 2017, pp. balance in cloud computing using optimization of bat swarm algorithm, J. Eng.
655–661. Sci. Technol. 15 (3) (2020) 2062–2076.

16
K. Raghavendar, I. Batra and A. Malik Decision Analytics Journal 7 (2023) 100200

[69] Hrushikesava Raju Sangaraju, M. Rao, N. Sudheer, P. Kavitharani, IOT based [73] Christos L. Stergiou, Kostas E. Psannis, Digital twin intelligent system for
home automation system with cloud organizing, Int. J. Eng. Technol. 7 (412) industrial IoT-based big data management and analysis in cloud, Virt. Real. Intell.
(2018) http://dx.doi.org/10.14419/ijet.v7i2.32.15728. Hardw. (ISSN: 2096-5796) 4 (4) (2022) 279–291, http://dx.doi.org/10.1016/j.
[70] N. Almurisi, S. Tadisetty, Cloud-based virtualization environment for IoT-based vrih.2022.05.003.
WSN: solutions, approaches and challenges, J. Ambient Intell. Hum. Comput. 13 [74] B. Raviprasad, Chinnem Rama Mohan, G. Naga Rama Devi, R. Pugalenthi,
(2022) 4681–4703, http://dx.doi.org/10.1007/s12652-021-03515-z. L.C. Manikandan, Sivakumar Ponnusamy, Accuracy determination using deep
[71] M. Al Masarweh, T. Alwada’n, W. Afandi, Fog computing, cloud computing and learning technique in cloud-based IoT sensor environment, Measurement: Sensors
IoT environment: Advanced broker management system, J. Sens. Actuat. Netw. (ISSN: 2665-9174) 24 (2022) 100459, http://dx.doi.org/10.1016/j.measen.2022.
11 (2022) 84, http://dx.doi.org/10.3390/jsan11040084. 100459.
[72] G. Soniya Priyatharsini, A. Jyothi Babu, M. Gnana Kiran, P.J. Sathish Kumar, [75] N. Sai Lohitha, M. Pounambal, Integrated publish/subscribe and push-pull
C. Nelson Kennedy Babu, Aleem Ali, Self-secured model for cloud based IOT method for cloud based IoT framework for real time data processing, Measure-
systems, Measurement: Sensors (ISSN: 2665-9174) 24 (2022) 100490, http: ment: Sensors (ISSN: 2665-9174) (2023) 100699, http://dx.doi.org/10.1016/j.
//dx.doi.org/10.1016/j.measen.2022.100490. measen.2023.100699.

17