EDGE COMPUTING IN IOT DEVICES

A seminar submitted in partial fulfillment of the

requirement for the award of degree of

MASTER OF COMPUTER APPLICATIONS

of
Visvesvaraya Technological University

By

SAHANA S
4MH22MC080

DEPARTMENT OF COMPUTER APPLICATIONS

MAHARAJA INSTITUTE OF TECHNOLOGY MYSORE
Belawadi, Srirangapatna Tq, Mandya – 571477
August 2023
 MAHARAJA INSTITUTE OF TECHNOLOGY MYSORE

Department of Computer Applications

Belawadi, Srirangapatna Tq, Mandya – 571477

CERTIFICATE

This is to certify that SAHANA S bearing 4MH22MC080 has completed Seminar

– 22MCA29 presentation entitled “Edge Computing in IOT Devices” as a partial

fulfillment for the award of Master of Computer Applications degree, during the

academic year 2023.

Head of the Department

 DECLARATION

I, SAHANA S, student of 2nd semester MCA, Maharaja Institute of Technology,

Mysore, bearing USN 4MH22MC080 hereby declare that the seminar entitled

Edge Computing in IOT Devices has been carried out by me and submitted the

report in partial fulfillment of the requirements for the award of the Degree of

Master of Computer Applications by the Vivesvaraya Technological University

during the academic year 2023. This report has not been submitted to any other

Organization/University for any award of degree or certificate.

Name: SAHANA S

Signature:
 ACKNOWLEDGEMENT

The seminar report on “Edge Computing in IoT Devices” is the outcome of

guidance, moral support and devotion bestowed on me throughout my work. For
this I acknowledge and express my profound sense of gratitude and thanks to
everybody who have been a source of inspiration during the seminar preparation.

First and foremost, I offer my sincere phrases of thanks with innate humility to our
Trustee Dr. T VASUDEV & Principal Dr. B G NARESH KUMAR, who has
been a constant source of support and encouragement. I feel deeply indebted to our
HOD Prof. MANJUNATH B for the right help provided from the time of
inception till date.

I would take this opportunity to acknowledge my mentor Ms. DEEPIKA D, who

not only stood by me as a source of inspiration, but also dedicated their time for me
to present the seminar on time.

I would also thank Prof. AMOS R for helping me in completion of this seminar on
time with his knowledge and support.

I would be failing in endeavor, if I do not thank our teaching and non-teaching

staff of Department of MCA, who helped us directly or indirectly in every aspect
of my seminar work.

Lastly, I would vote my thanks to my parents, friends and well-wishers who have
helped me throughout this work.

With Deep Reverence,

SAHANA S
 ABSTRACT

The number of smart devices connected to the internet is increasing manifold on a

daily basis, which in turn generates Big Data. This data requires a robust and
reliable cloud storage to save the data and protect it from unauthorized access.
Further this data requires huge processing power to drive to any tangible and useful
information from it. Many business processes are eyeing technologies to improve
efficiency, performance and reduce cost of operation of the IoT devices. The
leading ways now days is by using Edge Computing, i.e. processing data out of the
main cloud or at its edge. The new technology has the potential to deal and wipe
off the concerns of response time, bandwidth cost saving, duration of battery life
and most significantly safety and privacy of vital data of the organization.

The technologies that help this cutting-edge model to function properly. Since
Edge Computing involves a heterogeneous architecture, it requires to adapt to a
few technological recommendations for optimal performance. In this context, this
report reviews the latest hardware technology trends tied to Edge Computing
developments, and points out technical challenges implementing this innovative
computing model. In particular, we analyze how High Performance Computing
and Cloud Computing infrastructures can be efficiently organized to design an
Edge Computing-based framework able to tackle cutting-edge issues solved by
Artificial Intelligence techniques. Finally, this report presents selected real-world
applications of the Edge Computing paradigm across multiple domains affecting
our daily life, i.e., healthcare, smart city and grids, industry 4.0 and public safety.
 TABULATION

Sl No Particulars Page No

1 Introduction 1–6

 Edge Computing

 Internet of Things (IoT)

 Edge Computing in IoT Devices

 Why Edge Computing?

 Evolution of Computing System

2 Literature Review 7–9

3 Overview Of Edge Computing in IoT Devices 10 – 18

 Edge Computing Architecture

 Concepts of Edge Computing

 Working of Edge Computing with IoT

 Benefits and Challenges

4 Use Cases 19 – 25

5 Future Scope 26 – 26

6 Conclusion 27 – 27

7 Bibliography 28 – 29
 EDGE COMPUTING IN IOT DEVICES

CHAPTER 1

INTRODUCTION

Cloud computing is one of a very few technologies which altered the way we survive, learn and
study to a gigantic extent. Services such as Facebook, Twitter and Google which uses Software
as a Service (SaaS), are being extensively used in our daily life. Internet of Things (IoT) was first
tested in management industry for supply chain. But sooner than later it was adapted in other
fields as well, such as, healthcare, smart cities, smart parking, smart water system, transportation
and many others. IoT is most simple terms in “making computer/chip/smart device sense
information from its surroundings without any involvement of humans”. With the introduction of
Edge computing, Cloud processes has decreased to a greater extent. In 2019, more than 45% of
the data generated by IoT smart devices was processed and analyzed at the edge of the network.

EDGE COMPUTING

The data is produced i.e., “at the edge” of the network. This helps computations to be performed
near to the source where the data has been generated. It includes two streams, i.e., Upstream and
the downstream. Upstream means that the data is travelling from the data source (IoT device) to
the Cloud. This stream works on the behalf of the IoT services. Downstream means that the data
is travelling from the cloud to the IoT devices. This stream works on behalf of the cloud services.
Both these streams are equally important for the working of the Edge as well as cloud paradigm.
“Edge” can be defined as any network resource with computing power in the path between the
IoT device and the cloud. Data is stored locally instead of the cloud. Edge computing is the
technology that enables computation to be performed right at the place For instance, user’s smart
phone acts as an edge between smart wearable devices like smart watches, fitness trackers and
the cloud. A router/gateway installed in a smart home environment is the edge between the
various devices like smart AC, smart lights and the cloud. A cloudlet denotes an edge between a
cell phone apparatus and the cloud. The basic difference between Edge Computing and Fog
Computing is that, Fog Computing looks after the infrastructure of the network while Edge

Department of MCA Page 1

 EDGE COMPUTING IN IOT DEVICES

computing concentrates on the “Things”. It is better to infer that in the coming future,
future Edge
Computing can have a strong impact on the sustainable society as of Cloud Computing.

Edge computing makes the cloud computing avail and benefits closer to the end-user,
end and it's
illustrated by speedy processing and application response time. Surve
Surveillance,
illance, virtual reality, and
real-time
time traffic monitoring are just a few examples of internet
internet-enabled
enabled applications that demand
quick processing and reaction times. The applications indicated because processing in the Edge
network to be enabled. Cloudlets, Fog computing, and Mobile Edge computing are three edge
computing technologies that can be used to solve cloud computing difficulties.

INTERNET OF THINGS (IOT)

IoT (Internet of things) is an emerging global Internet

Internet-based
based technical architecture facilitating
facilitat the
exchange of goods and services in global supply chain networks has an impact on the security
and privacy of the involved stakeholders.

The IoT introduces a step change in individuals’ quality of life by offering a lot of new
opportunities to data access, specific services in education, security, health care or transportation
among others. On the other hand, it will be a key to increase enterprises’ productivity by offering
a widely distributed, locally intelligent network of smart devices and new sservices
ervices that can be
personalized to customer needs. The IoT brings benefits from improved management and
tracking of assets and products, it increases the amount of information data and allows the
optimization of equipment and use of resources that can be translated into costs saving.
Moreover, it offers the opportunity to create new smart interconnected devices and explore new
business models.

Department of MCA Page 2

 EDGE COMPUTING IN IOT DEVICES

IoT systems allow users to achieve deeper automation, analysis, and integration within a system.
They improve the reach of these areas and their accuracy. IoT utilizes existing and emerging
technology for sensing, networking, and robotics.

IoT exploits recent advances in software, falling hardware prices, and modern attitudes towards
technology. Its new and advanced elements bring major changes in the delivery of products,
goods, and services; and the social, economic, and political impact of those changes.

EDGE COMPUTING IN IOT DEVICES

Gadgets that are connected to the internetwork generate massive data which remains a challenge
in managing and analyzing those collected data. This paradigm can be solved by edge computing
by pushing the data from a centralized system to local edge centers that are close to the source.
These data are analyzed in relative positions with low cost and better efficiency. The edge
devices also solve low connectivity and transferring the cost of data. In the case of industrial IoT,
edge devices are self-regulated and dynamically respond to changes where the data are collected,
analyzed, and implemented robustly

WHY EDGE COMPUTING?

Currently, the entire world is going towards digitalization, and lots of data is produced in various
fields. Moreover, in most cases, this data needs to be processed in a short time to facilitate the

Department of MCA Page 3

 EDGE COMPUTING IN IOT DEVICES

present technology (real-time applications). A few years back, cloud technologies have been
introduced, gradually reducing the need for small and medium-scale companies and research
institutes to own a computer to do the computations. Nevertheless, the end-users still need to
send and receive the data to and from the location where the machine is located. In contrast,
Edge Computing is an alternative option for doing computations where the data is located, and is
especially suited for real-time applications. In particular, part of the IoT might require short
response time, private data, and Big Data, which could be challenging for the network. However,
Cloud Computing cannot handle few of these challenges. Figure 1a and 2 show the paradigm and
schematic model of Edge Computing. Edge Computing is not a direct competition to Cloud
Computing or supercomputers, but it is certainly sharing computational burden with cloud
technology and supercomputers. If the present trend continues, more robust and energy-efficient
small/embedded machines will improve the computations in the future.

The following items provide information about why Edge Computing is needed: The state-of-
the-art

Push from the Cloud Services: In general, Cloud Computing has proven to be very efficient in
terms of computation, but in some situations, there has to be an alternative solution to avoid data
transfer bottlenecks. Edge Computing is solving this problem. For example, a Boeing 787
produces 5 Gigabytes (GBs) of data every minute, and transferring this data to a satellite or the
ground is not efficient for data processing.

Push from the Internet of Things: Presently, electronic devices such as LEDs, surveillance
cameras, and air quality sensors are part of the IoT, and they produce and consume a lot of data.
In future, there will be even more electronic devices that will be connected to the IoT. It is not
feasible to process all the data in the cloud due to the bandwidth and latency. This means some
of the data need to be processed at the level of Edge devices. Moreover, privacy is a big concern
for cloud solutions. Edge Computing can minimize this concern by restricting the data within the
Edge.

Change from a Data Consumer to a Producer: A device at the Edge not only consumes the data
from the cloud, but it also produces the data and uploads the data to the cloud. Watching a
YouTube video from your mobile phone, using Facebook and Instagram, are examples where
Edge users pull the data from the cloud. At the same time, Edge devices produce the data, such

Department of MCA Page 4

 EDGE COMPUTING IN IOT DEVICES

as taking pictures or recording videos. When the Edge users try to upload this data to the cloud, it
could be a lot of data depending on the resolution. This would occupy even more bandwidth for
the uploading.

EVOLUTION OF COMPUTING SYSTEM

Computing has come a long way from where it started. At the very beginning, a computer could
only perform one task at a time. Several distributed computers had to run in parallel when
performing multiple tasks, and distributed systems were formed by connecting those computers,
which usually communicated and coordinated their actions through message exchange. Then,
personal computers and multitasking operating systems emerged and made it possible to run
multiple tasks on the same computer.
This enabled the systems’ developers to build and run an entire system within one or more
connected computers. As the price of computing power and storage fell, organizations all over
the world started using distributed systems. The breakthrough time of distributed systems came
when Internet-based companies became so large that they needed to build distributed systems
that spanned across the World as data centers. The standardization of the concept of distributed
computing led to the development of other models, including cluster computing, autonomic
computing, utility computing, and grid computing. Engineers and developers then started to
think about a way to create multiple virtual computers within the same machine. This led to the
concept of virtualization by which the same computer could act as multiple computers all
running at the same time. On top of these predecessor models, the construction of the cloud
computing model was a natural step forward from grid computing. Cloud computing leveraged
the existing models and used the virtualization concept to provide users with new services
including new features and characteristics.

That was a good idea, but it was not the best option when it came to the utilization of the
resource of the host computer. Running multiple operating systems required using additional
resources that were not needed when running one operating system. This led to the invention of
containers in which multiple programs required separate runtime when using the same host
operating system kernel. Containers act as virtual machines without the need for a separate
overhead operating system. Due to its innovative characteristics and services, the cloud
computing model has captured significant attention among individual users, academia,

Department of MCA Page 5

 EDGE COMPUTING IN IOT DEVICES

industries, and even governments. Cloud computing services are offered by providing access to a
wide range of infrastructures hosted on cloud data centers. Despite the increasing usage of cloud
computing, its inherent issues remain unresolved. These include unreliable latency, lack of
mobility support, location-awareness, as well as privacy and security issues. Other models have
emerged in an attempt to solve these problems, but edge computing is still the most promising
state-of-the-art computing system that continues to evolve.
The term ‘edge computing’ first appeared in 2002, to state that from a business perspective,
applications should be served and moved from the cloud data center to the network edge. The
term was then used in 2004 to describe a system that distributed program methods and the
corresponding data to the edge of the network in order to enhance the system’s performance.

fig 1.3

 1950-60s: It was then the first computer introduced.

 1970s: IBM introduced the Mainframe which was then Mainframe Era.
 1980s: MS Office, OS etc came into existence which was Client Server Era.
 1990s: It was the Internet-based Era where www used widely by people which was of
Google.
 2000s: It was Social-Computing Era where cloud computing started.
 Now: It is starting with Edge Computing Era.

Department of MCA Page 6

 EDGE COMPUTING IN IOT DEVICES

CHAPTER 2
LITERATURE REVIEW

This section presents a comprehensive survey on the edge computing applications and
potentialities embedded with IoT devices in diverse areas. The section also discusses edge
computing technologies solving many real-time issues with IoT devices.
Shufen Wange discusses on development of smart homes using IoT devices. These IoT devices
monitor the internal status of the home, regulates the home environment providing home safety
and proper security. As everything cannot be uploaded in the cloud especially some video data,
edge computing monopolized for processing home data through the gateway reducing the
overflow of household data, prevent the leakage of data improving the privacy of the edge node.
M. Chen et al. devised Edge and cognitive computing (ECC) that endows smart healthcare
systems when the patients are in an emergency. The edge devices assigned according to the
patient's health status. On assigned edge devices that process robustly with low latency, cognitive
computation is carried out. The cognitive data engine collects the patient's data, and the cognitive
resource engine collects the information relevant to edge, Cloud, and other network resources.
The data is sent to these allocated resources according to the risk level and requirement of the
patient.

Zhang et al. proposed video analysis in edge computing. This video analysis is named as Edge
video Analysis for Public Safety framework (EVAPS). With this edge computing framework, the
video analysis workload is distributed optimally on both the edge nodes and the cloud Centre.
The edge node prevents unnecessary data transmissions to the cloud.

L.U. Khan et al. proposed a smart city scenario using edge computing. In this system, the data is
taken from the environment. Vehicles run applications and transmit data to the edge computing
servers via roadside units (RSU) for further analysis and processing. The RSUs and edge servers
are interrelated and finally connected to the cloud. For seamless connectivity and security, RSUs
are deployed. The main objective of the proposed work is to reduce the cost of network
deployment, fulfilling the quality of service.

B. R. Stojkoska and K. Trivodaliev proposed an IoT and edge computing based smart home with
a three-tier framework. The unit comprises smart homes tier, nano grid tier and microgrid tier.

Department of MCA Page 7

 EDGE COMPUTING IN IOT DEVICES

The first tier consists of a household item that enables wireless communication. The data is
collected at the sink where it is locally processed and stored. In the second grid, sinks of smart
home communicate with one another. This communication is enabled through mesh, cluster, or
star. Sinks of various smart homes interact via IoT gateways enabled by the edge computing for
reliable computation.

Richard Olaniyan et al. define the Opportunistic Edge computing (OEC) paradigm. It also
furnishes the management framework to construct, organize and monitor the infrastructure of the
scalable edge that also involves the stakeholders. OEC utilizes the broker to launch pools of
resources with the contributions of the end-user at the edge. Thus, OEC pools are used to bring
the cloud nearby for a predefined duration. The author also compares the OEC model with that
of the existing models. The experimental results depict the merits and demerits of OEC over
cloud computing and fog computing. The paper also summarizes the resource management
factors in OEC environments.

Yunkon Kim and Eui-Nam Huh explore data caching concepts in edge computing. The state-of-
the-art relays on machine learning, reactive and proactive caching. A lightweight caching
algorithm, namely EDCrammer (Efficient Data Crammer), is implemented to manage computing
resources. This algorithm uses an enhanced PID (Proportional Integral-Differential) controller
for data streaming. EDCrammer helps in streaming data traffic, reduce the uplink load in the
cloud by providing high-quality video services. This model cache data efficiently, and the
caching rate is controlled amidst the cloud and the edge nodes. The algorithm produces a hit ratio
of about 96% with a cache capacity of around 1.5MB. Thus, with the edge node, the user can
receive better Quality of Service (QoS) by decreasing the load congestion on the data centre.

Y. Liu et al. explored the computation offloading using edge computing. The main objective is to
provide a better quality of service. The system framed a novel mechanism consisting of some set
of interactions carried over amidst the cloud and local edge for the Stackelberg game. In this
game, cloud server operator (CSO) serves as a leader that produces payments to end-server
owners (ESOs), and this ESO is considered as a follower determining the computation offload
for CSO. This model promises efficient computation offloading and extends its cooperation
among ESOs.

Department of MCA Page 8

 EDGE COMPUTING IN IOT DEVICES

J. Yang et al. proposed a new model of smart IoT toys. The author deployed a new prototype for
a data exchange accounting system based on edge computing. This prototype was designed with
Hyper ledger Fabric-based blockchain system. Exchange and payments of data are made amidst
demanders and suppliers through blockchain consortium. Smart contracts are deployed for
validating the data exchange records and billing between the peers. The proposed system
provides secure and reliable interactions between smart toys and IoT devices. This edge
computing-based data exchange (EDEC) comprises five phases, namely registration, the release
of data products, order generation, the transmission of data, accounting details and finally
payment. The transactions are all recorded and stored on edge computing.

Z. Zhao et al. deployed a three-step implementation approach to reduce the number of edge
servers for improving the throughput between Edge nodes and IoT devices. The proposed model
consists of three concepts such as discretization, utility metric and deployment algorithm. In the
first phase, the network is divided into small sections where the candidate node is determined. In
the next phase, the utility metric evaluates the performance of every candidate node by validating
the link quality and correlation. To achieve maximum throughput, the best node is deployed in
the network. The proposed work describes certain concepts of edge computing and provides at
most contributions in terms of real time applications and use cases in the IoT environment.

Department of MCA Page 9

 EDGE COMPUTING IN IOT DEVICES

CHAPTER 3

OVERVIEW OF EDGE COMPUTING AND IOT

EDGE COMPUTING ARCHITECTURE

Fig 3.1

Edge computing brings the resources of computation, storage, and networking significantly
closer to the devices, users, and applications. It is viewed as one of the significant technologies
associated with the Internet of Things and artificial intelligence in the next generation of
networks. It enables the data to travel at remarkably high speed without transferring data to the
cloud or data centre. Edge computing is multilayered distributed architectures that balance the
workload amidst edge layer, edge cloud, edge network and enterprise layer. Edge computing
comprises three nodes: Device edge, Local edge, and Cloud. Figure 3.1 illustrates the overview
of edge architecture with the relevant components within each node. Industry solutions/Apps can
be considered as duplicate components that exist both in the device edge and local edge. Specific
workloads are integrated either with the device edge and local edge, and they can be dynamically
migrated amid of the nodes either manually or automatically. Local edge node manages multi-
cloud management, network services.

 The device edge is the place where the device is occupied. These devices include
cameras, sensors and any physical devices gathering data and interact with edge data.

Department of MCA Page 10

 EDGE COMPUTING IN IOT DEVICES

Some applications such as AI models with deep learning, video analytics are managed on
these edge devices.

 Local edge encloses both applications and the network. The applica
applications
tions of the device
edge reside in this node. The non
non-runable
runable applications on the device edge are executed at
local edge. Such as IoT processing. Some virtualized and containerized network layers
will run on this local edge.

 Cloud is an environment where everything is brought together. It can either run at

premises or as a public cloud. The applications that are not able to run in either of the
nodes are handled by this cloud using orchestration layers.

Edge computing can also be defined in terms of comput

computing
ing power and latency. It allows us to
perform necessary computations with low latency with high computing
computing-intensive
intensive calculations.
Cloud, edge node, edge gateway and edge devices are the most essential layers in handling
application and transmitting data
ta aat the edge devices. Figure 3.2
.2 illustrates the layers of edge
computing architecture.

Fig 3.2

The purpose of the cloud is to serve as long term storage and coordinates the resources in the
case of irregular tasks. The edge node by default acts as downst
downstream.
ream. These nodes possess high

Department of MCA Page 11

 EDGE COMPUTING IN IOT DEVICES

computation power, 2q manages the network traffic. On the other hand, edge gateway manages
computations, and it serves as a platform for lower-level devices such as mobile phones, cars,
and sensors. In contrast, edge devices are composed of limited resources that include single
sensors and traffic lights.

Cloud Layer

Although edge computing was introduced to address network congestion and latency problems
commonly found in cloud computing, cloud computing in fact still plays an important role in the
entire edge computing architecture. We can say that cloud computing and edge computing
complement one another. Through the edge layer described in the next section, the entire system
determines if data needs to be processed in the cloud layer. If that is the case, edge servers will
pass data to the cloud layer for complex processing.

Edge Layer

This layer mainly consists of edge servers, and when compared to the cloud layer, the edge layer
contains edge servers that are larger in quantity and more vastly deployed. Therefore, through
distributed edge computing, the edge layer can process data that is closer to the data source and
address latency problems found in cloud computing. The edge layer can be considered the core
in the entire edge computing architecture. After data from the device layer is analyzed and
processed in the edge layer, data is transmitted to the cloud layer for subsequent processing and
analysis.

Device Layer

Amongst the three layers, the device layer contains the most devices. Ranging from devices that
are as small as our mobile phones or computers to ones that are as large as buses and factories,
these devices are all examples of components in the device layer. Through their sensors, devices
in the device layer collect and capture data used to help products achieve the purposes they are
designed for. Although components in the cloud and edge layers possess better computing
power, the devices in the device layer can still perform data analyses, processing and storage
tasks which require negligible computing power, as well as process data closest to the data
source in almost real-time.

Department of MCA Page 12

 EDGE COMPUTING IN IOT DEVICES

CONCEPTS OF EDGE COMPUTING

Fig 3.3
The smartness of Edge computing is that edge devices collect and share data without transporting
to another end of the server. It brings the data close to the point of interaction. The concept of
edge computing is stated as a computing model that executes at the edge of the network. The
cloud server is considered as downlink data, and IoT is justified as uplink data and edge to edge
computing is deliberated as arbitrary computing.
Edge computing can store and processing data significantly faster when compared to other
devices. It is implemented for real-time applications. In edge computing, algorithms run on an
edge server, or it increases the capacity of the smart phones by running algorithms via smart
phones. The storage resources include cloudlets, data centre and fog nodes. And it operates at the
edge of the network devices that remain nearing to the sensors or mobile devices. The main
objective of edge is to transmit the network of the cloud, storage, resources to the respective
network edges providing brilliant services at the edge of the network devices to meet the
emergency needs of the industries and factories. The edge computing is designed to meet the
requirements of low latency with high bandwidth on the network edge. Now, edge computing is
considered for many research purposes. The proposed work discusses the architecture,
potentialities, benefits, use cases, state-of-the-art applications, and challenges of edge computing
in IoT devices.

Department of MCA Page 13

 EDGE COMPUTING IN IOT DEVICES

WORKING OF EDGE COMPUTING IN IOT DEVICES

Fig 3.4

Edge computing allows data from Internet of Things (IoT) devices to be analyzed at the edge of
the network before being sent to a data center or the cloud.
We already knew about edge computing architecture in fig 3.2, which explains the edge
computing completely. Fig 3.4 gives the outline of the working of the edge computing by
collecting data from IoT devices and sending them to the edge for local processing and then
sending the necessary data to the cloud or the data center.

Fig 3.5
Fig 3.5 gives us the complete inner look of the working of the edge computing. Collecting and
processing can be seen in two phases i.e., Operational Technology and Information Technology.

Department of MCA Page 14

 EDGE COMPUTING IN IOT DEVICES

Operational Technology

a. Data Generation: creating or producing new data. This can be done through various means,
such as collecting data from sources, conducting surveys, performing experiments, or
generating data through algorithms and simulations.
b. Data Sensing: The high intelligence technology for analyzing customer behavior in order to
improve the market planning process according to the current customer needs and market
value.
c. Data Collection Aggregation: The process of summarizing a large pool of data for high
level analysis.
Then these data which is collected is passed to edge via edge gateways which act as a window
between the computing and data collection and aggregation.

Information Technology

a. Early Data Analytics: an analysis technique to analyze and investigate the data set and
summaries the main characteristics of the dataset. Main advantage of EDA is providing the
data visualization of data after conducting the analysis.
b. Deep Data Analytics: is the application of sophisticated data processing techniques to yield
information from large and typically multi-source data sets that may contain not only
structured data but also unstructured and semi-structured data.

Department of MCA Page 15

 EDGE COMPUTING IN IOT DEVICES

BENEFITS AND CHALLENGES OF EDGE COMPUTING

BENEFITS OF ENABLING EDGE COMPUTING FOR THE IOT

 Reduced Response Time / Latency

Running Facial Recognition at the edge rather than the cloud has helped researchers to
complete the recognition of any face with reduced response time by 169 to 900ms, that is,
the system was able to recognise the face in time ranging from 169 to 900ms less than
that of cloud system.
 Reduced Energy Requirement
Researchers used cloudlets to transfer the computing tasks for wearable devices. In this
case the response time was reduced by 80 to 220ms. The energy consumption was also
recorded to be reduced by 30-40% when using edge.
 Increases Data Security
As the data in edge computing is distributed to the devices only or at the nearest node, it
is difficult to attack each device to steal the data.
 Reduced Operational Cost
Edge helps in reducing the data traffic cost as well as the cloud data storage requirements,
which helps in reducing costs. At the same time, connection issues will not be that
problematic since the devices can work autonomously.
 Unrestricted Scalability
Unlike cloud, in edge computing the user has the option to increase their IoT network as
and when required, referencing to the storage available to the user is not essential.
 Better App Performance
Since the data is being processed at the edge, apps like Facebook, Instagram, photos app
give better performance to the users. It reduces the lag time, and the users get seem less
app navigation when still the images/videos are still getting uploaded.

Department of MCA Page 16

 EDGE COMPUTING IN IOT DEVICES

 Low Bandwidth Requirement

Giving all the computational work to the cloud proved to be very efficient since the cloud
had tremendous computing power. But with the increase in the number of devices
seeking the services of the cloud, the bandwidth requirement also increases, which is
basically on a stand still. As the devices increase, bandwidth is becoming the bottleneck
in the network. The big data is not able to reach the cloud in the desired time frame.

CHALLENGES FACED BY EDGE COMPUTING WHEN ENABLED IN IOT DEVICES

 Privacy & Security of data

IoT devices produce big data on daily basis. If these devices are deployed in a home
environment or in the field of healthcare, the loss of such data generated can lead to
privacy intrusion. For instance, with the readings of electricity and water meter of a
home, one can easily guess whether the house is vacant or people are staying there. If this
data goes in wrong hands, theft or burglary can be carried out. We take the scenario of
home Wi-Fi network security. Out of 43,90,00,000 household Wi-Fi routers 49% of the
routers are unsecured and more than 80% still have the routers set at default passwords.
Finding and implementing more efficient tools are the need of the hour, which will help
protect data privacy and secure the data.
 Reliability
It is the key challenge for the edge computing paradigm. Every electronic device has to
fail. There can a number of reasons for failure but most of the times it is very hard to
find. For example, malfunctioning of an AC can be due to compressor failure, power
cable defect or bad battery in temperature controller. Edge network can determine that the
AC has failed, but knowing the exact cause of failure will be very helpful in diagnosing
it. Edge OS should also be able to maintain a topology of the complete system which will
make network management and problem detection easier.
 Efficiency
Monitoring the performance of the edge device in real-time in the consumer point of view
from edge to cloud becomes difficult. By providing technology to view end-to-end

Department of MCA Page 17

 EDGE COMPUTING IN IOT DEVICES

monitoring needs to be deployed, this helps in transactional performance from edge to

cloud.
 Scalability
Increasing the scope of everything IT personnel deals with by adding more connected
devices at the edge. Edge computing requires a scaling-up of all IT disciplines, including
computation, network, storage, management, security, and licensing. When moving apps
to the network edge, businesses must keep the following in mind: Edge computing affects
everything IT touches, not just more equipment in a remote place.
 Speed
In cloud computing, the cloud is the main entity which decides where the computing is to
be performed. The devices feed their data to the cloud and all the programming and the
codes are in the cloud which is universal to all the devices. But in case of Edge
computing each IoT device or node has its own code/program to analyse and generate the
data. They might even have heterogeneous programs which might not be compatible to
each other. So, a universal standard of programming should be set for IoT devices which
should be accepted worldwide. Edge computing is a method of processing data locally,
close to users and devices. This saves bandwidth and reduces latency, resulting in the
high-speed digital experiences people have come to expect.

Department of MCA Page 18

 EDGE COMPUTING IN IOT DEVICES

CHAPTER 4
USE CASES

USE CASES OF EDGE COMPUTING ENABLED APPLICATIONS

1. MEDICAL APPLICATIONS
Overview: HPC at the Edge for medical imaging merges HPC/AI and medical sensing
technology in order to provide precision medicine through the use of real-time advanced
monitoring and analysis of a patient’s medical data to detect early pathologies while lowering the
risk of privacy breaches by keeping the data on site. This granular, yet massive amount of patient
data can be analyzed at the Edge, transformed, and then only pertinent data is sent to the cloud
such as alerts or data stripped of information that could lead to the patient’s privacy being
compromised. Medical Imaging at the Edge using HPC/AI removes the latency and dependence
on Cloud Computing resources, as well as reduces the patient’s digital footprint by limiting how
many systems have access to data. AI used in medical imaging provides tools that augment the
clinician’s intelligence in a way where they are able to provide better care at reduced costs.
Figure 4.1 illustrates the digital development in healthcare and how Edge Computing is being
used in healthcare.

Fig 4.1

Department of MCA Page 19

 EDGE COMPUTING IN IOT DEVICES

Examples:
Case 1: CT/MRI Scanning Technology
Centre for Clinical Data Science partnering with GE Healthcare, NVIDIA, Nuance, and DASA
to build HPC/AI technology that is integrated into CT, MRI, and Workstation machines which
utilize AI that is trained on vast data sets of diagnostic medical data.
Case 2: Handheld diabetic retinopathy diagnostic camera
Through NVIDIA’s virtual accelerator inception program, Taiwanese medical firm MiiS has
built a highly portable handheld diagnostic tool that combines NVIDIA Jetson TX2, a high-
resolution camera, Edge Computing architecture and GPU-powered AI algorithms to perform
instant screening of diabetic retinopathy.

Security Concerns: The EU has a strict legal framework in place to ensure consumer protection
and to protect personal data and privacy, minimizing risks to confidentiality and integrity of data.
Moreover, the EU GDPR sets certain rules and regulations for how the data is being processed
and handled in EU. In addition, there are specific rules that pertain to sectors such as healthcare
that will continue to apply to AI and medical devices: Creation of comprehensive documentation
and record keeping of data sets used for training and testing, programming and training
methodologies such as the processes and techniques used to build, test, and validate AI systems
especially those used to ensure the system is not biased in a way that could lead to prohibited
discrimination arising from the usage of AI, auditing abilities of how a patient’s medical data is
being used as well as who is accessing it.

2. SMART CITY
Overview: In the future cities will have sensors that will collect various data, for example, in
transportation, medical health, and urban security. Moreover, urbanization is rapidly increasing.
According to the UN, it is estimated that, by 2050, over 6 billion people will be living in the
cities. In the future, to have sustainable development in the town, a smart city is an excellent
solution. This might help to solve the problems that may arise in food supply, medical care,
transportation, culture and entertainment in the cities. These sensors will usually generate a large
volume of data, and this data should be processed quickly. Sending these data to the cloud will
need faster data movement (latency and data traffic in the network), and privacy. Therefore,

Department of MCA Page 20

 EDGE COMPUTING IN IOT DEVICES

these generated data should be processed closer to where it is produced. In general, Edge devices
have limited computing and storage, so it is also necessary to integrate multiple computing
models. A few cases of Edge Computing used in a smart city are listed below.

Fig 4.2
Examples:
Case 1: Closed-Circuit Televisions
CCTV’s are nowadays typically installed in almost all private and government premises. These
CCTVs will capture the movements of the objects. This will ensure the safety protocol in the
given premises. For this reason, the data collected by the CCTVs should be processed quickly.
To enable this, CCTVs should be connected to the Edge device through the LAN connection.
Edge devices can also use the latest technology called special image processing chip to process
videos more efficiently. Figure 4.2 shows an example of a smart city.
Case 2: Smart Home
A smart home is controlled by lots of sensors in lighting, kitchen, television, and surveillance,
and its technology is also rapidly developing with the help of IoT [59]. Again, to make it more
efficient (minimization of the latency) and effective (privacy of home data), the generated data
should be processed quickly where it is generated. Edge Computing will be playing an essential
role in facilitating that objective.

Security Concerns: While the user adoption for IoT technologies violating the boundaries of
private areas is surprisingly high, as testify for instance by the success of voice assistant like

Department of MCA Page 21

 EDGE COMPUTING IN IOT DEVICES

Amazon Echo/Alexa or Google Home, this raises several security concerns for spying
intelligence which remains an open challenge.

3. INDUSTRIAL / MANUFACTURING APPLICATION

Overview: Industry combines Edge HPC with AI in industrial automation environments. It aims
towards waste reduction, work reduction, and worry reduction in the work space. It is used for
connecting machines-to-machines and machines-to-people in a way where on-demand
production environments, equipment, and workers can quickly and intelligently react to
dynamically changing factory floor/environmental conditions. Certain industrial applications
may need to react quickly to real-time changing environmental conditions which may be
uncovered in data too voluminous to be sent to the cloud, such as image recognition data that
guides a robotic arm to interact with an object on a moving assembly line or creates alerts if
dangerous conditions arise. Moving data offsite for analysis may also incur transmission
latencies, which exceed the reactions times required for industrial applications, such as being
able to shutdown an assembly line if a foreign object interferes with the industrial process. This
is all assuming that the industrial site is even able to acquire a high-speed network connection
due to geographic constraints.

Fig 4.3

Department of MCA Page 22

 EDGE COMPUTING IN IOT DEVICES

Examples:
Case 1: Welding Quality Assurance
Rexroth, a Bosch company, has developed Weld Spot ML analytics software which operates on a
smart Edge that is located close to welding controllers. Until now, destructive tests were the only
reliable way to test if a spot weld was done according to quality specifications. Weld Spot
analytics software contains an AI engine which employs ML algorithms to detect anomalies and
provides this information to welding engineers through a UI that displays a variety of data and
analyses.
Case 2: Worker Ergonomics safety
A novel real-time spatio temporal Pyramid Graph Convolutional Network trained on video of
warehouse workers performing typical repetitive duties and their associated movements was
integrated with a traditional ergonomic risk index to assess the potential of musculoskeletal
disorders in the warehouse. This system can send alerts and warnings based on video analysis of
actions that are above a certain risk threshold.

Security Concerns: Limiting the amount of data sent to the cloud for processing also limits your
attack surface from malicious actors. There is a much lower risk of your data being intercepted,
tampered with, and stolen when it stays on-premise. Networking automated machinery poses a
threat where compromised machines can be manipulated either directly or by tainting real-time
and training data that your algorithms are using to control manufacturing processes.

4. SMART GRID & PUBLIC SAFETY

Overview: Electricity is one of the primary sources for humans to conduct most of the activities
in daily life. In recent years, special emphasis has been placed on how electricity is produced and
distributed to facilitate better economic, technical, and environmental reports. In particular, how
it is generated, distributed, and controlled, and monitored through digital instruments. The smart
grid is a term that refers to how the whole electricity production and distribution are controlled
by the smart digital instruments (for example, sensors) and embedded systems. Figure 4.4 shows
an example of Edge Computing in the smart grid. Over the past years, surveillance security has
been playing an important role in our daily life, for example, ATM centre. Most of the

Department of MCA Page 23

 EDGE COMPUTING IN IOT DEVICES

surveillance security is based on the visual feed, where this feed needs to be analyzed quickly
using AI/ML/DL for better security reasons without taking much time with accuracy. And also,
sometimes, there is some high risk of data being manipulated or leaked over the network. The
following cases show how Edge Computing will improve or tackle this problem.

Fig 4.4

Examples:
Case 1: Smart Grids
Such infrastructures will benefit in numerous ways by adopting Edge Computing. For example,
Edge Computing will make electricity to be bi-directional. This methodology allows the
customer to be both consumers and producers of energy. This means that it will enable the
customers to produce renewable energy (i.e., solar power and biofuels) and sell back their
excessive production to other consumers in the accessible market created by the Edge
computing-enabled smart grid.
Case 2: Visual Detection
During the process of visual detection, there could be a threat that might come from the
firmware, direct physical access, and visual layer-based attacks. Sometimes these threats can be
identified or not. In order to eliminate these threats, Edge Computing offers many solutions
through AI/ML. For example, in face spoofing attacks, one can use the online frame forgery

Department of MCA Page 24

 EDGE COMPUTING IN IOT DEVICES

detection technique. Furthermore, in many cases, at the Edge, it would be inefficient to process
the high-resolution video frames. There has been an algorithm defined in AI/ML to overcome
this problem, for example, the super-pixel-based technique.

Security Concerns: The automatic tracking of citizen movements and usage facilitated by Edge
Computing architectures raises a serious concern with regards privacy preserving rights, which
have to analyzed in the context of a degraded security climate due to the increase in terrorists
attacks hitting in the last decades EU countries among many others. For sure the ongoing
developments within the Edge Computing paradigm could be of great help to protect critical
infrastructures as smart grids, or sustain global decisions improving the global security of our
continents while preserving our privacy.

Department of MCA Page 25

 EDGE COMPUTING IN IOT DEVICES

CHAPTER 5
FUTURE SCOPE

In the technological perspective, edge computing is a providential framework enriching smart

applications with expeditious computing and storage resources. Cloud is steadily deployed for
smart devices with better computation and resource management. Due to its latency issues,
computing and storage resources are transmitted and centralized nodes to the edge node, thereby
reducing the burden of cloud computing. Thus, most of the services are pushed towards the edge,
as edge computing ensures shorter response time with reliability.
IoT Integration: The Internet of Things (IoT) is a growing field that generates a vast amount of
data. Edge computing will continue to play a crucial role in IoT by allowing for real-time
analysis and response to the data generated by IoT devices. As IoT technology continues to
advance, edge computing will become even more integrated into the IoT landscape, providing
faster and more accurate insights into IoT data.
Increased Security: As edge computing continues to gain popularity, security will become a top
priority. Companies will need to ensure that their edge computing solutions are secure, and that
the data they process is protected. This will drive the development of new security measures,
such as encryption and secure data storage, to protect data at the edge.
The evolution of edge devices with enhanced processing, computation, analytics, and storage
capabilities poses a new challenge: deciding on the best combination of technologies for each
deployment based on the processing requirements, the desired service level, and practicality.
So what does the future hold for edge computing? Intel aims to integrate one trillion transistors
into a single device by 2030, which is a testament to the limitless possibilities in this field. In the
future, we might also see the development of eco-wellbeing-edge, a zero-carbon deployment
model with the ability to assess users’ stress levels and well-being.
In conclusion, the future of edge computing is bright, with increased adoption, improved
performance, IoT integration, and increased security. Edge computing will continue to play a
crucial role in processing and analyzing data, allowing businesses to make informed decisions in
real-time. As edge computing technology continues to evolve, businesses can look forward to
even more sophisticated and efficient solutions in the future.

Department of MCA Page 26

 EDGE COMPUTING IN IOT DEVICES

CONCLUSION

This discusses some applications and potentialities of Edge Computing in IoT devices. To have a
deep insight into edge computing, this report surveys some additional concepts of edge
computing such as its architecture, its overview, etc. This report enumerates some of the edge
computing use cases in various domains. Nevertheless, this report also provides some
complementary aspects of edge computing with that of cloud computing. A comparison table has
been provided concerning the benefits of edge over cloud. And the challenges of edge computing
and addressing techniques have clearly been described. This report offers systematic concepts of
edge computing from the architecture to the challenges. The in-depth description includes data
storage, latency and time-management as edge computing has become a research-oriented model
or domain. This report also sketches about the challenges of edge computing. Applications of
Edge computing in IoT devices are furnished in this report. This report investigates the outlook
of edge computing. Though it faces some difficulty, edge computing will be the future network
in the real world which provides continuous development in the internet, AI, and also plays a
vital role in human society and the geological environment.

Department of MCA Page 27

 EDGE COMPUTING IN IOT DEVICES

BIBLIOGRAPHY

[1] S Mangesh, J Indumathi, Radha RamMohan, International Journal of Computer Networks

and Applications (IJCNA), Volume 7, Issue 5, September – October (2020), page 146 – 149.

[2] Niveditha V R, P Shanmuga Prabhu, International Journal of Computer Networks and

Applications (IJCNA), Volume 7, Issue 5, September – October (2020), page 153 – 154.

[3] Ezhilmathi Krishnasamy, Sebastien Varrette, Michael Mucciardi, Partnership for Advanced
Computing in Europe, Prace, page 3 – 4.

[4] Wazir Zada Khan, Univesity of Wash, Future Generation Computer Systems, February 2019,
at: https://www.researchgate.net/publication/331362529, page 25 – 29.

[5] Ezhilmathi Krishnasamy, Sebastien Varrette, Michael Mucciardi, Partnership for Advanced
Computing in Europe, Prace, Use Cases, page 14 – 17.

[6] Mehtab Alam, Wesleyan Journal of Research, Vol 14 No 07 March 2021, EDGE
COMPUTING AND ITS IMPACT ON IOT, page 211 – 215.

[7] Ihtiram Raza Khan, Wesleyan Journal of Research, Vol 14 No 07 March 2021, EDGE
COMPUTING AND ITS IMPACT ON IOT, page 216 – 219.

[8] S.Anitha Jebamani, S Godfrey Winster, SRM Institute of Science and Technology, A Survey
of Edge Computing in IOT devices, page 11 – 13.

[9] Elkhatib, Y., Porter, B., Ribeiro, H.B., Zhani, M.F., Qadir, J. and Rivière, E., 2017. On using
micro-clouds to deliver the fog. arXiv preprint arXiv:1703.00375.

Department of MCA Page 28

 EDGE COMPUTING IN IOT DEVICES

[10] Fortinet, https://www.fortinet.com › resources › cyberglossar.

[11] IBM Developer, https://developer.ibm.com>articles>iot-vs-edge-com

[12] Wang, R., Yan, J., Wu, D., Wang, H. and Yang, Q., 2018. Knowledgecentric edge
computing based on virtualized D2D communication systems. IEEE Communications Magazine,
56(5), page 32-38.

[13] Coady, Y., Hohlfeld, O., Kempf, J., McGeer, R. and Schmid, S., 2015. Distributed cloud
computing: Applications, status quo, and challenges. ACM SIGCOMM Computer
Communication Review, 45(2), page 38-43.

[14] Mouradian, C., Naboulsi, D., Yangui, S., Glitho, R.H., Morrow, M.J. and Polakos, P.A.,
2017. A comprehensive survey on fog computing: State-of-the-art and research challenges. IEEE
Communications Surveys & Tutorials, 20(1), page 416-464.

[15] Kang, J., Yu, R., Huang, X. and Zhang, Y., 2017. Privacy-preserved pseudonym scheme for
fog computing supported internet of vehicles. IEEE Transactions on Intelligent Transportation
Systems, 19(8), page 2627-2637.

Department of MCA Page 29