Académique Documents
Professionnel Documents
Culture Documents
AbstractOver the past few years, the world is being introduced to the fascinating emerging technologies such as IoT (And
Security), Cloud Computing, Data Analytics. Internet of Things is
the network of embedded devices (sensors,actuators,etc.) with the
objective of collecting and exchanging data anytime, anywhere,
always. So many connected devices produces large chunks of
data and interaction of data which demands and seeks solution
to cloud technology (Solution to processing large amount of data
with the help of virtual server). Moreover, with the increasing
penetration of Internet in human life, we are producing zetta
bytes of data carrying meaningful insights. Data analytics allows
us to process this raw data, analyze it and come to meaningful
conclusion with opportunities. Providers of all these services (and
others as well) deal with discrepancy of demand and capacity.
Thus, the capacity management is required with the help of
performance modeling.This survey paper covers wide view of all
these emerging technologies and the challenges associated with it.
It is very important to consider the challenges and cost of those
challenges in these emerging technologies, because we are going
to emerge along with these technologies. This paper contains
enlightening view of challenges from which one can have more
insights to the solution to it and to the technology itself.
I. I NTRODUCTION
IoT: It is the network of physical objects embedded with
electronics, actuators, softwares, sensors etc. The objective is
to enable this network of objects to collect and exchange data
with 3As (Always, Anywhere, Anytime) of IoT. The devices
are called thing with environment of Devices, Network
and Application (DNA of Internet). It gives opportunity to
integrate physical world into computer based systems. In
environment, communication systems are embedded with
RFID, WSN, Bluetooth, Wi-Fi and Cloud Technology.
Data Analytics: Data analytics is about processing
raw data, modeling and reaching to meaningful and useful
conclusion that could help in Decision making and strategy
formation. Too many devices connected to network produces
very large chunks of data with hidden perception which
makes it important to process that BIG DATA to get the
insights of a system and to explore new possibilities with the
derived conclusions.
B. Data Analytics
Cognitive biased data may lead to unreliable conclusions.
BIG Data comprises of huge chunks of information with lot
of variety. Every bit of information may not be neutral.
Computational capacity and storage: Big Data demands
huge processing power and storage for incoming data as well
as resultant data. Moreover, the response time should be high
(Many organization asks for conclusions within few days of
giving data to data analysts). This real time demand in both
ways is very crucial.
Processing Speed: Requirement of real time data
visualization demands fast processing of huge chunk of
data. Processing speed will be compromised with increasing
rate of big data.
Heterogeneity of data makes it more complex to clean
(preprocessing), processing, modeling and visualization.
In general, the quality of dataset is inversely proportional to
its size. Big data might not always carry meaningful insights.
Huge data volume, huge variety and high velocity of data are
major issues with Big data
No guarantee of 100 percent accurate conclusions (Or
predication). Because of stochastic nature of real world,
conclusions might not be absolutely reliable (e.g. Weather
forecasting, election results based on social media activities
of voters)
? Cost of Challanges:
Inaccurate conclusion through DA may do large impact on
its applications. Reliability could be at stake when the data
is untrustworthy or because of the faulty processing of BIG
DATA.
Huge data sets makes every relationship significant which
could affect the probability of hypothesis testing.One of the
Vs in BIG DATA is variety. The more variables are there
in the data makes it very difficult to model and come to a
concrete conclusion.
? Cost of Challanges:
Vulnerability of network and synchronization issues largely
impacts on security and reliability of IoT.
Issues with communication protocols raises questions on universal acceptance of IoT because of the residing compatibility
issues
Human-in-loop incorporation could be very crucial when
it comes to health care,pervasive surgery, automobile. Dependency is at stake as the system is not programmed to deal with
change or learn effectively covering all the permutation and
combination.
? Cost of Challanges:
Unreliable future forecasting of demand or service
level requirements may lead to unnecessary and huge
investments, causing compromised spending in other
necessary development. 90 percent of real time traffic is
Markovian distributed which could lead to nondeterministic
performance modeling. There will always be trade-off
between the different solution approach between technology
view and business view. These two different approaches could
create compromised system level investments.
D. Cloud Computing
Security and privacy of data is major concern on two
levels. On user side as well as on the service provider
(Storage side). Both ends are vulnerable to the security threat.
(e.g. leaked celebrity images from Apple iCloud on 31ts
August, 2014)
Performance and bandwidth cost: Cloud computing demands
very high bandwidth becuase of the virtual server are involved
as far as storage is concerned. Moreover, in Public cloud
too many users accessing simultaneously requires very high
bandwidth.
Software licensing. Companies have to allow their customers
to use the software in cloud environment. Companies can
either sale their product in environment which is dependent
on the number of machines or can use Cloud licensing asking
for verification by vendors.
Government regulations (Migration of BI solutions of
companies to a cloud infrastructure outside the border)
? Cost of Challanges:
Money can be saved on business maintenance but for
data-intensive applications money needed for bandwidth is
high. Applications which collects or exchange huge data
requires high bandwidth for less response time and Qos.
In order to make cloud more secured (Which is required)
the access time decreases. As cloud is huge network,
the layered security approach could be one option.Thus,
query goes through different security requirements (firewall,
encryption,decryption, 2-3 step verification processes). These
security requirements decreases the access time of the service.
R EFERENCES
[1] Perera, Charith, et al. A survey on internet of things from industrial
market perspective. Access, IEEE 2 (2014): 1660-1679.
[2] John A. Stankovic Research Directions for the Internet of Things.
IEEE