Proceedings of the Third International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC 2019)

IEEE Xplore Part Number:CFP19OSV-ART; ISBN:978-1-7281-4365-1

A Survey of Fog Computing: Fundamental,

Architecture, Applications and Challenges
Ravva.Sai Sanketh Dr.Y.MohanaRoopa Panati.Viswa Narendra Reddy
Department of Computer Science and Department of Computer Science and Department of Computer Science and
Engineering Engineering Engineering
Institute of Aeronautical Engineering Institute of Aeronautical Engineering Institute of Aeronautical Engineering
Dundigal, Hyderabad Dundigal, Hyderabad Dundigal, Hyderabad
[email protected] [email protected] [email protected]

Abstract—Fog computing is the advancement or extension of

cloud computing paradigm. It is the advanced distributed system, C. Multi Access Edge Computing
where the network keeps the data close and reduces the time to The multi access edge computing improves the cloud
deliver the information to the end users. This research work computing efficiency. The main purpose of nomadic edge
presents the discussion about fog computing architecture by computing is to bring running applications and information
explaining its working nature with the support of various layers. near to the end user.
The limitations of the cloud computing and how fog computing
overcomes cloud computing drawbacks and its application in
D. Mobile Cloud Computing
various fields are also discussed.
The Mobile cloud computing contains both mobile
Keywords— Cloud computing ; fog computing; mobile cloud computing and cloud computing. It improves the processing
computing; edge computing; mobile edge computing; and computational resources to the end users
E. Fog Computing
I. INTRODUCTION
As the devices that perform on the brink of the virtual edge The advanced version of cloud computing is fog
get smarter and extra efficient, they continuously capture computing. It overcomes all drawbacks and limitations of
information to ease the human life so to establish connection cloud computing it is the distributed system in between
between the devices in the traditional system because it end user and cloud it delivers the information required for
supports its users in many ways by reducing the project cost, the end user quickly than cloud[4].
back up the information, and connects the automotive software
to hardware gadgets. But there are some disadvantages in cloud
computing they are the connection between the devices,
privilege issues [1]. To overcome the above-mentioned
problems we use fog computing. This Fog computing is also
called fogging or fog network it is the redistributed system it is
close to the end user so it is faster in delivering data to the user
with reducing delay time. It provides storage, data, compute;
applications are shared between the cloud and the source of
data. So that it can deliver the information to the end user
quickly [2]. The fog computing also decreases the size of
information sent to the cloud for analyzing and saving the
information.

II. FOG COMPUTING OVER VIEW

Some similar technologies to the fog computing are
cloudlet, multi-access computing, private cloud computing.
A. Cloud-Let Computing Fig 1: Platform of Fog, Mobile Cloud, Cloud, Edge, Mobile Edge
Computing
A cloudlet is a data center that is connected to the cloud
the main cause of cloudlet is assisting aid-intensive and user III. CHARACTERISTICS OF FOG COMPUTING
collective apps on mobile devices with less delivery period.
A. Monitor Traffic With Ease
B. Private Cloud Computing
According to a survey, 30 billion gadgets connected
A private cloud computing is that works by shared internationally, in future, it might soar up to the large number
computing resources rather than having a private server. It
by 2020[5-8]. These billions of gadgets include play-station,
reduces the cost of the project [3].

978-1-7281-4365-1/19/$31.00 ©2019 IEEE 512

Authorized licensed use limited to: Hochschule Landshut. Downloaded on December 02,2024 at 09:22:38 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the Third International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC 2019)
IEEE Xplore Part Number:CFP19OSV-ART; ISBN:978-1-7281-4365-1

pills, and smart phones used for generating and send communication between gadgets in the platform and to control
information to other end gadgets depending upon the features them. The underlying layer of fog computing architecture is
and hardware used may additionally accumulate statistics in the physical layer. This layer is responsible for establishing
computing [9-10] the connection between gadgets or devices in a single platform
and exchanges the information between them. The physical
B. Acceptance Towards IOT Jobs
layer contains gadgets, terminals, sensors, virtual sensors. In
The smart gadgets are increasing day-by-day the main this layer, sensors are decentralized to sense the nearby
motive of these smart gadgets are to know or sense the locations data faster and send this information to a superior
surroundings of the gadget. So these gadgets have to sense the layer of the architecture, and the nodes of this layer are
surroundings and update as fast as fogging provides less controlled depending upon its functions. The second layer in
latency time it is used in IoT [11-13]. the architecture of fog computing is monitor layer it keeps
C. Reducing Storage Space watch the usage of resources, functioning of nodes and
sensors. The next layer of fog computing architecture is pre-
Fog computing might be an extremely good choice for
processor layer it is liable for managing the record data the
avoiding the redundant information and useless information
analysis of information is accomplished [7]. In this layer
send to other gadgets in the total platform may reduce the size
minimization of information is done by cutting down the
of the storage area.
unwanted information it reduces the useless communication.
IV. THE ARHITECTURE OF FOG COMPUTING NETWORK
The next layer is storage layer is liable for storing the
As stated earlier, fog spreads cloud to the threshold information most records are saved in fog on a compendiary
so it is called edge computing. The fog computing is to style. For storing the data for more time, the cloud is
connect nodes directly to network assets which producing the appropriate because it has many sources [8]. As soon as the
records-leaving at the node level. It is very virtualized network information is sent to cloud fog no more save the data. The
this is liable for supplying storage space and networking next layer is the privilege layer many application of fog
between the end users in addition to the control it indicates a computing needs the security measures. This layer safeguards
standard form where the information while sending it to inclined channels. After
information is ready to send to cloud the transport layer is
called on after transmitting the information to the cloud the
difficulty in middle platform reduces. The fog sources are
placed in the middle of the cloud and IoT gadgets [15-17]. The
fog can remove the unwanted information so it reduces the
latency it provides good streaming in mobile nodes, which
includes transferring car with help of provides and get
admission to factors positioned as consequence.

Fig 2: Layered Architecture

As stated earlier, fog spreads cloud to the threshold so it is
called edge computing. The fog computing is to connect nodes
directly to network assets which producing the records-leaving Fig 3: Tire Level Architecture
at the node level. It is very virtualized network this is liable for V. LIMITATIONS OF CLOUD
supplying storage space and networking between the end users
in addition to the control it indicates a standard form where in While comparing with preceding conventional computing
specific times of committed fog offer sources close to the paradigm, cloud offers numerous benefits, but it also has some
network [6]. The architecture of fog is making easy drawbacks and demanding situations

978-1-7281-4365-1/19/$31.00 ©2019 IEEE 513

Authorized licensed use limited to: Hochschule Landshut. Downloaded on December 02,2024 at 09:22:38 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the Third International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC 2019)
IEEE Xplore Part Number:CFP19OSV-ART; ISBN:978-1-7281-4365-1

C. Reducing The Storage Space

A. Security And Privacy Issues In reducing the storage space the fog computing will be the
top-notch choice to avoiding irrelevant or unrelated data to
Data security is the primary issue associated with non-
move over the total platform [16]. This can reduce the storage
public records and private records of corporations. The client
space and decrease the Postpone.
should depend upon the cloud service provider to access
his/her information D. Support For The Internet of Things
The outline of fog computing may be carried out in a
B. Recovery And Lower Backup different field, the Internet of Things (IoT) is one of its gifts.
The fog computing has an important task in IoT applications
Data safety is a crucial component of the cloud. A few
as the autonomous vehicle, intelligent traffic signal
times in disaster situations the recuperation system is quite
scheduling.
sluggish

C. Hardware Issues VII. FOG COMPUTING VS CLOUD COMPUTING

Immense pace net should be provided to handle this The concept of cloud and fog computing are similar but
complex machine so some technical problems getup at some still, [8] they vary in some aspects they are represented in the
time of high demand below table.
Table 1. parameter wise comparisons of cloud
D. High Latency
and fog computing
For put off the latency time, switch to fog which allows
retaining the information near the customers in preference to PARAMETERS CLOUD FOG
storing them in data centers which are far.
COMPUTING COMPUTING
E. Information Isolation Location of server In under the Near the local
Mostly Data isolation issue emerges in the multi-occupant Nodes internet Network edge
utilization mode, where the distinctive clients' virtual
machines are co-situated on the equivalent hard disks or a
Distance from Multiple number Only one hop
similar server. Here the dangers incorporate to legitimately
Client and server hops
isolate capacity or memory between various clients.
Computing Medium Low
F. Information Lock-In
Latency
The absence of standard APIs confines the relocation of Delay Jitter High probability Very low
utilization and administrations between clouds. With the probability
ascent of the cloud, the issues of data movement, portability
Security in Less secure, More secure,
and merchant lock-in circumstance will increment
computing Undefined Can bedefined
Location Not available available
VI. ADVANTAGES OF FOG COMPUTING OVER CLOUD Awareness
COMPUTING
Vulnerability High probability Very low
A. Solid Distributed Platform
probability
Fog computing gives an aspect to the platform, which is
Geographical Centralized Dense and
present at the numerous locations to extend directly for the
cloud services. This solid distributed platform infrastructure distribution Distributed
facilitates to address and examine the massive information Number of Few Very large
quickly, and travel along the complete wide area network server nodes
might be constrained due to controllers can guide vicinity- Real time Supported Supported
based total demands [14].
interactions
Kind of last mile Leased line Wireless
B. High Support For Portability connectivity
There is fantastic growth in the number of records as well Mobility Limited support Supported
as gadgets. The fogging machine helps for managing the
massive facts and records and give higher as well as the
quicker manner for getting admission to and to examine the VIII. CHALLENGES
information.
The some of the challenges of fog computing are discussed
as follows.

978-1-7281-4365-1/19/$31.00 ©2019 IEEE 514

Authorized licensed use limited to: Hochschule Landshut. Downloaded on December 02,2024 at 09:22:38 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the Third International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC 2019)
IEEE Xplore Part Number:CFP19OSV-ART; ISBN:978-1-7281-4365-1

A. Virtualization Technology Selection D. Security and Privacy

Virtualization is the primary strategy to give hypervisor We concede that security and protection ought to be
virtualization method while para drop secluded conditions in considered in each phase of the fog computing stage structure
fog computing and furthermore the primary factor of fog node [3]. What's more, we view it as one of the greatest difficulties
execution. Along these lines, an instinctive inquiry here is faced by fog computing. To conquer this, we have to apply get
"hypervisor to the container, which one should we pick?" As to control and interruption location framework, which need
we probably are aware [9], cloudlet is using is utilizing a support from each layer of the stage.
lightweight arrangement: a container for example OS-level
virtualization. The plan decision is made in an unexpected
way since their equipment types have diverse abilities. Be that IX. FOG COMPUTING APPLICATIONS
as it may, one weakness against container-based virtualization
A. Smart Home
is the loss of adaptability [10]. For instance, it can't have
various types of visitor working frameworks on a solitary The advancements in the Internet of Things had lead to
foundation hub. Hence, we lean toward hypervisor raise in the number of smart gadgets used in the home. The
virtualization systems as opposed to container-based gadgets from various sellers are difficult to work together, but
virtualization. some jobs which need a large quantity of analyzing,
computing, storing the data. Example: video analysis. To
overcome these difficulties we introduce fog computing, it
B. Latency Battle combines all components into one network and runs all smart
There are numerous components presenting high home programs [17]. To use domestic safety application, for
idleness of use or service execution on fog computing stages. instance, extensively deployed sensors encompass the video
High dormancy will demolish the client experience recorder, audio recorder etc. The fog computing also gives
furthermore, fulfillment, since fog computing is focusing at strong security. If the fog computing is installed once every
delay-delicate applications and services [11]. There are a few sensor is connected to the user, then the respective software is
potential outcomes to acquire inactivity in fog computing: installed in independent virtual machines.

Information conglomeration: The geo-conveyed B. Autonomous Vehicle

nature of fog computing worldview confirms that there will be The fog computing may be incorporated into automotive
a deferral if information accumulation isn't done before depending on its features. Autonomous vehicle system can be
information handling. In any case, there are numerous differentiated into sort infrastructure primarily conventional
approaches to alleviate this issue, for example, applying drive or self-drive. The fog computing nodes are used to
information apportioning/separating and using region in the communicate the data to/from vehicles. Every fog can
chain of importance to diminish calculation volume on the communicate with the customer in same fog or different fog
higher layer. network. There are many applications depending on vehicular
Resource provisioning: There will be a postponement fog computing like intelligent traffic signal and for sharing
in provisioning resource for specific errands, particularly for protective measures.
asset constrained fog nodes [12]. We may require cautiously C. Secure Medical Status
structured booking by utilizing the need and mobility model.
Node versatility, churn, and failure. fog computing needs Health status control has been touchy trouble given that
to be versatile to node portability, stir, and failure. Both health records consist of precious and personal records thereby
framework screen and area service will cooperate to give data with the help of fog. The user can able to take ownership of
to help on picking moderation methodologies, for example, his medical status. These fitness data can be stored in node
registration, rescheduling, and replication. level in a personal maintaining way which helps him to
directly share this data to authorize the person like a doctor
C. System Management and test labs.
The system the board will be a weight for fog computing D. Face Detection Algorithm
except if we receive the rewards of applying SDN and NFV
methods. Be that as it may, the consistent mix of SDN and By using fog computing we can design the application for
NFV into fog computing isn't simple, also, will absolutely be a face detection. In this first user should download the
major test. The troubles come from re-structure the south- application in his/her smart phone and he should take his/her
bound, north-bound and furthermore the east-west-bound APIs photo by using this application. Then this information is sent
to incorporate important fog computing natives. Credulous and stored in a remote server. While using face recognition to
reconciliation may not meet the plan objectives of idleness and unlock his/her device then it will match the current photo with
effectiveness. the previous photo and photo in the remote server, which you
gave at starting if it matches then your device will be unlocked
Since the fog computing gives low latency it is used to build
this application.

978-1-7281-4365-1/19/$31.00 ©2019 IEEE 515

Authorized licensed use limited to: Hochschule Landshut. Downloaded on December 02,2024 at 09:22:38 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the Third International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC 2019)
IEEE Xplore Part Number:CFP19OSV-ART; ISBN:978-1-7281-4365-1

Table 2. Face recognition performance using Fog computing XI. REFERENCES

Type of computing Face recognition Response time in [1] Thahiris De Souza Alvera et.al, “From Knowledge Based Vision
time in ms ms Systems to Cognitive Vision Systems: A Review, International
Conference on Knowledge Based And Intelligent Information And
Cloud 2.502 900.869 Engineering Systems, Kes2018, 3-5 September 2018, Belgrade,
Serbia 34-40.
Fog 2.491 167.771 [2] Chen Y et.al, “IBM Watson: How Cognitive Computing Can Be
Applied to Big Data Challenges in Life Sciences Research.” Clin
there April 2016 Vol.38(4) Pp: 688-701
E. Augmented Reality [3] Handers et.al, “How and why aff ective and reactive virtual agents
will bring new insights on Social cognitive disorders in
The applications of augmented reality are profoundly schizophrenia? An illustration with a virtual card Game paradigm.”
latency and intolerance uniform exceptionally little Feb 26, 2015 - Published online 2015 Mar 30. doi:
postponements accordingly can harm the client experience. 10.3389/fnhum.2015.00133. PMCID: PMC4378306
Subsequently, fog computing can possibly turn into a [4] Rodrigo Alberto Britto et.al, “The Impact of Flight Delays on
Passenger Demand and Consumer Welfare” Researchgate.net,
noteworthy performer in the Augmented reality space. An 2016, Pp 1-28.
augmented brain-computer interaction game dependent on a [5] Hsien-Tsung Chang et.al, “Automatic Travel Itinerary Planning
linked data and fog computing. At the point an individual System for Domestic Areas” Comput Intell Neurosci. 2016; 2016:
plays the game, crude surges of information gathered by EEG 1281379. Published online 2015 Dec 29. doi:
10.1155/2016/1281379
sensors are produced and grouped to identify the mental [6] Catherine Sundling, “Travel Behavior Change in Older Travelers:
condition of the player. Mind state arrangement is among the Understanding Critical Reactions to Incidents Encountered in Public
most computationally substantial signal handling tasks, yet Transport”, Int. J. Environ. Res. Public Health 2015. Int. J. Environ.
this should be completed in real time. The framework utilizes Res. Public Health 2015, vol 12, Pp14741-14763
[7] Logesh Ravi et.al, “A Collaborative Location Based Travel
both fog and cloud servers, a blend that empowers the Recommendation System through Enhanced Rating Prediction for
framework to perform nonstop continuous mind state the Group of Users”, Hindawi Publishing Corporation
arrangement at the fog servers while the characterization Computational Intelligence and Neuroscience Vol 2016.pp1-11
models are tuned routinely in the cloud servers dependent on [8] Mao Ye et.al, “Modeling the Commuting Travel Activities within
Historic Districts in Chinese Cities”, Hindawi Publishing
the EEG readings gathered by the sensors. Corporation Computational Intelligence and Neuroscience Vol
2016. Article ID 1291358, pp1-28
A wearable cognitive assistance framework dependent on [9] Cong Bai et.al, “Dynamic Bus Travel Time Prediction Models on
google glass gadgets that helps individuals with decreased Road with Multiple Bus Routes”, Hindawi Publishing Corporation
Computational Intelligence and Neuroscience Volume 2015. Article
mental keenness. As a result of the idea of cognitive gadgets ID 432389, ppo:1-9
with constrained assets, the computing concentrated remaining [10] Jin Fan,” Cognitive control in majority search: a computational
burdens of this application should be offloaded to an outside modeling approach” Original Research Article published: 09
server. In any case, this offloading must give fresh, ongoing February 2011 doi: 10.3389/fnhum.2011.00016 pp1-10
[11] Tomasz Smole et.al,” Most evidence for the compensation account
reactions, neglecting to do which will be impeding to the of cognitive training is unreliable” Memory & Cognition 2018,
client experience. Offloading the process concentrated tasks to vol.46: pp:1315–1330
the cloud brings about an impressive latency, in this way the [12] Blanka Klimova “Computer-Based Cognitive Training in Aging”,
author makes utilization of adjacent gadgets. These gadgets MINI REVIEW December 2016 Vol 8, Article 313, pp1-6
[13] Y. Mohana Roopa et.al,” Component-based Self-adaptive
may speak with the cloud for deferral tolerant jobs like error Middleware Architecture for Networked Embedded Systems”,
detailing and logging. The previously mentioned works are International Journal of Applied Engineering Research ISSN 0973-
regular uses of fog computing in that they perform latency 4562 Vol 12, Number 12 (2017) pp. 3029-3034
critical examination at the very edge and idleness tolerant [14] Y. Mohana Roopa et.al “Road and Traffic Enforcement System
Using GPS Enabled Mobile Cloud Computing” Indian Journal of
calculation at the cloud – accordingly depicting fog as an Science and Technology, Vol11, issue no44, pp-1-6
expansion of cloud. [15] Y. Mohana Roopa, et.al,” Fog Computing for Smart Cities”, IEEE
International conference Communication and Electronics Systems
ICCES 2019 on July17-19, 2019 organized by PPG Institute of
X. CONCLUSION Technology, Coimbatore, India. ISBN: 978-1-7281-1261-9, Page
No: 910-915.
In this research work, we have given a brief overview [16] Kumar, T. S. (2019). Efficient resource allocation and qos
on fog computing and then we have discussed about the fog enhancements of IoT with fog network. Journal of ISMAC, 1(02).
computing architecture that explains about the layers of fog 101-110.
[17] Karthiban, K., and S. Smys. "Privacy preserving approaches in
computing architecture. Then we have introduced the cloud computing." In 2018 2nd International Conference on
limitations of cloud computing and how fog computing solves Inventive Systems and Control (ICISC), pp. 462-467. IEEE, 2018.
the drawbacks of cloud computing. As fog computing is the
emerging technology we had many issues to be solved I order
to incorporate it in many applications like smart home, smart
grid, virtual machine, health care, autonomous vehicle, and
face recognition.

978-1-7281-4365-1/19/$31.00 ©2019 IEEE 516

Authorized licensed use limited to: Hochschule Landshut. Downloaded on December 02,2024 at 09:22:38 UTC from IEEE Xplore. Restrictions apply.