Accelerat ing t he world's research.

A Review-Fog Computing and Its

Role in the Internet of Things
IJERA Journal

Related papers Download a PDF Pack of t he best relat ed papers 

T he Int ernet of T hings (IoT ): An Overview

Ant ar Abdul-Qawy, E. Magesh

Cyber Physical iot

August o Gomes

Design and Simulat ion of a Smart Home managed by an Int elligent Self-Adapt ive Syst em
IJERA Journal
Vinod Pande et al. Int. Journal of Engineering Research and Application www.ijera.com
ISSN : 2248-9622, Vol. 6, Issue 10, ( Part -4) October 2016, pp.7-11

RESEARCH ARTICLE OPEN ACCESS

A Review- Fog Computing and Its Role in the Internet of Things

Vinod Pande*, Chetan Marlecha**, Sangramsing Kayte ***
*(Department of Computer Science & IT, Dr. Babasaheb Ambedkar Marathwada University, Aurangabad,
India)
**(Department of Computer Science & IT, Dr. Babasaheb Ambedkar Marathwada University, Aurangabad,
India)
***(Department of Computer Science & IT, Dr. Babasaheb Ambedkar Marathwada University, Aurangabad,
India)

ABSTRACT
Fog computing extends the Cloud Computing paradigm to the edge of the network, thus enabling a new breed of
applications and services. Dening characteristics of the Fog are: a) Low latency and location awareness; b)
Wide-spread geographical distribution; c) Mobility; d) Very large number of nodes, e) Predominant role of
wireless access, f) Strong presence of streaming and real time applications, g) Het-erogeneity. In this paper we
argue that the above characteristics make the Fog the appropriate platform for a number of critical Internet of
Things (IoT) services and applications, namely, Connected Vehicle, Smart Grid , Smart Cities, and, in general,
Wireless Sensors and Actuators Net-works (WSANs).
Keywords: Fog Computing, Cloud Computing, IoT, WSAN, Software Dened Networks, Real Time Systems,
Analytics.

I. INTRODUCTION the platform that supports Fog services. The

The pay-as-you-go Cloud Computing following section takes a close look at a few key
model is an ecient alternative to owning and applications and services of interest that substantiate
managing private data centers (DCs) for customers our argument in favor of the Fog as the natural
facing Web applications and batch processing. component of the platform required for the support
Several factors contribute to the economy of scale of for the Internet of Things. In the fourth section we
mega DCs: higher predictability of massive aggre- examine analytics and big data in the context of
gation, which allows higher utilization without applications of interest. The recognition that some of
degrading performance; convenient location that these applications demand real-time analytics as well
takes advantage of inexpensive power; and lower as long-term global data mining illustrates the
OPEX achieved through the deployment of interplay and complementary roles of Fog and
homogeneous compute, storage, and networking Cloud. We conclude with comments about the state
components. of the Fog Computing and discussion of future work.
Cloud computing frees the enterprise and
the end user from the speciation of many details. II. THE FOG COMPUTING
This bliss becomes a problem for latency-sensitive PLATFORM
applications, which require nodes in the vicinity to Fog computing, also known as fogging, is a
meet their delay requirements. An emerging wave of model in which data, processing and applications are
Internet deployments, most notably the Internet of concentrated in devices at the network edge rather
Things (IoTs), requires mobility support and geo- than existing almost entirely in the cloud. That
distribution in addition to location awareness and concentration means that data can be processed
low latency. We argue that a new platform is needed locally in smart devices rather than being sent to the
to meet these requirements; a platform we call Fog cloud for processing. Fog computing is one approach
Computing [1], or, briey, Fog, simply because the to dealing with the demands of the ever-increasing
fog is a cloud close to the ground. We also claim that number of Internet-connected devices sometimes
rather than cannibalizing Cloud Computing, Fog referred to as the Internet of Things (IoT).
Computing enables a new breed of applications and In the IoT scenario, a thing is any natural or
services, and that there is a fruitful inter-play man-made object that can be assigned an IP address
between the Cloud and the Fog, particularly when it and provided with the ability to transfer data over a
comes to data management and analytics. network. Some such things can create a lot of data.
This paper is organized as follows. In the Cisco provides the example of a jet engine, which
second section we introduce the Fog Computing they say can create 10 terabytes (TB) of data about
paradigm, delineate its characteristics, and those of its performance and condition in a half-hour.

www.ijera.com 7|P age

Vinod Pande et al. Int. Journal of Engineering Research and Application www.ijera.com
ISSN : 2248-9622, Vol. 6, Issue 10, ( Part -4) October 2016, pp.7-11

Transmitting all that data to the cloud and  Large-scale sensor networks to monitor the
transmitting response data back puts a great deal of environment, and the Smart Grid are other
demand on bandwidth, requires a considerable examples of inherently distributed systems,
amount of time and can suffer from latency. In a fog requiring distributed computing and storage

 Very large number of nodes, as a consequence

computing environment, much of the processing resources.
would take place in a router, rather than having to be
transmitted of the wide geo-distribution, as evidenced in
sensor networks in general, and the Smart Grid

 Support for mobility. It is essential for many

III. CHARACTERIZATION OF FOG in particular.
COMPUTING
Fog Computing is a highly virtualized Fog applications to communicate directly with
platform that provides compute, storage, and mobile devices, and therefore support mobility
networking services between end devices and techniques, such as the LISP protocol 1, that
traditional Cloud Computing Data Centers, typically, decouple host identity from location identity,

 Real-time
but not exclusively located at the edge of network. and require a distributed directory system.
Figure 1 presents the idealized information and interactions. Important Fog
computing architecture supporting the future IoT applications involve real-time interactions rather

 Predominance of wireless access.

applications, and illustrates the role of Fog than batch processing.

 Heterogeneity. Fog nodes come in dierent form

Computing.
Compute, storage, and networking
resources are the building blocks of both the Cloud factors, and will be deployed in a wide variety
and the Fog. Edge of the Network, however, implies of environments Interoperability and federation.
a number of characteristics that make the Fog a non- Seamless support of certain services (streaming
trivial extension of the Cloud. Let us list them with is a good example) requires the cooperation of
pointers to motivating examples. dierent providers. Hence, Fog components must
Edge location, location awareness, and low be able to interoperate, and services must be


latency. The origins of the Fog can be traced to early federated across domains.
proposals to support endpoints with rich services at Support for on-line analytic and interplay
the edge of the network, including applications with with the Cloud. The Fog is positioned to play a sign
low latency requirements (e.g. gaming, video cant role in the ingestion and processing of the data
streaming, augmented reality). close to the source. We elaborate in section 4 on the
interplay between Fog and Cloud regarding Big
Data.

A. Fog Players: Providers and Users

It is not easy to determine at this early stage
how the different Fog Computing players will align.
Based on the nature of the major services and

 Subscriber models will play a major role in the

applications, however, we anticipate that:

Fog (Infotainment in Connected Vehicle, Smart

 The Fog will give rise to new forms of

Grid, Smart Cities, Health Care, etc.)

competition and cooperation between providers

angling to provide global services. New
incumbents will enter the arena as users and
providers, including utilities, car manufacturers,
Figure 1: The Internet of Things and Fog public administrations and transportation
Computing agencies.

 Geographical distribution. In sharp contrast to B. Fog Computing and the Internet of Things
the more centralized Cloud, the services and In this section we demonstrate the role the
applications targeted by the Fog demand widely Fog plays in three scenarios of interest: Connected
distributed deployments. The Fog, for instance, Vehicle, Smart Grid, and Wireless Sensor and
will play an active role in delivering high Actuator Networks.
quality streaming to moving vehicles, through The Connected Vehicle deployment
proxies and access points positioned along displays a rich scenario of connectivity and
highways and tracks. interactions: cars to cars, cars to access points (Wi-

www.ijera.com 8|P age

Vinod Pande et al. Int. Journal of Engineering Research and Application www.ijera.com
ISSN : 2248-9622, Vol. 6, Issue 10, ( Part -4) October 2016, pp.7-11

Fi, 3G, LTE, roadside units [RSUs], smart trace The information on is not unidirectional
lights), and access points to access points. The Fog (from the sensors to the sink), but bi-directional
has a number of attributes that make it the ideal (sensors to sink, and controller node to actuators). In
platform to deliver a rich menu of SCV services in a subtler, but sign cant way, it becomes a closed-
infotainment, safety, trace support, and analytics: loop system, in which the issues of stability and
geo-distribution (throughout cities and along roads), potential oscillatory behavior cannot be ignored.
mobility and location awareness, low latency, Latency and jitter become a dominant concern in
heterogeneity, and support for real-time inter- systems that require rapid response.
actions. A smart trace light system illustrates the S.S. Kashi and M. Shari [4] survey the
latter. The smart trace light node interacts locally contributions in the coordination of Wireless Sensor
with a number of sensors, which detect the presence and Actuator Networks (WSANs). They point out
of pedestrians and bikers, and measures the distance that in one architectural choice, the WSAN consists
and speed of approaching vehicles. It also interacts of two networks: a wireless sensor net-work and a
with neighboring lights to coordinate the green trace mobile ad hoc network (MANET). T. Banka et al [6]
wave. Based on this information the smart light stress that emergent applications demand a higher
sends warning signals to approaching vehicles, and bandwidth, collaborative sensing environment. Their
even modes its own cycle to prevent accidents. Re- experience is rooted in the CASA (Collaborative
coordinating with neighboring STLs through the or Adaptive Sensing of the Atmosphere) project.
chest ration layer of the Fog follows any medication CASA [5], a multi-year, multi-partner initiative led
of the cycle. The data collected by the STLs is by UMASS, deployed a network of small weather
processed to do real-time analytics (changing, for radars, integrated with a distributed processing and
instance, the timing of the cycles in response to the storage infrastructure in a closed-loop system to
trace conditions). The data from clusters of smart monitor the lower troposphere for atmospheric
trace lights is sent to the Cloud for global, long-term hazards like tornados, hailstorms, etc. Zink et al [3]
analytics. provide technical details of the deployment.
The characteristics of the Fog (proximity
C. Smart Grid and location awareness, geo-distribution,
Smart Grid is another rich Fog use case. hierarchical organization) make it the suitable
We defer section 4 a discussion of the interplay of platform to support both energy-constrained WSNs
Fog and Cloud in the context of Smart Grid. and WSANs

D. Wireless Sensors and Actuators Networks E. Open Challenges and Future Directions
The original Wireless Sensor Nodes The proposed Cloud centric vision
(WSNs), nicknamed motes [2], were designed to comprises of a flexible and open architecture that is
operate at extremely low power to extend battery life user centric and enables different players to interact
or even to make energy harvesting feasible. Most of in the IoT framework. It allows interaction in a
these WSNs involve a large number of low manner suitable for their own requirements, rather
Bandwidth, low energy, low processing than the IoT being thrust upon them. In this way, the
power, small memory motes, operating as sources of framework includes provisions to meet different
a sink (collector), in a unidirectional fashion. requirements for data ownership, security, privacy,
Sensing the environment, simple processing, and and sharing of information.
forwarding data to the static sink are the duties of
this class of sensor networks, for which the open
source TinyOS2 is the de-facto standard operating
system. Motes have proven useful in a variety of
scenarios to collect environmental data (humidity,
temperature, amount of rainfall, light intensity, etc).
Energy constrained WSNs advanced in
several directions: multiple sinks, mobile sinks,
multiple mobile sinks, and mobile sensors were
proposed in successive incarnations to meet the
requirements of new applications. Yet, they fall
short in applications that go beyond sensing and
tracking, but re-quire actuators to exert physical
actions (open, close, move, focus, target, even carry
and deploy sensors). Actuators, which can control
either a system or the measurement process itself,
bring new dimensions to sensor networks.

www.ijera.com 9|P age

Vinod Pande et al. Int. Journal of Engineering Research and Application www.ijera.com
ISSN : 2248-9622, Vol. 6, Issue 10, ( Part -4) October 2016, pp.7-11

machine-to-machine (M2M) interaction, collects,

process the data, and issues control commands to the
actuators. It also lters the data to be consumed
locally, and sends the rest to the higher tiers. The
second and third tier deal with visualization and
reporting (human-to¬machine [HMI] interactions),
as well as systems and processes (M2M). The time
scales of these interactions, all part of the Fog, range
from seconds to minutes (real-time analytics), and
even days (transactional analytics). As a result of
this the Fog must support several types of storage,
from ephemeral at the lowest tier to semi-permanent
at the highest tier. We also note that the higher the
tier, the wider the geographical coverage, and the
longer the time scale. The ultimate, global coverage
Figure 2: Roadmap of key technological
is provided by the Cloud, which is used as repository
developments in the context of IoT application
for data that that has a permanence of months and
domains envisioned
years, and which is the bases for business
intelligence analytics. This is the typical HMI
Some open challenges are discussed based
environment of reports and dashboards the display
on the IoT elements presented earlier. The
key performance in dictators.
challenges include IoT specific challenges such as
privacy, participatory sensing, data analytics, GIS
based visualization and Cloud computing apart from V. CONCLUSIONS
the standard WSN challenges including architecture, We have outlined the vision and dened key
energy efficiency, security, protocols, and Quality of characteristics of Fog Computing, a platform to
Service. The end goal is to have Plug n‘ Play smart deliver a rich portfolio of new services and
objects which can be deployed in any environment applications at the edge of the network. The
with an interoperable backbone allowing them to motivating examples peppered throughout the
blend with other smart objects around them. discus-sion range from conceptual visions to existing
Standardization of frequency bands and protocols point solution prototypes. We envision the Fog to be
plays a pivotal role in accomplishing this goal. a unifying platform, rich enough to deliver this new
A roadmap of key developments in IoT breed of emerging services and enable the
research in the context of pervasive applications is development of new applications.
shown in Figure:2, which includes the technology We welcome collaborations on the
drivers and key application outcomes expected in the substantial body of work ahead: 1) Architecture of
next decade [7]. The section ends with a few this massive infrastructure of compute, storage, and
international initiatives in the domain which could networking devices; 2) Orchestration and resource
play a vital role in the success of this rapidly management of the Fog nodes; 3) Innovative
emerging technology. services and applications to be supported by the Fog.

IV. ANALYTICS, AND THE REFERENCES

[1]. F. Bonomi. Connected vehicles, the internet
INTERPLAY BE-TWEEN THE FOG
of things, and fog computing. VANET
AND THE CLOUD 2011, 2011.
While Fog nodes provide localization, [2]. M. Bowman, S. K. Debray, and L. L.
therefore enabling low latency and context Peterson. Reasoning about naming systems.
awareness, the Cloud provides global centralization. ACM Trans. Program. Lang. Syst.,
Many applications require both Fog localization, and 15(5):795825, November 1993.
Cloud globalization, particularly for analytics and [3]. G. Forman. An extensive empirical study of
Big Data. We touched upon this point earlier in feature selection metrics for text
reference to smart track light. Here we consider classication. J. Mach. Learn. Res.,
Smart Grid, which data hierarchies help illustrate 3:12891305, Mar. 2003.
further this interplay. [4]. B. Frohlich and J. Plate. The cubic mouse:
Fog collectors at the edge ingest the data a new device for three-dimensional input.
generated by grid sensors and devices. Some of this In Proceedings of the SIGCHI conference
data relates to protection and control loops that on Human factors in computing systems,
require real-time processing (from milliseconds to CHI 00, pages 526531, New York, NY,
sub seconds). This rst tier of the Fog, designed for USA, 2000. ACM.

www.ijera.com 10 | P a g e
Vinod Pande et al. Int. Journal of Engineering Research and Application www.ijera.com
ISSN : 2248-9622, Vol. 6, Issue 10, ( Part -4) October 2016, pp.7-11

[5]. M. J. Sannella. Constraint satisfaction and

debugging for interactive user interfaces.
2003.
[6]. P. Tavel. Modeling and simulation design.
2007.
[7]. wikipedia.org
[8]. H. Sundmaeker, P. Guillemin, P. Friess, S.
Woelfflé, Vision and challenges for
realising the Internet of Things, CERP-IoT
– Cluster of European Research Projects on
the Internet of Things, 20

www.ijera.com 11 | P a g e