Amid heightened competition to sell data analysis software, closel.docx

Amid heightened competition to sell data analysis software,
closely held SAS Institute says it will fend off threats from
rivals IBM, SAP, and Oracle
Jim Goodnight, chief executive of SAS Institute, doesn't mince
words. During a Nov. 10 talk at the Churchill Club, a gathering
of Silicon Valley business people, the head of the closely held
software maker disparaged Wall Street analysts, questioned the
value of marketing staff, and told a joke involving Bill Clinton
and the Pope that some people of faith might have found
offensive.
Goodnight is equally blunt about his competitors in the $9
billion market for data analysis software. In the past two years,
IBM (IBM), SAP (SAP), and Oracle (ORCL) have bought
companies that compete with SAS in providing software that
can predict trends, identify profitable customers, reduce risk, or
shave costs. Goodnight views the mergers as a chance to make
hay. "We always look for any company that's acquired to be
messed up for a year," Goodnight says in an interview.
Goodnight and the rest of SAS will need to keep backing up the
bold talk with results. The company has held up well during a
recession that crimped rivals' growth. At SAS, sales this year
will be unchanged from 2008, when they stood at $2.26 billion,
according to Goodnight. Profit may decline by 4%. By contrast,
revenue at German software giant SAP is expected to decline by
8% this year, and for Oracle, the large U.S. business-software
company, sales slipped 1% for the 12 months ended Aug. 31.
SAS specializes in software that helps companies extract
insights from growing volumes of data. That's attractive because
it is expanding even as demand for other kinds of software, such
as applications that streamline business processes, has declined.
SAS products are deeply embedded
SAS's growth comes partly because it can charge premium
prices for its software, which lets statisticians predict future
scenarios based on historical data. Versions of SAS products

tailored for specific industries can start at $1 million to run a
unit of a company, with annual subscription fees of 20% to 30%
of the initial cost. "That makes the product somewhat pricey,"
compared with other vendors', says Bill Hostmann, an analyst at
market research company Gartner (IT).
Sales of data analysis software increased 22%, to $8.8 billion,
in 2008, according to Gartner. SAS had about 15% of the
market, the same as Oracle's share. SAP held a 24% share of the
market and IBM 11%.
SAS products are deeply embedded in the companies that use
them. The software employs its own programming language and
the statisticians and actuaries who use it are trained on the
system and highly familiar with it. That makes it harder for
customers to switch providers.
IBM is at the forefront of companies hoping to lure SAS clients
that may be looking for an alternative. About a quarter of SAS's
revenue comes from copies of its software installed on IBM
mainframe computers. That gives Big Blue an opportunity to
convert customers to its products. "They're betting they can
poach a lot of SAS's business off the mainframe," Hostmann
says of IBM.
Over the past five years, IBM has spent $12 billion to acquire
13 companies in data analysis and to beef up its own research
and development in that arena. IBM paid $1.2 billion at a hefty
premium for SAS competitor SPSS in a deal announced July
28.In April, IBM formed a group of 4,000 "business analytics"
consultants, which the company expects to yield $2 billion in
services revenues next year.IBM paid $4.9 billion for business
intelligence software maker Cognos in 2007.
Oracle and SAP have been buying, too
IBM's sales force is trying to persuade would-be customers that
Big Blue can outperform SAS in providing the hardware,
software, and consulting services they need to incorporate data
analysis into their operations. "We want to position SAS as a
niche play in statistics and data mining," says Rob Ashe, a
general manager at IBM. "They only cover a small part of the

domain."
Oracle and SAP have also branched into the analytics market to
add revenue growth. In 2007, Oracle bought Hyperion
Solution
s for $3.3 billion, and SAP bought Business Objects for $6.8
billion.
At the low end of the market, SAS also is contending with the
popularity of an open source programming language called R,
which is used by Google (GOOG), Stanford University, and
other prominent organizations to build statistical applications.
SAS has several advantages as it tries to jog sales growth and
fend off competition. Its software tends to be quick for
customers to install and start using, although SAS hires some
consultants from Accenture (ACN) and other firms to help
customers get off the ground. SAS products for small companies
can cost as little as tens of thousands of dollars to install, says
senior vice-president Jim Davis. In 2010, the company's sales of
new software licenses may grow by a double-digit percentage,
although overall revenue might not increase, according to
Davis.
Goodnight scorns buyers, Wall Street
Customers are finding new uses for SAS software, thanks to its
ability to collect data from various sources and compile it for

analysis. Shell uses the software to forecast the life expectancy
of pumps on oil-drilling platforms in the North Sea. Banking
giant HSBC employs it to screen credit-card transactions for
fraud. Department store Kohl's (KSS) bought SAS to maximize
profits from markdowns at its roughly 850 stores. "These are
enormous computational problems," Goodnight says.
SAS is using its $1 billion in cash partly to expand its Cary,
N.C., headquarters. Next year, SAS plans to open a $70 million
data center to expand into software delivered over the Web.
Goodnight, 66, co-founded SAS in 1976 at North Carolina State
University and owns two-thirds of its shares. John Sall, a co-
founder and SAS executive, owns the other third. SAS is worth
$10 billion to $12 billion now, says Goodnight, a billionaire
who in 2009 ranked 33 on the Forbes list of wealthiest
Americans. The CEO is loath to sell. "I've been asked by two
major companies if we'd sell, and I said, 'Yes, for $20 billion,'"
he says. "I have a lot of respect for the people who work at
SAS. I don't want to see them get laid off because the company
gets acquired."
During his Churchill Club talk, Goodnight was even plainer.
"We're still making a lot of money this year. Why should I lay
off 3,000 people so we can say our profits went up?" he said.
"I'm not going to listen to a bunch of 28-year-old kids on Wall
Street telling you what your profit should be down to the
penny."

SAS's independent streak has served the company well for more
than three decades. But the pressure is on Goodnight and his
team to prove that it can navigate a wave of surging
competition.
2019 International Conference on Machine Learning, Big Data,
Cloud and Parallel Computing (Com-IT-Con), India, 14th -16th
Feb 2019
978-1-7281-0211-5/19/$31.00 2019 ©IEEE 129
A Detailed Study of Azure Platform & Its Cognitive
Services
Ankita Verma, Dhutima Malla, Amrit Kaur Choudhary, Vasudha
Arora
Department of Computer Science & Engineering
Manav Rachna International Institute of Research and Studies

Faridabad, India
[email protected], [email protected], [email protected],
[email protected]
Abstract—Windows Azure is the Microsoft's conveyed
registering contraption. In this paper, we will learn about the
prelude to the Azure framework. The scattered enlisting stage
outfits clients with assets and associations for an affiliation. We
would in like way talk about azure associations: Register,
structuring, Stockpile, Fabric Controller Mesh, Cellphone,
Holder, Index, Problem-Solving, A.I + M.L, I.O.T,
Combination,
Identity, Safety, DevOps, Migrate and The executive's aids.
Microsoft’s Azure is the standard circled figuring stage which
draws in the client to send and work assets with speed of
adaptability. Our basic work in this paper is to go up against
recognizing confirmation and face attestation utilizing "Face
API" where we will use our own one of a kind codes, for
example, to perceive and see the powers of the all-inclusive
community.
Keywords— Windows Azure, Resources, Services,
Deployment, Cloud Computing, Face Detection, Face API

I. INTRODUCTION
Conveyed registering is an unused headway. It is the update
of parallel figuring, dispersed enrolling system arranging, and
is the blend and advancement of Virtualization, Utility
dealing with, Software-as-a-Service (SaaS), Infrastructure-as-
a-Service (IaaS) and Platform-as-a-Service (PaaS). In the
method for the most current couple of years, apportioned
enlisting viewpoint has seen a to an incredible degree
extraordinary pass closer to its social occasion and it has
changed into a precedent in the quantifiable improvement
district as it ensures fundamental cost decreases and new
fitting undertaking reachable to its purchasers and
providers.[10] Scattered enlisting is a creation manual for
dessert on and continue programming and is being gotten by
the business, for instance, Google, IBM, Microsoft, and
Amazon. A couple of mannequin purposes and stages, for
instance, the IBM ―Azure Cloud structure, the Google App
Engine, the Amazon Cloud, and the Elastic Computing
Platform. Dissipated making prepared is viewed as the
running with an upgrade that will influence distinctive leveled
affiliations and how they manage their IT establishments. The
progress and organization that cloud connection and
association models offer are a key territory of research.
Appropriated processing uses Azure as a phase for its

distinctive organizations, for instance, Analytics, AI, Machine
Learning, IOT, Face API, etc. and engages customers to send
and work resources with the speed of versatility.
Azure is a foundation for fleeing data in the cloud. Instead of
giving programming that Microsoft customers can present and
run themselves in solitude PCs. Nowadays, Azure is an
association: Customers utilize it to sprint apps and stockpile
facts & figures on web-accessible devices rued by Microsoft
Corporation. Associations are being offered to the
relationships, customers or both by these applications, Azure
Platform is a web-based distributed technology. Extremely
systematic, changeable and compatible capacity can be
utilized to amass unused apps to remain constant for the cloud
or revive live apps with cloud-based cutoff administrations.
Azure is a Distributed framework that gives progress, for
empowering the Microsoft’s Azure Platform.
Azure proceeds running in Microsoft on machine server
farms. The target of Windows to equip engineers with on-ask
for figure & restrain arrange to have, scale, and regulate web
or cloud applications. The arrange itself surrenders and keeps
it exceptionally possess a duplicate of Windows. Designers
based exclusively on making applications that proceed
running on Windows. Windows backings a solid headway
episode by compromising with its IDE. Windows is an open

organize that underpins both Microsoft and non-Microsoft
tongues and conditions. [2] Azure invites outcast disobedient
and vernaculars, for an occasion, Obscure, python, ruby, and
php. Azure gives a Windows-based figuring and constraint
condition within the distributor. Escaping information within
the distributor is a proper choice for some conditions. The
differing bit of Azure collaborates to form this conceivable.
Microsoft’s Azure Tables provide adaptable, accessible, and
firmly collected as collecting tables. A clear interface to
fixing the tables are granted by Microsoft’s Azure. The tables
are versatile to many of the pieces of stuff and terabytes of
realities may separate transversely named records adjacent
metadata. Solid constrain and development of messages for an
application and are the foremost for the most part seen
clarification behind Web and Laborer portion correspondence
are granted by Microsoft’s Azure [3].
In this paper, we are going conversation around the Azure
Components, Cloud Benefit Suppliers, Cloud Architecture,
Azure Cloud Storage Model and Demystifying the Use of
Face API in Windows Azure. By overcoming with the ideas
of Azure Cognitive Services in this paper, we will build up
software in our next paper in which we will put together our
model to identify n number of appearances precisely.

130
II. WINDOWS AZURE COMPONENTS [3]
i. Computer: Windows procedure can run a gathering of
vocations. Whatever an application does, regardless, it needs
to be done as no short of what one occupation. Azure at that
factor typically runs a variety of examples of each movement,
utilizing worked in load changing as per spread needs
crosswise over them.
ii. Storage: The 2nd stage in Azure is a limit. We have three
journalists perfect here – Blobs, that take after records,
Tables, which are entered and well-shaped limit, and lines,
which let Web Parts and Specialist Parts, provide for one
another.
iii. Fabric Controller: The 3rd sort out in Azure is the
connect controller or App connect. It handles articulations and
association – between Microsoft Windows Azure
applications, and in development from the servers. Ready to
express that we got the opportunity to have that expansive
SQL Server or Database structure and enabling access to an
Azure app, & we will not permit customers of an app in our

framework.
CLOUD ARCHITECTURE:
Dispersed considering can be separated along with two
regions, the client and the distributor. By and huge, the client
is associated with the distributor by the methodology for the
net It is additionally productive for a relationship to own an
individual conveyance where a client associates with an
intranet. Regardless, the two conditions are dim other than the
usage of an individual and unbolted framework or
distribution. The customer sends asking the distributor and the
distributor gives the association. Inside the distributor, a focal
server oversees administering the composition and from
different perspectives fills in as the working redirection
organize of the specific distribution orchestrate.
III.TOOLS AVAILABLE EVERY COMPANY BASED ON
CLOUD:
Distributing merchants are passing on the plague of unused
instruments to help adventure IT produce, buy, supervise,

screen, change and track cloud organizations. These gadgets
are proposed to help IT officials free up their budgetary plans
and their staff so both can be used towards progressively key,
a rule of business adventures.
1. Cloud ability: Cloud ability is a money related
administration device for checking and dissecting all cloud
costs over an association.
2. Cloudyn: These instruments are intended to help corporate
IT from over-purchasing Amazon cloud assets.
3. AtomSphere: AtomSphere is a cloud-based programming
Platform as a Service utilized by clients that need to
incorporate their different cloud-based applications with one
another and with on-commence applications.
4. Enstratius: Enstratius gives cross-stage cloud framework
administration for open, private and crossover mists that can
be firmly lined up with an endeavor’s administration and
security prerequisites.
5. Cloud Hub: Conveyed as a bundled incorporation
experience, Cloud Hub and Mule ESB are based on open
source innovation to give brisk, dependable application joining
without seller secure.
AZURE CLOUD STORAGE MODEL:
Azure ability "gives chose, monotonous limit inside the

distributor". Microsoft will likely make amassing that is solid
and secure, flexible and helpful. Azure ability awards us to
keep facts for whenever assignment and to save any level of
bits of awareness. [4] Data can be saved in Windows Azure in
4 unique ways i.e.: Queues, SQL Azure Database, Tables,
Blobs.
Blobs: These are handed-down for coordinated estimations
and take after records on our troublesome drive.
Tables: To draw in support to work with a fact in a much
increasingly basic fine-grained method, Azure storing offers
tables. A table has no described graph or perhaps, properties
can have forms.
Queues: This amassing is used to trade messages between
squares of code.
Figure 1: Windows Azure components [1]
Fig 2. Cloud Architecture [10]
131

SQL Azure Database: We are capable make utilization of
SQL Azure Databases when we require social storing and use
the negotiate-SQL code we are starting at by and by have.
SQL Azure offers a basic subset of SQL Server's
accommodation, which fuses counting, as an administered
cloud advantage.
CLOUD DEPLOYMENT MODEL:
Inside the distributing association show up, sorting out,
orchestrate, storing, and programming foundation is given as
associations that go up or down surprising upon the
captivated. The distributing demonstrate has four guideline
sending models which are:
i. Private Cloud: Private cloud is another term that a few
dealers have begun late used to portray duties that emulate
passed on enrolling on private structures. It is set up inside an
alliance's inward parts wind datacenter. [12] For E.g.
Eucalyptus Frameworks.
ii. Public Cloud: Open cloud depicts dissipated enlisting inside
the customary standard sense, whereby resources are
continuously provisioned on a fine-grained, self-advantage

present over the Web, by construes of web applications/web
relationship, from an off-site page removed provider who
offers resources and bills on a fine-grained utility dealing with
the introduction. [13] For E.g. Microsoft Azure & Google
App Motor
iii. Crossover Cloud: Crossover cloud other than known as
Crossbreed cloud could be an individual merchant identified
with at scarcest one outside circulating affiliations, for the
most part coordinated, provisioned as a specific unit, and
included by a tied down a structure. [14] It gives virtual IT
strategies through a blend of both open and private veils of
mist.
CLOUD COMPUTING SERVICES:
Platform As A service: Platform as a service [4] (PaaS) is a
sending & advancement condition inside the cloud that
conveys straightforward cloud-based applications to complex,
cloud-empowered applications. PaaS is intended to help the
total web application lifecycle of the building, testing,
conveying, overseeing, and refreshing. PaaS incorporates a
total foundation of servers, stockpiles, organizing, and
middleware improvement apparatuses like business

knowledge administrations (BI), database administration
frameworks, and so forth [4].
i. Software As A Service: Software as a service [4] (SaaS)
allows clients to associate with and make use of cloud-based
applications over the Internet. Basic models are email,
calendaring and office instruments, (for example, Microsoft
Office 365). SaaS gives a whole programming association
which you buy on a compensation as-you-go premise from a
cloud specialist co-op.
ii. Infrastructure As A Service: Infrastructure as a service [4]
(IaaS), a moment figuring foundation, provisioned and
oversaw over the Internet. Rapidly scale here and there with
the request and pay simply for what you utilize. IaaS reasons
you to preserve away from the fee and multifaceted nature of
buying and dealing with your personal bodily servers and
another data center framework [4].
iii.
DEMYSTIFYING THE USE OF FACE API IN
WINDOWS AZURE:

The Face API Service, a cloud-based organization that gives
the most dynamic face computations. It is a biometric
programming application arranged to do especially perceiving
or checking a man by differentiating and examining structures
subject to the person's facial structures.
Fig3. Windows Azure Cloud Storage Model [3]
Fig 4. Cloud Deployment Model [4]
Fig 5. Microsoft Cloud Computing Architecture (Source
Microsoft) [4]
Fig 6. Face API in Azure [7]
132

FACE DETECTION:
It perceives some places around one human faces in a
photograph and get returned go up against rectangular shapes
for where inside the photograph the appearances are, close-by
defy characteristics which join M.L-ground needs for gesture
highlights. The defy property features accessible are Age,
Feeling, Sexual introduction, Posture, Grin and Facial Hair
adjacent 27 markers for each stands up to inside the image.
[7] The Face API sees up to 64 personal appearances with
higher accuracy oppose an area in a picture. In like manner,
the picture can be appeared by the record in bytes or huge
URL. Face square shape (left, best, width, and stature)
exhibiting the face district in the picture is returned adjacent
each perceived face. On the other hand, defy area expels a
movement of face-related properties, for instance, present,
sex, age, head present, facial hair, and glasses.
For Example:
Detection result [16]:
JSON:
[
{
"faceId": "2090b697-46a0-497b-95d0-4a965f6dfffd",
"Rectangleface": {

"top": 184,
"left": 358,
"width": 157,
"height": 157
},
"Attributeface": {
"hair": {
"bald": 0.06,
"invisible": false,
"hairColor": [
{
"color": "brown",
"confidence": 1.0
},
{
"color": "black",
"confidence": 0.86
},
{
"color": "red",
"confidence": 0.51
},
{
"color": "other",
"confidence": 0.07

},
{
"color": "gray",
"confidence": 0.05
},
{
"color": "blond",
"confidence": 0.04
}
]
},
"smile": 1.0,
"headPose": {
"yaw": 3.2,
"pitch": 0.0,
"roll": 12.1,
},
"age": 28.0,
"gender": “female”,
"facialHair": {
"moustache": 0.0,
"beard": 0.0,
"sideburns": 0.0
},

"glasses": "No Glasses",
"makeup": {
"eyeMakeup": true,
"lipMakeup": true
},
"emotion": {
"anger": 0.0,
"contempt": 0.0,
"disgust": 0.0,
"fear": 0.0,
"happiness": 1.0,
"neutral": 0.0,
"sadness": 0.0,
"surprise": 0.0
},
"occlusion": {
"foreheadOccluded": false,
"eyeOccluded": false,
"mouthOccluded": false
},
"accessories": [],
"blur": {
"blurLevel": "low",
"value": 0.08
},

"exposure": {
"exposureLevel": "goodExposure",
"value": 0.57
},
"noise": {
"noiseLevel": "low",
"value": 0.29
}
},
"faceLandmarks": {
"pupilLeft": {
"x": 405.7,
"y": 220.9
},
"pupilRight": {
"x": 477.4,
"y": 237.3
},
"noseTip": {
"x": 439.9,
"y": 271.3
},
"mouthLeft": {
"x": 394.1,
"y": 287.8

},
"mouthRight": {
"x": 463.9,
"y": 304.6
Fig 7. Example of Face Detection [16]
133
},
"eyebrowLeftOuter": {
"x": 378.2,
"y": 198.9
},
"eyebrowLeftInner": {
"x": 432.8,
"y": 210.6
},
"eyeLeftOuter": {
"x": 392.7,
"y": 219.4
},

"eyeLeftTop": {
"x": 408.0,
"y": 217.8
},
"eyeLeftBottom": {
"x": 404.0,
"y": 227.2
},
"eyeLeftInner": {
"x": 418.9,
"y": 226.2
},
"eyebrowRightInner": {
"x": 467.8,
"y": 219.6
},
"eyebrowRightOuter": {
"x": 509.9,
"y": 227.0
},
"eyeRightInner": {
"x": 467.0,
"y": 237.6
},
"eyeRightTop": {

"x": 483.1,
"y": 234.0
},
"eyeRightBottom": {
"x": 479.8,
"y": 245.1
},
"noseLeftAlarOutTip": {
"x": 413.7,
"y": 265.0
},
"noseRightAlarOutTip": {
"x": 465.0,
"y": 277.2
},
"upperLipTop": {
"x": 433.8,
"y": 291.3
},
"upperLipBottom": {
"x": 432.0,
"y": 297.4
},
"underLipTop": {
"x": 426.4,

"y": 309.5
},
"underLipBottom": {
"x": 423.2,
"y": 321.7
}
}
}
]
FACE/EMOTION RECOGNITION:
Face affirmation is extensively used in various circumstances
including security, regular UI, picture content examination
and organization, convenient applications, and apply self-rule.
Four face affirmation limits are given: defy affirmation,
finding equivalent appearances, stand up to the social event,
and individual recognizing proof [7] .
The Face API as of now additionally organizes feeling
affirmation, reestablishing the sureness over a relationship of
affections for each face in the image, for instance, shock,
disdain, irritate, fear, euphoria, neutral, pain and surprise.
These feelings are grasped to be differently and all around
talked with exact outward appearances.
For Example:

Detection result [16]:
1 face detected
JSON:
[
{
"faceRectangle": {
"top": 56,
"left": 123,
"width": 59,
"height": 59
},
"scores": {
"anger": 6.932031E-08,
"contempt": 1.34474765E-09,
"disgust": 9.843048E-08,
"fear": 1.65876812E-10,
"happiness": 0.9999997,
"neutral": 4.23310347E-08,
"sadness": 2.851677E-10,
"surprise": 1.01985542E-07
}

}
]
1. Face Verification: Face API affirmation plays out a
check against two distinguished appearances or approval
from one perceived face to one individual question [7].
2. Finding Similar Faces: Given a goal recognized face and
a game plan of cheerful appearances to look for with, the
organization finds a little course of action of
countenances that look most like the objective face [7].
3. Face Grouping: Given one course of action of cloud
faces, defy gathering API normally parcels them into a
couple of social affairs considering closeness. [7] Each
get-together is a mixed-up suitable subset of the main
darken confront set and contains relative faces. Also, all
Fig 8. Example of Face Recognition [16]
134

the appearances in a comparable get-together can be
considered to have a place with a comparative individual
inquiry [7].
4. Face Identification: Face API can be used to perceive
people in perspective of a recognized face and a people
database (portrayed as a Large Person Group/Person
Group). Make this database early, which can be adjusted
after [7].
IV. CONCLUSION
Scattered figuring a massive change to web age; it felt the
invigorating impact from the upstream of the industry to
clients. Windows Azure gives an Aperture figuring and
control condition inside the cloud. In this paper, we find that
how they go up against request/perceiving insistence get
into the universe of current advancement. By and large the
upsides of the usage of going up against API inside the
Microsoft Azure that how it makes things present day.
Microsoft's Windows Azure make with its foundation and
stages affiliations will help & steer the improvement culture
move from on-begin application makers to cloud engineers.
Azure gives an orchestrating and constraint condition inside

the distributor. SQL Azure gives a social DBMS inside the
cloud, adjoining sorting out and information
synchronization.
The Windows Azure programming show can be obliging for
anyone who should make less unpredictable to immediate,
increasingly accessible, and progressively flexible
applications.
As Microsoft Azure gives a phase to develop any of its
organizations. Henceforth in the wake of inquisitive about
on Azure, we will furthermore get significant into the
emotional organizations i.e. Face Application Program
Interface (API) Service. It is furthermore isolated into Face
Detection with properties, Face Recognition and Face
Storage. Face Detection attributes are Emotion Recognition,
Face Identification, Similar Face Search and Face Grouping.
We are intending to use these Face API's to develop an
application in which the participation of the whole class can
be taken by just clicking a social event picture.
ACKNOWLEGEMENT
We might want to truly convey our caring gratefulness to Dr.
Prateek Jain, Accendere Information Administration

organizations for having any kind of effect and guiding us in
this paper course of action.
REFERENCES
[1] “Windows Azure Platform: An Era for Cloud Computing”,
Madhurima et al, / (IJCSIT) International Journal of Computer
Science and Information Technologies, Vol. 2 (2), 2011, 621-
623.
[2] Dong Xu, “Cloud Computing: An Emerging Technology”,
International Conference on Computer Design and Applications
(ICCDA 2010), Volume-1, Pgs (100-104).
[3] “Cloud Computing: Windows Azure Platform”, Volume 3,
No. 1,
January 2012, Journal of Global Research in Computer Science.
[4] “Windows Azure Paas Cloud: An Overview”, International
Journal of Computer Application Issue2, Volume 1 (February

2012), Issn: 2250-1797.
[5] Rimal, B., Choi, E., and Lumb, I. (2009). A Taxonomy and
Survey of Cloud Computing Systems. In Fifth International
Joint
Conference on INC, IMS and IDC, pages 44–51. IEEE.
[6] [6] Sujay. R, “Hybrid Cloud: A New Era”, International
Journal of
Computer Science and Technology, Vol. 2, Issue 2, June 2011.
[7] Windows Azure platform:
http://www.microsoft.com/windowsazure/
[8] David Chappell, A white paper on “Introducing The Azure
Services Platform: An Early Look at Windows Azure, .Net
Services, Sql Services, And Live Services”, Oct 2008.
[9] Moother, J.; Bhatt, V. (2009): A cloud computing solution
for
universities: Virtual computing lab: case study of North
Carolina
state university, www.google search engine.
[10] “Cloud Computing: Overview & Current Research

Challenges”,
IOSR Journal of Computer Engineering (IOSR-JCE), ISSN:
2278-
0661, ISBN: 2278-8727Volume 8, Issue 1 (Nov. - Dec. 2012),
PP
14-22,
[11] S. Zhang, S. F. Zhang, X. B. Chen, and X. Z. Huo, “The
Comparison between Cloud Computing and Grid Computing,”
2010 International Conference on Computer Application and
System Modelling (ICCASM), pp. V11-72 - V11-75, DOI= 22-
24
Oct. 2010.
[12] B. R. Kandukuri, R. Paturi V, A. Rakshit, ―Cloud Security
Issuesǁ, In Proceedings of IEEE International Conference on
Services Computing, pp. 517-520, 2009.
[13] A Platform Computing Whitepaper. ―Enterprise Cloud
Computing: Transforming IT.ǁ Platform Computing, pp6, 2010.
[14] Global Netoptex Incorporated. ―Demystifying the cloud.
Important opportunities, crucial choices.ǁ pp4-14. Available:
http://www.gni.com [Dec. 13, 2009].

[15] Vaquero, L.M.; Radero-Merino, L.; Linder, M. (2009): A
break in
clouds towards a cloud definition, SIGCOMM Comput.
Communication Rev. 39, pp.50-55.
[16] Face Application Programming Interface:
https://goo.gl/images/HDtBHJ/smiling_human_faces_image
[17] SQL Azure: http://www.microsoft.com/en-
us/sqlazure/default.aspx
[18] T. Zou, R. Bras, M. Salles, A. Demers and J. Gehrke,
“ClouDiA: a
deployment advisor for public clouds”, Proceedings of the 39th
International Conference on, Very Large Data Bases, pp. 121-
132,
2012.
[19] P. Kalagiakos “Cloud Computing Learning,” 2011 5th
International Conference on Application of Information and
Communication Technologies (AICT), Baku pp. 1 - 4, DOI=12-
14
Oct. 2011.
[20] B. Grobauer, T. Walloschek, and E. Stöcker,

“Understanding
Cloud Computing Vulnerabilities,” 2011 IEEE Security and
Privacy, pp.50-57, DOI= March/April 2011.
International Conference on Advances in Computing,
Communication Control and Networking (ICACCCN2018)
ISBN: 978-1-5386-4119-4/18/$31.00 ©2018 IEEE
64
Virtualization in Cloud Computing- a Study
Manjeet Singh
Directorate of Education
Government of NCT of Delhi, India
E-Id: [email protected]

Abstract— The recent past have experienced tremendous
rise and popularity of technologies like cloud computing
and virtualization. Cloud computing is a construct that
allows us to access applications that actually reside at a
location other than your computer, most often, this will be
a distant data center. Many organizations have already
started implementing these technologies to further reduce
costs through improved resource utilization. Cloud
Computing and virtualization enable users to use
applications on internet and intranet. In Cloud computing
many of the services relies heavily on virtualized
hardware. Because of this, we carefully reviewed existing
virtual hardware services in order to both understand how
the process works and to find the service that is best suited
to Cloud computing and virtualization in different
situations.
Keywords—Cloud Computing; Virtualization; Virtual Machine;
Resource Utilisation; Data Center.
I. INTRODUCTION
In Cloud computing, virtualization is the basis of
delivering Infrastructure as a Service (IaaS) that separates

data, network, applications and machines from hardware
constraints. Cloud computing is becoming popular among IT
businesses due to its agile,flexible and cost effective services
being offered at Software, Platform and In-frastructure level.
Software as a Service (SaaS) allows users to access
applications hosted by different vendors on Cloud via internet.
Platform as a Service (PaaS) enables developers to code, test
and deploy their applications on IaaS. In infrastructure as a
Service (IaaS) model, Cloud provider’s offer services such as
computing, network, storage and databases via internet. IaaS is
the base of all Cloud services with PaaS and SaaS both built
upon it. The primary features of IaaS are elasticity and
virtualization. Virtualization enables a single system to
concurrently run multiple isolated virtual machines (VMs),
operating systems or multiple instances of a single operating
system (OS) [1].
Cloud Suite relies heavily on virtualized hardware.
Because of this, we carefully reviewed existing virtual
hardware services in order to both understand how the process
works and to find the service that is best suited to Cloud Suite.
Virtualization is one of the oldest areas of interest in
computer science. One of the most successful examples of
virtualization is the Java Virtual Machine (J V M) [3]. Using

the J V M allows the Java programming language to produce
identical results regardless of the instruction set architecture
(ISA) that is used by the hardware platform. This is because,
for all intents and purposes, the compiled Java code is running
on the same machine hyphen albeit in a virtual setting. As
long as a system has a J V M installed that can translate Java
byte code into the appropriate ISA, then the results will be
identical.
The Java virtual machine allows one piece of virtual
hardware to be present on multiple discrete systems. Software
like V M Ware allows multiple virtual hardware systems to be
present on a single piece of physical hardware. Creating
multiple virtualized hardware images on a single piece of
physical hardware is one of the cornerstones of cloud
computing. These concurrent virtual images are made possible
by a hypervisor. Various types of virtualization exists such as
Client virtualization, Storage virtualization and Server
virtualization in cloud computing.
II. NEED, BENEFITS AND CONS OF VM MIGRATION
VM Migration is at the core of cloud computing. The
needs and benefits of virtualization are as follows [2]:

Multiple Systems: With virtualization we can run
multiple applications even on different operating
system on the same physical hardware.
Power Management: We can improve the power
utilization.
Improve Resource Utilization: The resource utilization
is increased with the virtualization.
Load Balancing: Load can be distributed on available
resources.
Maintenance and Servicing.
It improves system availability.
It is scalable, as per the requirement with time.
Security is provided by running multiple servers in
isolation.
The major hurdle with the virtualization is that the network
load during the migration increases which results in the

increase of communication cost. One side effect of this is that
performance of application running on virtual machine
degraded.
III. VIRTUAL MACHINE PLACEMENT TECHNIQUES
The various algorithms implemented earlier for VM
Placement in data centers are First Fit, Next Fit, Random Fit,
65
Least full First, Most Full First, First Fit Decreasing (Single
Dimension), First Fit Decreasing Product, First Fit Decreasing
Sum, Dot Product, Minimizing Angle. First fit, next fit and
random fit algorithms are simple algorithms. These algorithms
do not consider the resource capacities needed for placing
virtual machine and not consider the resource consumptions in
physical machines. Least full first, most full first, first fit

decreasing (single dimension) algorithms are designed based
on CPU capacity. First fit decreasing product, first fit
decreasing sum and dot product algorithms are designed for
multi dimensional resources. In these algorithms,
physical/virtual machines’ resource volumes are calculated
using all resources capacities. In minimizing angle algorithm
all resources are involved directly for placing virtual
machines. In this algorithm different resource consumptions
are balanced so that physical machines are minimized. This
algorithm concentrates on balancing resources instead of
maximizing resource usage of physical machines in each
dimension. After all these Virtual Machine Placement
algorithms, Modification of the Best Fit Decreasing (BFD)
algorithm is implemented in cloud data centers for placement
of Virtual Machines.
IV. SOME MAJOR CLOUD AND VIRTUALIZATION
SERVICE PROVIDER
Google (Google App Engine)
Google App is great for all types of applications such as
Business, Consumer, Marketing, Mobile, and Website [4].
Easy to Build: By using the familiar development

tool it will allow the user to buildup and get started
app quickly
Easy to Scale: It can serve any number of users and
automatically responds to data traffic change.
Easy to Manage: Administer app through the
centralize app web console always actives and
works, can’t need software installation, server or
backup etc.
Easy to Storage: GAE advanced storage
infrastructure gives benefits to companies to store
and access their data
Highly Scalable.
Flexible Security and sharing.
Fast data access.
Reliable Storage Access.
Table 1: Comparison of service providers over various matrices
Google [4] Amazon [6] Microsoft [5]
Cloud

Provider Google App Engine Amazon EC2 Azure
Service
Mode PaaS /SaaS Provider PaaS/ SaaS Provider and IaaS PaaS/
SaaS Provider and IaaS
Develop
Support Web service
Web service;
Windows console application;
Virtual Instance
Virtual machine
Feature Easy to build; Easy to scale; Easy to manage,
inexpensive.
Elastic; Completely Controlled;
Flexible; Designed for use with
other Amazon Web Services
Reliable; Secure; Easy to Start.
Always up. Always on; Open;

Unlimited servers. Unlimited
storage. Powerful Capabilities.
Compute
model[65]
Predefined application structure and
framework; programmer-provided
“handlers” written in Python, all
persistent state stored in MegaStore
(outside Python code);
Automatic scaling up and down of
computation and storage; network
and server failover; all consistent
with 3-tier Web app structure
x86 Instruction Set Architecture
(ISA) via Xeon VM; Computation
elasticity allows scalability, but
developer must build the machinery,

or third party VAR such as Right
Scale must provide it
Microsoft Common Language
Runtime (CLR) VM; common
intermediate form executed in
managed environment;
Machines are provisioned based
on declarative
descriptions (e.g. which “roles”
can be replicated); automatic load
balancing
Storage MegaStore/BigTable
DynamoDB;SimpleDB;Rel ational
Database Service (RDS);
ElastiCache
Azure storage service

Support
SQL Cloud SQL
-
SQL Azure
Availability 100% 99.97% 99.95%
Free Usage Limitation usage in 1 GAE unite Micro-Linux-
instance 3 month for free
Table 1 shows comparison of these service providers over
various matrices.
66
Microsoft (Windows Azure)

Microsoft windows azure is an open cloud platform
which enable user to build deploy and manage applications
across global network of Microsoft datacenters. The user can
build applications by using any kind of language, tool or
framework.net, node.js, java, php, it give facility to its
customer that whatever language you want to install you can
install and use it [5].
Always up and always on: Microsoft window azure
claim that its deliver services 99.95% to its users and
enable its users to build and run highly available
application without focusing on its infrastructure.
Open: MWA is open source, it give facilities to its
customer for using any kind of language, framework,
or tool to build their applications.
Unlimited Server and Unlimited Storage: MWA
enable its customer to easily scale their applications
to any size. Window Azure is available on multiple
datacenters around the world, which enable its users
to deploy their application near to their customers.
Powerful Capabilities: MWA can deliver flexible
cloud computing platform which can assure any

application need. Its distributed caching and CDN
service allow user to minimize latency and deliver
great application performance anywhere in world [5].
Amazon EC2(aws.amazon.com)
Compute: Amazon is scale to meet user application
demands whether it’s one server or large cluster [6].
Amazon Elastic Compute Cloud (EC2).
AEC is web service which provides resizable
compute capacity in cloud. Amazon EC2 web service
allows users to obtain and build up with minimal
friction. Amazon EC2 provides complete control to
their users for computing resources and let them run
on Amazons computing environment.
Networking: In networking Amazon customize and
control user network resources, both inside and
outside the cloud [6].
V. VIRTUAL LABS
Creating a cloud based ad hoc laboratory system, such as

Cloud Suite, requires a variety of different technologies. The
primary technology that is necessary for Cloud lab is hardware
virtualization. Fortunately there are numerous examples
available to us. We will perform an examination of hardware
virtualization services in general and a specific exploration of
the services offered by Amazon.com. In addition to
virtualization we will also explore: existing virtual lab
environments, templating systems, remote storage solutions,
and finally, A P Is and Frame Works [7].
i. Amazon.com Elastic Compute Cloud
On August 25, 2006, Jeff Bezos announced the limited
beta of Amazon Elastic Compute Cloud [8], referred to as
Amazon E C 2. Amazon E C 2 is a service that allows users
to, essentially, rent computational resources on an as needed
basis. This type of service is made possible by using the Xen
hypervisor, Xen allows Amazon E C 2 to create Virtual
Private Servers as demand requires. These virtual servers
make use of operating systems that have been optimized for
paravirtualization. Once the end user has customized the
virtual machine to their liking they are then able to save it as
an Amazon Machine Image, or A M I.

An A M I is a system image that has been created to be
used with E C 2. Users may customize these A M Is with their
own software packages allowing an A M I to be tailored for
any task. The Amazon Simple Storage Service (S 3) [9] is
used extensively in the code written for Cloud Suite. S 3
allows a user to upload and download files using H T T P
requests. Amazon Simple Storage Service, like the rest of
Amazon Web Services, is not free semicolon however at the
time of writing, Amazon offers a free tier of service available
for the one year [9].
Amazon, and various 3rd parties, provide S D Ks that
allow developers to interact with Amazon Web Services. The
framework for Cloud Suite is written using S D Ks for P H P,
Python (Boto) and Unix forward slash Linux command line.
However S D Ks exists for multiple platforms and languages.
The provided S D Ks greatly ease interaction with the cloud
services provided by Amazon.
ii. Google Compute Engine
Amazon is far from the only available option for cloud
services. Google recently announced ‘Google Compute
Engine’ [10] , a service similar to Amazon E C 2, in addition
to the long standing Google App Engine. In the next section

we will take a closer look at cloud services offered by
providers other than Amazon.com.
iii. Heroku
Heroku {11] deservers special consideration because it is
one of the best examples of a platform for executing arbitrary
programs in the cloud. Heroku allows the end user to develop
an application and leave the running of that application up to
Heroku. The developer pushes code to Heroku using the Git
version control system. Once the code is present, Heroku
relies upon H T T P requests to interface with the client
application.
iv. Google App Engine
Google App Engine (G A E) [12] is another service similar
to Heroku semicolon however, Google App Engine places a
number of restrictions on what code developers may use.
Unlike an Infrastructure as a Service provider, G A E places
limits on what code may be executed as well as only allowing
the proprietary Google Query Language as a data store.
v. Eucalyptus
Extensive research on cloud technology has been

performed at University of California, Santa Barbara (U C S
B). Two of the most well known products of that research are
Eucalyptus and App Scale.
Eucalyptus is an open source emulator for Amazon E C 2
that can be run on local clusters square [13] . Since the initial
release of Eucalyptus, it has grown to become a commercial
success, and is included in the Ubuntu operating system.
vi. App Scale
67
AppScale is an open source implementation of several
cloud A P Is, including Google App Engine. App Scale can be
deployed over Amazon E C 2, Eucalyptus, or an Ubuntu
image. App Scale uses virtualization to provide a uniform

experience across multiple cloud platforms.
vii. Coursera - Online Education
With the growth of high speed internet access to distance
learning has become increasingly more accessible. In addition
to universities supplementing their courses with online
offerings, many universities are also making classes available
online for free. Coursera is a central location where those
seeking to improve their education may participate in free
online classes from sixteen universities including: Caltech,
Stanford, and Princeton. Additionally many universities have
begun offering free classes online for no credit. One such
offering from University of California, Berkley made use of
Heroku to allow students to publish a web based application
and see results in real time. The Massachusetts Institute of
Technology also offers free online courses. It is important to
note that these online courses typically do not confer any sort
of degree or certificate, Coursera being notable in that some of
it ‘courses do offer an electronic certificate of completion
[14]. Courses offered online could make great use of a service
like Cloud Suite to provide their students with an interactive
lab component.
viii. Khan Academy

Khan Academy is a not for profit online educational
resource. Khan Academy offers videos explaining a variety of
topics. In addition to videos Khan Academy also offers a web
based set of tools for developing mathematical proficiency. In
late 2012 Khan Academy introduced Computer Science
curriculum that offers programming instruction and
demonstration. The programing curriculum is based on the
Java Script language and is overseen by John Resig [15]. The
program is designed to demonstrate the fundamentals of
computer science to someone with little to no experience.
ix. University of Hawaii Virtual Lab
In 2007 the Department of Educational Technology at the
University of Hawaii at Mano published a paper detailing an
experiment where an online biology class made use of a C D
hyphen ROM based virtual wet lab. The study compared the
use of a virtual lab to what the study referred to as a ‘face to
face’ lab. The data analysis showed that, while students found
the virtual labs to be useful, face – to -face lab time was more
valuable to the overall learning experience than a purely
computer-based lab. The study was primarily focused on the
effectiveness of the virtual lab and did not give great detail on
the technology [16].

x. Navy Virtual Lab
The Naval Postgraduate School (N P S) [17] has developed
a distance learning solution that allows non - resident students
to perform signal processing laboratory assignments. The
Electrical and Computer Engineering department has
developed an innovative mix of hardware and software to
allow students to access lab equipment regardless of their
location. The collection of signal generators and field
programmable gate arrays allow students to perform
experiments in real time.
VI. CONCLUSION
Cloud computing is basically a framework that allows us
to access applications that actually reside at a location other
than your computer. Virtualization is a technique to provide
abstraction between the hardware and software of a computer
system. In virtualization we create a virtual version of a
resource (generally computer system or application running on
it). The virtualization and cloud computing makes the
resources available to a remote location. Some resources are
shared between multiple users and accessed at the same time.
The technique uses the resource sharing and increase the

system utilization.
The Computer is virtualized in a clustered computing
environment controlled by Cloud computing to fully utilize
the computing and storage. The cloud will provide dynamic
resource allocation to ensure effective system utilization
through carefully monitoring of resource workload and
available physical resources, and providing managed solutions
on a continual basis. The most important aspect of
virtualization is to develop a platform that students can access
remotely without time and location constraints. The Cloud
provides a segregated virtual environment without affecting
any system outside of the cloud as well as other virtual
systems within the cloud.
REFERENCES
[1] Security Aspects of Virtualization in Cloud Computing.
Available from:
https://www.researchgate.net/publication/273950406_Security_
Aspects_
of_Virtualization_in_Cloud_Computing
[2] R M Sharma The impact of Virtualization in Cloud
Computing

International Journal of Recent Development in Engineering and
Technology Website: www.ijrdet.com (ISSN 2347-6435(Online)
Volume 3, Issue 1, July 2014).
[3] Oracle “The Structure of the Java Virtual Machine” http: //
docs .oracle
.com / javase / jvms / se7 / html / jvms - 2.html
[4] “Google cloud services – App Engine”. [Online]. Available:
http://www.google.com/enterprise/cloud/appengine/.
[5] “Windows Azure: Cloud Computing, Cloud Services , Cloud
Application Development”. [Online]. Available:
http://www.windowsazure.com/en-us/.
[6] “Amazon Web Services”. [Online]. Available:
http://aws.amazon.com/.
[7] Drew Alex C, Cloud Suite- Ad Hoc Laboratories Using
Cloud Resource,
California State University, Channel Island.
[8] EGI “European Grid Infrastructure” n.d. Web. http://
www.egi.eu/about/
[9] Amazon Web Services “Amazon S3 Pricing”
http://aws.amazon.com/s3/

[10] Google Cloud Platform “Google Compute Engine” http
cloud.google.com/products /compute - engine .html.
[11] Robinson, Tom. “Re: How does Heroku Work”
http://www.quora.com/Scalability/.
[12] Google Developers “Google App Engine General
Questions” https://
developers .google .com / appengine / kb / general.
[13] Eucalyptus “Why Eucalyptus” http:// www.eucalyptus.com
/ why –
eucalyptus.
[14] Coursera, “Support Center” http:// help.coursera.org/
customer/ portal/
articles / 557884 - will-each-class-that-i-complete – 02.
[15] Kahn Academy. “Talks and Interviews” http://
www.khanacademy.org/talks-and-interviews/v/khan-academy-
computer-science-launch.
[16] Stuckey - Mickell, Tracey, Stuckey - Danner, Bridget,
Taylor, Brandon.

“ Virtual Labs in the Online Perceptions and Implications for
Policy and
Practice “ , TCC 2007 Proceedings<http: / / etec.hawaii.edu /
proceedings / 2007 / stuckey.pdf greater than sign.
[17] Cristi, Roberto. “ EC3400 Digital Signal Processing
FFPGA Laboratory
“ http: / / faculty.nps.edu / dl / eo3404 dl lab / n.d.

Amid heightened competition to sell data analysis software, closel.docx

Recommended

Recommended

More Related Content

Similar to Amid heightened competition to sell data analysis software, closel.docx

Similar to Amid heightened competition to sell data analysis software, closel.docx (20)

More from daniahendric

More from daniahendric (20)

Recently uploaded

Recently uploaded (20)

Amid heightened competition to sell data analysis software, closel.docx