1. See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/355064336
Computing Paradigms: An Overview
Conference Paper · August 2021
DOI: 10.1109/ASIANCON51346.2021.9545007
CITATION
1
READS
885
2 authors:
Neha Sharma
Chandigarh Group of Colleges
9 PUBLICATIONS 9 CITATIONS
SEE PROFILE
Chander Prabha
Chitkara University
22 PUBLICATIONS 38 CITATIONS
SEE PROFILE
All content following this page was uploaded by Chander Prabha on 22 April 2022.
The user has requested enhancement of the downloaded file.
3. From this Edge, Fog and Cloud ecosystem, the author of
[12] explored different aspects of system architecture,
application features, and platform abstractions, and also
highlighted the latest capabilities of the Edge and Fog layers,
such as physical and application agility, privacy protection,
and an emerging runtime environment.
To address the issue of large-scale access to cloud
resolution resources in processing, connectivity, and storage,
the author presented a fog computing-based face recognition
and resolution strategy [14]. Most other biometrics can also
benefit from this framework. To create a model for
producing face identifiers, the author employs the proper
face representation method for face identification.
The author [17] enhances the omnipresent neuroimaging
system environment. This paper explicates its pervasive
scheme design and presents its favourable technique,
including protocols for publishing/subscribing machine-to-
machine, multi-tier fog/cloud computing framework, and
connected web information.
The idea of a holistic view of imperative computing
expands the direction of resources from computing resources
on the information acquiring and reprocessing devices in
emergency situations. This system's layered structure is
introduced in detail and its realignment is proposed in the
event of difficulty. Using Urgent Service Profiles, this
procedure is aligned with large-scale computing. Practical
work carried out under the ISMOP project validated the
proposed approach [18].
Within the Fog Computing architecture, Author [19]
applied these methods in a new way, so that all the methods
could be joined with a particular understanding that is
merely available at the edge (Fog) devices. With network
edge, particular knowledge, better dynamic adaptation to the
customer conditions (like the status of the network and load
of device computation) can also be achieved
Another group of authors [20] in their work carry out a
review of the expansion of iot and their assimilation with
cloud computing for improved and more useful user service
provisioning and efficient resource consumption. This
integration or coordination work, referred to here as the
Cloud of Things (CoT), involves certain key challenges
III. FOG COMPUTING AND SIMILAR COMPUTING
PARADIGMS
This segment compares fog computation and similar
computation paradigms to show how FC can be useful in a
diversity of aspects. Furthermore, this section clarifies how
various computing models can boon the existing and future
environment of linked devices. Fog computing (FC) is
compared to cloud computing, and also with other connected
computing paradigms.
A. Cloud Computing
Cloud computing refers to the dissemination of various
services over the Internet. Various resources provided by cc
are Storage capacity, database, servers, and applications [27].
Cloud computing provides two types of models the first one
is the cloud deployment model and the second one is the
cloud service model.
Fig. 1. Cloud Computing Models [27].
B. Edge Computing (EC)
It is the latest technology that enables the processing of
an enormous quantity of information generated by Internet-
connected gadgets via the Internet of Things (IO) [28].
Computational information, services, and applications are
directed aside from the Cloud servers to the edge of the
network with edge computing. Edge computing systems can
be used by content suppliers and app developers bringing
services closer to the user.
Edge computing is described by high bandwidth, ultra-
low latency, and real-time access to network data that could
be used by a variety of applications. Edge computing shares
a number of characteristics with Cloud computing. On the
other hand, it has several defining properties that make it
stand out:
i) Geographical Distribution is Dense.
ii) Assistance with Mobility
iii) Location Sensitivity
iv) Closeness
v) Latency is low.
vi) Context
Cloud computing
model
Deployment
model
Service Model
Public Cloud
(Publically
access services)
Private Cloud
(Reserve for
specific
reorganization)
Hybrid Cloud
(Combination
of Public +
Private cloud)
Software as a
service
(On demand
services)
Infrastructure as a
service
(Rented servers)
Platform as a
service
(Platform for
creating software
that is delivered
via the Internet)
2
Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on April 11,2022 at 05:48:38 UTC from IEEE Xplore. Restrictions apply.
4. Fig. 2. Edge computing (EC) Architecture[28].
C. Fog Computing
Fog networking enhances the conventional cloud-based
computing model and the network edge services. It supports
distinct services at the network edge like computing,
controlling, and transmission. There is a lot of difference
between the distribute premise model and the traditional
computational paradigm.
Over the years, the numerous fog networking model has
been proposed. They are generally derived from the essential
three-layer framework. It increases the resources of the cloud
to the network edge by proposing layers in between the
cloud and also the end devices. Fig. 1 represent the
framework of fog networking [8].
Fig. 3. Framework of Fog Networking [8]
1) First Layer
The first layer of fog computing is the term layer that is
close to the real environment. It resists distinct Internet of
Things devices like smart (phones, cars, vehicles, and cards),
mobile phones, and etc. The savvy gadgets have the ability
of computing gadgets; we can use these devices here as a
detection tool. All these devices are important for sensing,
real-time information and transform this information to the
fog layer for storage handling.
2) Second Layer
This layer is known as Fog layer. It is located on the
network. It is a mixture of a huge number of fog devices
such as base stations, gateways, and network devices. It is
also the middle layer of the fog computing framework. It
combines a large number of fog nodes like computer
network devices. The devices are usually shared between the
end gadgets and the cloud. They can be fixed at a constant
location. The edge devices easily associate with the fog
devices to access services. They have the capacity to
compute, communicate, and collected sensed information for
a time. The original time inquiry and latency delicate
application can be adapted in fog layer
3) Third Layer
It is the third layer of the cloud network model. This
layer supports immense efficiency servers and storage
gadgets, and also supports different application resources
like a smart (home, industry, and transformation). This layer
consists of effective computational and storage capacity that
is helpful for broad calculation surveys.
In spite of all, contrasting from the conventional cloud-
based computing model, neither all the computation nor the
data storage jobs go via the cloud. Here, each end node and
smart device is linked with another fog node with the help of
connectionless and connection-oriented technology.
The fog node could be related and interconnected by
connection-oriented and connectionless mechanism and
every fog device is connected within the cloud by internet
protocol core network. Following are the various advantages
that are associated with the fog computing such as
scalability, low latency requirements, network traffic
reduction [25].
IV. FOG COMPUTING
CHARACTERISTICS,CHALLENGES AND
APPLICATIONS
A. Fog Computing Characteristic
It is a virtual platform that serves the processing of
information, and networking resources within the end edge,
and the data center of conventional cloud computing,
however, not particularly placed at the network edge. The
main components of cloud-based computing and fog
networking, computing are Storage processing and
networking assets. Below, are the lists of various
characteristics of fog computing [9] :-
1) Minor Suspension and real-time interaction
Fog networking devices at the network edge regionally
collect the information create by detectors and gadgets, and
the information is processed and stored by the network edge
gadgets in the Local Area network. It automatically
decreases the information activity beyond the Internet and
supports quick immense feature limited resources provided
by endpoints. So, it permits minor waiting time and joins the
need of real time connection particularly for time delicate
application [11].
CLOUD
Edge Nodes
Internet
Users
LAN/WAN
Sensor/
Controller/
Devices
Cloud Data
Center
(Store all the
data)
LAYER 3
LAYER 2
LAYER 1
Fog Nodes
(Gateways,
base station)
Networking
Devices
(Storage
gadgets)
3
Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on April 11,2022 at 05:48:38 UTC from IEEE Xplore. Restrictions apply.
5. 2) Mobility Support
There are a number of mobile clouds computing devices
like smart phones, smart watches, and vehicles so that
geographical mobility is common at the first layer of the fog
networking, even though there are few gadgets that endure
fixed, like a red light camera that is significant for fog
networking to transfer Every end hubs or momentary gadgets
measure the enormous measure of information created by the
IoT, and really accomplishing portable information
examination [12].data precisely using cell phones. There is
no need to move information from the cloud to the base
station.
3) Secure Transmission Capacity
Fog networking enhances the data processing and
capacity, ability to the network gadgets to achieve
information preparing and putting away, linking with the end
gadgets, and the ordinary cloud. In a few application scheme,
decision-building is nearby accomplished in the fog devices,
instead of achieving with cloud. This benefit will become
increasingly powerful along with the growth of the amount
of information in the current generation [13].
4) Heterogeneity
The organization structure of fog processing is
excessively heterogeneous, which joins not just rapid
associations interfacing with the data place, yet in addition
connectionless advances, for example, WLAN, Wireless
fidelity,3G, and 4G, and so on Partner to edge hubs [14].
5) Geographical Allocation and decentralized
information expository
It comprises of an enormous number of extensively
disperse nodes, which have the capacity to record and decide
the areas of end edges so as to give the portability. Then
again, preparing and putting away information is supported
server farm removed from the end-user, the scatter system of
mist registering to give the proximity of data examination to
the client [14].
B. Fog Computing Challenges
There are different types of challenges in fog computing
that are listed below [10]
1) Structural problem
Various components from both the core network and the
edge may be utilized as promising Fog networking
framework and these elements is provided with distinct types
of processor, however, that is not utilized for the basic
reason figuring. Give the components regular reason
calculation aside from their conventional movements will be
exceptionally testing.
2) Service determine
Not all Fog gadgets are resources improved. Huge scope
applications progress in resources constrained gadgets isn't
exactly simple in contrast with the conventional server
farms. In this situation, a promising programming stage of
dispersing applications progress in Fog are should be
suggested.
3) Safety aspects
Fog figuring is portrayed upon regular system
administration components; it is massively powerless against
security assaults.
C. Fog computing enhance applications with minor
latency demands, thus it has the promising to be used in
several applications that is the delay delicate as in
medical care, prompt administrations and digital actual
frameworks. Present, we listed few applications of
computing [8].
Fig. 4. Fog Computing Applications [8]
1) Smart Environment
For smart living environment and the IoT applications,
the network is the key element and the latter is collected out
of intelligent devices (sensors, controller and inter-
connectors) and different types of processor. For orientation
and cooperation among smart devices, cloud assistant are
used [15].
2) Healthcare
Medical services have fascinated large amounts of
published works. A broad assortment of works regarding
control, disclosure, analysis, and assurance of wellbeing
dieses have been suggested as of late [16].
3) Smart Energy Framework
Energy framework is a power dispersion network that
sets up intelligent meters at the zone to gauge the constant
status information. Regarding energy achievement, energy
circulates energy burning-through and power charging. Keen
energy show to the utilization of organizational
administration innovations, and Internet of Things to
growingly convey power so as to lessen their expense just as
most elevated force, which incorporate choice achieving and
choice taking subsystem [12].
4) Improve reality, brain-machine integrates and
gaming
A lot of standard products and projects like Microsoft
Holo lens, and Sony smart eyeglass can be adapted in an
augmented real-world application. Augmented applications
mainly need huge bandwidth for information transfer and
high energy consumption to progress program streaming.
Considering, a little break in the reach of milliseconds can be
a misfortune the client experience and have a negative
reaction, low idleness is a need for the increased genuine
world and brain-related applications. It is an outstanding
model that can achieve such conditions. Improve reality
support by fog networking can diminish inactivity in both
handling and correspondence and exaggerate throughput
[17].
FC
Applications
Smart Environment Healthcare
Immediate
processing
Smart Energy Improve reality
4
Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on April 11,2022 at 05:48:38 UTC from IEEE Xplore. Restrictions apply.
6. 5) Immediate processing and other application
Fog networking is very appropriate for surroundings for
immediate processing. Applications that required instant
utilization of fog-based computing is in web improvement,
since all web appeals that the client makes initially going to
the server’s edge, which later gathers them from the network
core where the web server finds and regionally collects these
documents.
V. COMPARATIVE ASSESSMENT OF CLOUD
COMPUITNG ,EDGE COMPUTING & FOG
COMPUTING
An analysis on various computing paradigm such as
cloud computing, edge computing and fog computing done
in this section based on some parameters like security,
latency, load migration, scalability, similarities differences,
and drawbacks.
TABLE I. COMPARISION OF VARIOUS COMPUTING PARADIGM
Computing
Paradigms/Parameters
Cloud
Computing
[27]
Edge
Computing
[28]
Fog
computing[26]
Security Moderate High High
Latency High Low Moderate
Load Migration High Moderate Moderate
Flexibility High High high
Scalability High High Low
Bandwidth Moderate Moderate Moderate
Advantages More flexible
in user work
practice
High speed High Security
Disadvantages Risk of
information
confidentiality
More
storage
space
Complexity
VI. VARIOUS FOG COMPUTING ALGORITHMS
In this section various fog computing algorithms are
discussed based on distinct parameters such as security,
latency, and flexibility.
Numerous techniques are accessible to adjust the weight,
privacy, and cost in fog computing. In this section, some of
the algorithms and their drawbacks are discussed below: -
A. Advanced Encryption Standard (AES)
B. Job Scheduling algorithm (JSA)
C. Migration Modelling & Learning Algorithm (MMLA)
D. Decentralized algorithm for Randomized task
Allocation (DART)
E. Time–Cost aware Scheduling (TCaS)
A. Advanced Encryption Standard (AES)
It is deemed as more powerful as compared to the data
encryption algorithm. In the case of encryption of electronic
data, it is the most advanced and security standard. It uses a
symmetric key encryption method with the key size 128,
192, 256 bits. For the fog environment, AES may be
considered more appropriate and flexible. This algorithm has
been used by the author in the second layer of fog computing
to improve security.AES is more secure as compare to the
other security algorithm because AES 120 is considered to
be unbreakable [1].
B. Job Scheduling algorithm (JSA)
For latency-critical systems, it reduces the delay. Based
on the length, it schedules the jobs on the fog node and
reduces the total loop backlog and memory usage. Since the
projected algorithm SJF (Shortest job first) reduces the total
waiting period, tuples with larger lengths will starve. In the
future, the author will try to incorporate certain meta-
heuristic, hyper-heuristic, reinforcement learning-based
method, etc. for scheduling
C. Migration Modelling & Learning Algorithm (MMLA)
The present FC also lacks the mobility support system
when faced with diversified device consistency criteria for
many Smartphone devices. Such a system of mobility
assistance can be crucial, as in the industrial, internet, where
individuals, products, and computers can be moved. The
MMLA is proposed to fill in these gaps. As a large-scale
MDP issue, the container migration issue of mobile
application tasks in fog computing. First, we describe the
system model, whose cost function composed of latency,
consumption of power, and cost of migration. Then we
implemented the algorithms named as deep Q-learning-based
container migration. In the Q-network update, we enhance
random action selection in research and DNN training
program to accomplish quick decision making [22].
D. Decentralized algorithm for Randomized task
Allocation (DART)
The author suggested decentralized techniques to enable
gadgets to make local offload choices. Inspired by the
extensively considered architecture of nested fog computing.
In static mixed strategies, it is centered on an equilibrium job
scheduling. This model has a number of interesting
extensions. Firstly, a communication technique in which
gadgets allocate bandwidth with everyone could be
considered. The new approach is to determine the power
consumption of discharging [23].
E. Time–Cost aware Scheduling (TCaS)
The TCaS algorithm's main goal is to achieve a good
trade-off with both implementation time and monetary price
in order to complete a bag of Cloud-Fog system tasks.
Furthermore, this algorithm can adapt flexibly to the
requirements of multiple users in which someone at present
wants to prioritize the execution time and others want to
complete their tasks with a tight budget. By focusing on
maximizing many other priorities, such as time, delivery
costs, processing power, and energy usage, the author can
extend the scheduling issue to accommodate consumers [24
VII. CONCLUSIONS
The fog based computing is treated as the main
component in processing world, and there are a huge number
of the devices are connected. Fog computing model pushes
new applications and administrations for cloud to the
organization end. It enormously diminishes the information
transmission hour and the heap of organization
correspondence, Adaptability meets the request of real –time
delay delicate application and comfort, network transfer
speed blockage. In this paper, we clarified the theory of Fog
networking architecture, applications and obstacles and
observed in detail. We survey and described framework of
fog based computing and its qualities. Different application
conditions like Healthcare, Smart environment augmented
real world are introduced, Moreover, clarified fog computing
5
Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on April 11,2022 at 05:48:38 UTC from IEEE Xplore. Restrictions apply.
7. applications and different fog computing algorithms. In this
paper, various computing paradigms are discussed and the
comparisons of these techniques are done on the basis of
some parameters like security, latency, load migration,
scalability, similarities differences, and drawbacks .Fog
networking will give, as a more brilliant and green data
processing paradigm to help the growth of big data , and
IoT. This is a significant and interesting exploration region
which will influences future scholarly world and
participation.
REFERENCES
[1] Akhilesh, V., Ramya, P. & Jing, H., (2016). Security in Fog
computing through Encryption. I.J. Information Technology and
computer science, 5,28-36.
[2] Ivan, S. and sheng, & w., (2014). The Fog Computing paradigm:
scenario and security issues. Federated conference on computer
science and information system, 02,1-8.
[3] Shanhe, Y., Zijang, H., Zhengrui, Q. & Qun, L., (2015). Fog
Computing: Platform and Applications. Third IEEE Workshop on
Hot Topics in Web Systems and Technologies, 22,73--78.
[4] Jiang, Z., Douglas S., C., Mythili Suryanarayana, P., & Preethi, N.
(2013). Improving Web Sites Performance Using Edge Servers in
Fog Computing Architecture. IEEE Seventh International Symposium
On Service-Oriented System Engineering, 320-323.
[5] Nasir, A., Yan, Z., Amir, T., & Tor, S. (2018). Mobile Edge
Computing: A Survey. IEEE Internet Of Things Journal,, 05(01),
450-465.
[6] Le, G., XU, k., Meina, S., & unde, S. (2011). A Survey of Research
on Mobile Cloud Computing. 10Th IEEE/ACIS International
Conference On Computer And Information Science, 387-391.
[7] Michaela, I., Larry, F., Robert, B., Michael J., M., & Nedim, G.
(2018). Fog Computing Conceptual Model. National Institute Of
Standards And Technology, 1-8.
[8] Pengfei, H., Sahraoi, D., Huansheng, N., & Tie, Q. (2017). Survey on
Fog Computing:Architecture ,Key Technologies,Applications and
open issues. Journal Of Network And Computer Application, 27-42.
[9] Flavio, B., Rodolfo, M., Jiang, Z., & Sateesh, A. (2012). Fog
Computing and Its Role in the Internet Of Things. MCC, 13-15.
[10] Redowan, M., Ramamahanarao, K., & Rajkumar, B. (2018). Fog
Computing : A Taxonomy, Survey and Future Directions. Cloud
Computing And Distributed Systems(CLOUDS) Laboratory
Department Of Computiing And Information System., 103-130.
[11] Shanhe, Y., Cheng, L., & Qun, L. A Survey of Fog Computing :
Concepts, Applications and Issues.
[12] Prateeksha, V., & Yogesh, S. (2017). Demystifying Fog Computing:
Characterizing Architectures, Applications and Abstractions. IEEE
1St International Conference On Fog And Edge Computing (ICFEC),
115-125.
[13] Pengfei, H., Huansheng, N., Tie, Q., & Xiong IEEE Proof, L. (2016).
Fog Computing Based Face Identification and Resolution Scheme in
Internet of Things. Ieee Transactions On Industrial Informatics, 1-11.
[14] Hu, P., Huansheng, N., Tie, Q., Yanfei, Z., & Xiong, L. (2016). Fog
Computing-Based Face Identification and Resolution Scheme in
Internet of Things. IEEE Transactions On Industrial Informatics, 1-
11.
[15] Jianhua, L., Jin, J., Dong, Y., Marimuthu, P., & Klaus, M. (2015).
EHOPES: Data-centered Fog Platform for Smart Living. International
Telecommunication Networks And Applications Conference
(ITNAC), 308-313.
[16] Vladimiri, S., Ahmed, B., GHULAM, S., & Johannes, S. (2015).
Smart Items, Fog and Cloud Computing as Enablers of Servitization
in Healthcare. Sensors & Transducers, 185(02), 121-128.
[17] John, K., Gan, T., Chun-Kai, Y., & Yu-Te, W. (2014). Augmented
Brain Computer Interaction based on Fog Computing and Linked
Data. International Conference On Intelligent Environments IEEE,
374-378.
[18] Robert, B., Marek, K., Kwolek, B., Piotr, N., Tomasz, S., &
Krzysztof, Z. (2015). Holistic approach to urgent computing for flood
decision support. CCS International Conference On Computational
Science, 51, 2387–2396.
[19] Zhu, J., Douglas, S., Mythili, S., Preethi, N., Hao, H., & Flavio, B.
(2012). Improving Web Sites Performance Using Edge Servers in
Fog Computing Architecture. IEEE Seventh International Symposium
On Service-Oriented System Engineering, 320-323.
[20] Mohammad, A., ham Phuoc, H. and Eui-Nam, H., (2014). Smart
Gateway Based Communication for Cloud of Things. IEEE Ninth
International Conference on Intelligent Sensors, Sensor Networks and
Information Processing (ISSNIP), pp.1-6.
[21] Bushra, J., Mohammad, S., Israr, A. and Atta, U., (2019). A job
scheduling algorithm for delay and performance optimization in fog
computing. John Wiley & Sons, Ltd., pp.1-13.
[22] IEEE Transactions on Services Computing, (2015). Migration
Modeling and Learning Algorithms for Containers in Fog Computing.
14(08), pp.1-14.
[23] Sladana, J. and György, D., (2019). Decentralized Algorithm for
Randomized Task Allocation in Fog Computing Systems.
IEEE/ACM TRANSACTIONS ON NETWORKING,, 27(1),pp.85-
96.
[24] Binh, M., Huynh Thi, T., Tran, T. and Do, B., (2019). Evolutionary
Algorithms to Optimize Task Scheduling Problem for the IoT Based
Bag-of-Tasks Application in Cloud–Fog Computing Environment.
Applied science, 9, pp.1-20.
[25] Dastjerdi, A., Gupta, H., Calheiros, R., Ghosh, S. And Buyya, R.,
2021. Fog Computing: Principles, Architectures, And Applications.
Pp.61-75.
[26] Yousefpour, A., Fung, C., Nguye, T. and Kadiyala, K., 2019. All one
needs to know about fog computing and related edge computing
paradigms: A complete survey. JournalofSystemsArchitecture, 98,
pp.289-330.
[27] Jaspreet, S., Deepali, G. and Sharma, N., 2019. Cloud Load
Balancing Algorithms: A Comparative Assessment. Journal of
Computational and Theoretical Nanoscience, 16, pp.1-6.
[28] Wazir Zada, K., Ahmed, E., Hakak, S., Yaqoob, I. and Ahmed, A.,
2019. Edge Computing: A Survey. Research Gate, pp.1-40.
6
Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on April 11,2022 at 05:48:38 UTC from IEEE Xplore. Restrictions apply.
View publication stats