Empirical Investigation of Agile Software Development: A

Cloud Perspective


ABSTRACT
Technological advancements have contributed to more complex
software demands. The Agile approach to software development is
widely practiced by the software development industry as it offers
faster production with a promise of better software quality. It also
provides a flexible process to accommodate changes during the
software development life cycle, as per dynamic user requirements.
In this paper we characterize the significance of software
development models and the role of Agile methodologies in todays
dynamic world of technologies. The purpose of this paper is to
empirically investigate the choice between two of the Agile
methodologies, Scrum and Extreme Programming. The outcome of
the investigation based on secondary sources and limited primary
sources, reveals that current software industry practices tend to opt
for Scrum-based development. This work highlights the benefits of
bringing the cloud and Agile methods of software development
together, to fully realize the potential of the distributed paradigm.
Keywords
Software Development Life Cycle (SDLC), Agile Methods, Scrum,
Extreme Programming (XP), Cloud Computing
1. INTRODUCTION
The software development approach plays an important role in the
quality of software being developed. A non-systematic development
approach will produce low-quality software that may not meet user
expectations and also has a risk of incurring additional cost of
development. According to Sommerville [2] a software development
model provides a high-level abstraction of a set of best software-
development practices.
The software development process has progressed from the
traditional waterfall model to Agile processes with the shift from
static to dynamic industry demands and ever-advancing technologies.
Agile processes unlike the traditional models focus on incorporating
change within software during the process of software development
as they have shorter development cycles and higher levels of client
involvement.
In the present scenario of electronically-distributed software systems,
the concept of software-as-a-service (SaaS) has gained popularity.
Agile processes have a strong focus on collaboration among
developers and customers via feedback loops. The fact that the cloud
is seen as a business opportunity and platform to promote
collaboration makes it a more attractive option to implement Agile.
This work provides a comparison between two of the most-widely
accepted Agile methods, Extreme Programming and Scrum-based
development. It would enable the practitioners of Agile to choose
between these approaches. This paper also explores the idea of
combining the power of Agile and cloud. The rest of this paper is
organized as follows. Section 2 includes a literature review, followed
by section 3, which outlines the significance of software development
models. In Section 4, we throw light on Agile methodologies and
their practices. In Section 5, we analyze two widely-accepted Agile
models Scrum and Extreme Programming. In Section 6 we discuss
using the cloud for Agile development, leading to conclusions in
Section 7.
2. LITERATURE REVIEW
The early research on Agile focused on issues related to the adoption
of Agile methods. Less work in the literature investigates the
amalgamation of Agile with cloud. Andrea Magdaleno et al. [6]
presented a quasi-systematic review of reconciling software
development models. They found that there exist distinct
organization, group and process levels of reconciliation amongst
software development models.

The Standish Groups 2011 CHAOS report states that Agile methods
are three times as successful as the traditional approaches. In their
work, Felipe Carvalho et al. [21] outlined the fact that though all the
techniques that constitute the Agile group have overlapping subsets
of the features of the Agile manifesto [20], still they vary in their
relevant characteristics.
The work done by O. Salo and P. Abrahamsson [9] demonstrates a
clear trend towards higher adoption of Extreme Programming
practices in comparison to Scrum in embedded software development
organizations across Europe. Another empirical study conducted at
Microsoft by Begel, A. and Nagappan [4] revealed that one-third of
Anupriya Tuli
Amity University Uttar Pradesh
Noida, India
[email protected]

Nitasha Hasteer
Amity University Uttar Pradesh
Noida, India
[email protected]

Megha Sharma
Amity University Uttar Pradesh
Noida, India
[email protected]

Abhay Bansal
Amity University Uttar Pradesh
Noida, India
[email protected]

ACM SIGSOFT Software Engineering Notes Page 1 July 2014 Volume 39 Number 4
DOI: 10.1145/2632434.2632447 http://doi.acm.org/10.1145/2632434.2632447
the study respondents used Agile methodologies. This study shows
that Scrum is the most popular agile methodology practiced at
Microsoft. Asif Qumer et al. [26] presents a comparative analysis of
Extreme Programming and Scrum, using the 4-Dimensional
Analytical Tool (4-DAT). This comparison is based on scope, agility,
Agile values and software process. Their work highlights that Scrum
supports more Agile practices, but has less Agile phases than
Extreme Programming.
Software as a Service (Saas) has gained popularity in recent times
along with distributed software development systems. A team of
researchers at Collabnet [24] is working on a development platform
in the cloud to simplify the Agile issues in a cost effective way. The
work of Amit Dumbre et al. [14] demonstrates that Agile software
development practices are significantly enhanced with cloud
technologies and platform as a services such as Azure Platform as a
service, to provide an on-demand, scalable environment to achieve
faster software delivery.

3. SOFTWARE DEVELOPMENT MODELS
Software engineers aim to use software development models for
building fault free software that meets user requirements and is
delivered within the specified time limit and budget. The use of
software development models enables organized development of
software. Without these models, software development teams may
produce low-quality software, which is difficult to maintain and
violates the budget and time deadlines.
In 1968, during the first NATO Software Engineering Conference,
the term Software Crisis was coined to refer to the difficulty of
writing correct, understandable and verifiable software [18]. The
convergence of communication, computer science and GUI
encouraged the adoption of planned software development methods.
After 1968, technical advancements have made software
development more complex in nature, eliminating informal
approaches of software development, which were earlier being used
for developing simple software. This transformation gave birth to a
situation where hardware cost was decreasing and software cost was
increasing. Developing software using software development models
gave a solution that was required in the software industry to control
the software crisis [16].
Software development models can be categorized into plan-driven,
Agile and FOSS (Free and Open-Source Software) development
models. These models follow different approaches, but have the same
goal to enhance the end product by using a systematic software
development process. Agile development models help in building
flexible software and enable teams to respond to the unpredictability
of software development. The plan-driven models seek reliability,
predictability and stability. The FOSS development models develop
software products based on special characteristics. Among these
software development models we find similarities as well as many
differences. Plan-driven, Agile and FOSS development models
follow different approaches to software development [6].

4. AGILE SOFTWARE DEVELOPMENT
Agile can be defined as the software development methods where
tasks are carried out in small work phases, characterized by
retrospective analysis, contributing to frequent modification to the
software in accordance with the ever-changing requirements of
customers. The Agile processes have a group of software
development methods to execute iterative and incremental
development, by replacing high-level design with frequent redesign.
With rapid technological enhancements, the software being
constructed today is becoming more and more complex. To meet the
dynamic demands of customers and avoid a software crisis,
companies are adopting Agile methods. A Survey conducted by
Murphy et al. [22] shows an increase in Agile methods being
practiced within Microsoft from 2006 to 2012.


The Agile umbrella provides shelter to many software development
methodologies including Scrum, Extreme Programming, Crystal,
Kanban, FDD (Feature-driven development), DSDM (Dynamic
Systems Development Method), etc.
4.1 Extreme Programming (XP)
Sommerville [17] states that Extreme Programming is a planned and
iterative method to develop software with extreme levels of
involvement by customers in the construction and testing phases.
The features of XP that are uncommon to the rest of its counter
approaches are [17, 19]:
4.1.1 Requirements
These are expressed as scenarios by users, which are documented in
the form of Story Cards. The development team breaks each story
card into a series of small tasks, which are prioritized by the customer
for development.
4.1.2 Planning Game
The development team estimates the effort needed for
implementation of Story Cards. It is on the basis of this estimation
that the customer decides the scope and release time of the tasks.
4.1.3 Simple Design
XP does not accommodate changes in the design, which causes
degradation of the software structure. The developers nullify this
software structure degradation by performing refactoring. In
refactoring the software is immediately updated whenever there is a
scope for improvement.
4.1.4 Pair Programming
Programmers work in dynamic pairs encouraging communication,
group ownership and reduction of workload and working hours.
4.1.5 Test First Development
This is a unique concept where a test case is written prior to coding
for the task. XP is a software development model with testing being
an integral part.
4.1.5.1 The customer is responsible for developing acceptance test
cases and validations.
4.1.5.2 Acceptance testing is incremental in nature as series of test
cases are executed.
4.1.5.3 Testing is done using automated test harness tools.
Figure 1. Percentage of Agile methods practiced in Microsoft
over the years
ACM SIGSOFT Software Engineering Notes Page 2 July 2014 Volume 39 Number 4
DOI: 10.1145/2632434.2632447 http://doi.acm.org/10.1145/2632434.2632447





The French Scrum User Group [12], which is an affiliate of the
Scrum Alliance, conducted a national survey on Agile methods in
France in J une 2009. The survey, with the participation of 230
industry professionals from over 150 companies ranging from E-
commerce, banking, insurance, scientific innovation, software
development and IT service providers, showed that 52% of
companies were using XP. Andrew Begel et al. [15] conducted a
survey at Microsoft Corporation to study the perception towards pair
programming, one of the engineering practices advocated by XP.
Their survey indicates that 64.4% of respondents believe that pair
programming works well for them and produces higher quality code.
O. Salo and P. Abrahamsson [9] conducted a survey on the use of
Extreme Programming in embedded software development
organizations across Europe. They found that 54% of the
participating organizations applied XP.

4.2 Scrum Methodology
In the last decade of software development, there has been challenges
related to methodologies and best practices to deliver software.
People used to do work with the help of defined roles, defined
models, and defined practices to execute defined plans. In the
present times, there is a need for an approach through which
resource, schedule and cost constraints could be best managed. One
of the methodologies currently used is Scrum. It is a lightweight
methodology that uses fixed-length iterations known as sprints. A
sprint includes all aspects of work, i.e. designing, coding, testing, etc.
Scrum methodology has the following roles:
4.2.1 Product Owner
The product owner is responsible for product vision, maximizing
return on investment, prioritizing the product backlog, which is the
list of customer-centric features and also deciding whether to
continue development or not.
4.2.2 Scrum Development Team
The team is responsible for doing each and every task as assigned by
the product owner within the sprint deadline. A team of 5 -7 members
is made to carry out the development tasks.
4.2.3 ScrumMaster
The ScrumMaster acts as a facilitator who resolves obstructions and
also helps in creating a self-organizing environment.

Further, there are Scrum meetings, which are initiated by the
ScrumMaster, although the ScrumMaster has no decision-making
authority at these meetings. In the Sprint planning meetings, the
Product Owner assigns a priority to the tasks from the product
backlog items (created by Product Owner). Selected work is then
shifted from the product backlog to the sprint backlog. Team
members decide a particular place and time within the office
premises where they discuss their work on a daily basis. In a Sprint
review meeting after completion of each Sprint, the team gives a live
demonstration of the product increment to the Product Owner. The
Product Owner then decides whether to accept the task as complete
or not. Lastly, in Sprint retrospective meetings, all the team members
review their work completed during a particular Sprint. They
conclude whether to continue in the same manner or adopt a different
strategy.



A survey conducted by the French Scrum User Group [12] in 2009
showed that 86% of the companies were using Scrum. A Danube
Case study by Pat Elwer [13] discussed the adoption of Scrum by
Intel in 2008 for one of its products. The outcome of the project
showed a 66% reduced cycle time, performance to schedule,
improved morale of employees, and increased transparency.
According to this report, Scrum was considered as the best project
management framework to employ. Another survey on the use and
usefulness of Scrum in embedded software development
organizations across Europe by O. Salo and P. Abrahamsson [9]
showed that 27% of the participating organizations applied Scrum.
However, there are a few challenges that organizations may
encounter while adopting the Scrum methodology:
Cost to company may increase due to staff turnover.
Availability of skilled resources.
Ability to build applications from multiple locations in
projects with geographically distributed teams.
Test-driven development is difficult as it sometimes leads
to excess code generation.

5. COMPARISON BETWEEN SCRUM AND
EXTREME PROGRAMMING
The Agile manifesto [20] states a set of principles, which are
practiced using different software development models collectively
known as Agile processes. Although all of these development models
are under the Agile umbrella, they vary on different parameters of
Agile principles. Empirical study of both Scrum and XP variants of
the Agile methodology led to the formulation of the comparison
presented in Table 1.
Figure 3. Scrum meeting flow
Figure 2. Extreme Programming approach
ACM SIGSOFT Software Engineering Notes Page 3 July 2014 Volume 39 Number 4
DOI: 10.1145/2632434.2632447 http://doi.acm.org/10.1145/2632434.2632447

Table 1. Comparison between XP and Scrum













Parameter


Scrum-Based Development

Extreme Programming (XP)

Relevant Characteristic [21]


Lifecycle management

Software construction

Customer Involvement [10]


Not specifically onsite

Onsite customer
Requirement Documents

Managed Requirements in form of
ARTIFACTS: Product backlog, Sprint backlog


Managed Requirements in form of STORY
CARDS.

Project Management Process
[26,10]

Sprint planning meeting, daily Scrum meeting
and Sprint review

Planning game

Facilitator [10]

ScrumMaster

XP coach

Module Development Priority
[17]


Priority by Scrum product owner

Priority by customer
Design Principle [26]
Complex Design and believes in design for
change

Simplification of code and unanticipated change
implementation (refactoring)

Development Approach

Iterative (Sprint) +Incremental

Iterative +Incremental
Coding

Coding Standards are not used

Coding Standards are used

Working Hours [17]

Programmers have long working hours


Programmers work in pairs which reduces
workload and working hours (pair programming)

Testing

No formal testing approach


Test first approach, automated and incremental
acceptance testing


Product Ownership [10]

Product Owner is responsible for the product

Group responsibility
ACM SIGSOFT Software Engineering Notes Page 4 July 2014 Volume 39 Number 4
DOI: 10.1145/2632434.2632447 http://doi.acm.org/10.1145/2632434.2632447
Quick interviews were conducted with the representatives of nine
software development companies. It was found that only one out of
the nine companies that participated in the interview currently
implements XP, whereas the rest of them practice Scrum.




6. CLOUD FOR AGILE DEVELOPMENT
The usage of Agile software development is increasing due to many
benefits like increased productivity, better quality, more flexibility
and collaborative processes to accommodate changes. But there can
be many unanticipated issues like scalability that can arise in parallel
with the use of Agile software development. Also, providing and
managing development tools and infrastructure that support the Agile
process could be challenging for enterprises.Using Cloud Computing
would help Agile enterprises in this regard. By leveraging a cloud
development platform, an enterprise can start using development
tools that support the Agile process, cost effectively, without
provisioning and managing the development infrastructure [24].
Basically Agile depends on many different tools to support its
practices like feedback, transparency and prioritizing tasks. An Agile
organization requires development infrastructure and an integrated
toolset. All these requirements are fulfilled more easily with the help
of a cloud development platform. The following are few of the
benefits of bringing cloud and Agile together [14, 24]:
6.1 Distributed Application Development
The cloud supports distributed application development with teams
dispersed geographically. Any team member can have access to
information about the application at any time and at any place.
6.2 Data Sharing
Resources are required to share a huge amount of data between
different machines for many applications. The cloud plays a crucial
role by providing a medium to share large amounts of data frequently
as per the requirements.
6.3 Prioritizing Tasks
Continuous prioritizing and reprioritizing the tasks within a sprint
needs tracking and transparency. In a distributed system, there is a
requirement of a tool that can be accessed by all team members to
keep them updated. Deploying such a tool internally consumes
energy and takes time. With a cloud development platform,
organizations can provision Agile Management tools in minutes, and
can scale globally with thousands of users.
6.4 Transparency
Transparency comes through data capture. In a cloud-based Agile
development, organizations can have developer services that capture
and share data from all distributed code repositories, build and test
environments. By capturing the needed and important data,
organizations can measure performance, observe, supervise and
manage the Agile project.
6.5 Infrastructure
Agile software development requires Infrastructure to be in place
from the start of the project. Cloud helps by providing infrastructure
from the first day of the project. Additional overhead of software and
hardware upgrade is taken care of, by the Cloud provider.
6.6 Benefits Business Users
Cloud facilitates Agile practices in order to meet the users
requirements faster. The cloud supports the concept of pay per use,
easy deployment of applications to Go Live and also provides faster
time to market services and products.

7. CONCLUSION

Agile philosophy is an answer to many important issues of new
software reality ranging from distributed software development to
software as a service (SaaS). The increase in adoption of Agile
practices with time has shown the promising aspect of this approach.
Agile brings in added value and ensures better returns for developing
software and services on cloud. The salient feature of the Agile
approach is collaboration between customers and organizations. The
power of Agile can increase with Cloud-based development. The idea
of the amalgamation of Agile and cloud not only solves the
scalability issues, but also augments existing Agile practices to
deliver software faster.
8. REFERENCES
[1] Da Silva, E.A.N., Lucredio, D. 2012. Software Engineering for
the Cloud: A Research Roadmap. In
0
1
2
3
4
5
6
7
8
9
XP Scrum
No. of companies practicing Scrum/XP as per a
quick telephonic survey of 9 companies.
no.ofcompanies
Agile
Method
Company Name
XP
Vismaad
(http://vismaad.com/software.htm)
Scrum
Computer Sciences Corporation
(http://www.csc.com/in)
Toshiba Software Pvt. Ltd.
(http://www.toshiba-tsip.com/tsip/)
Mercer
(http://www.mercer.co.in/home)
Accenture Services Pvt. Ltd.
(www.accenture.com)
Avalon Technologies, Inc.
(avalontech.net)
Kellton Tech
(www.kelltontech.com)
Deloitte Touche Tohmatsu Limited
(https://www.deloitte.com/in)
Figure 4. Scrum is being practiced more as compared to XP
Table 2. Companies practicing XP/Scrum
ACM SIGSOFT Software Engineering Notes Page 5 July 2014 Volume 39 Number 4
DOI: 10.1145/2632434.2632447 http://doi.acm.org/10.1145/2632434.2632447
Software Engineering (SBES), (2012) 26th Brazilian
Symposium on, 71- 80. DOI=10.1109/SBES.2012.12.
[2] Sommerville, I. 2004. Software Engineering, Addison Wesley, 7
Ed.
[3] Highsmith, J . and Cockburn, A. 2001. Agile Software
Development: The Business of Innovation. Computer, vol. 34,
120-122. DOI=10.1109/2.947100
[4] Begel, A. and Nagappan, N. 2007. Usage and perceptions of
Agile software development in an industrial context: An
exploratory study. In
Empirical Software Engineering and Measurement,
(Washington, 2007), 255 - 264. DOI=10.1109/ESEM.2007.12.
[5] Yau, s.s., and An,H.G. 2011.Software Engineering Meets
Services and Cloud Computing. Computer ,Vol 44, no 10 (Oct.
2011). DOI=10.1109/MC.2011.267
[6] Magdaleno, A.M., et al., Reconciling software development
models: A quasi-systematic review, J ournal of Systems and
Software, v.85 no.2, 351-369 ( February,
2012).DOI=10.1016/j.jss.2011.08.028
[7] Rising, Linda, J anoff, N.S. 2000. The Scrum Software
Development Process for Small Teams, IEEE Software, vol. 17,
no. 4 (J uly/Aug. 2000), 26-32. DOI=10.1109/52.854065.
[8] Williams, L., Cockburn, A. 2003.
Agile software development: it's about feedback and change,
Computer, Vol.36, no.6 (2003), 39 43.
DOI=10.1109/MC.2003.1204373.
[9] Salo, O. and Abrahamsson, P. 2008. Agile Methods in European
Embedded Development Organizations: a survey study of
Extreme Programming and Scrum, IET Software, vol. 2
(February, 2008), 58-64. DOI =10.1049/iet-sen:20070038.
[10] MSDN Library, Agile Principles and Values by Jeff Sutherland,
Microsoft Developer Network White Paper.
http://msdn.microsoft. com/en-us/library/dd997578.as
[11] Standish Group 2011. The Crisis in Software: The Wrong
Process Produces the Wrong Results, The Standish Group
Report (2011).
www.controlchaos.com/storage/S3D%20First%20Chapter.pdf .
[12] French Scrum User Group 2009. A National Survey on Agile
Methods in France. (J une 2009) www.frenchsug.org
[13] Elwer, P., et al.. 2008. Agile Project Development at Intel: A
Scrum Odessey, Danube Case Study of Intel Corporation.
http://Scrumtrainingseries.com/Intel-case-study.pdf
[14] Dumbre, A., et al. 2012. Practicing Agile software development
on the Windows Azure Platform., White Paper Infosys (2012).
www.infosys.com/cloud/.../practicing-Agile-software-
development.pdf
[15] Begel, A., Nagappan, N. 2008. Pair programming: what's in it
for me? , In Proceedings of the Second ACM-IEEE international
symposium on Empirical software engineering and
measurement, Kaiserslautern, Germany.
DOI=10.1145/1414004.1414026
[16] Schach, S. 2007. Software Engineering, Tata McGraw Hill, Ed.
7, 4-6.
[17] Sommerville, I. 2005. Software Engineering, Pearson, Ed. 7, 26,
418 430.
[18] Randell, B. 1996. The 1968/69 NATO Software Engineering
Reports.http://homepages.cs.ncl.ac.uk/brian.randell/NATO/NAT
OReports.
[19] Beck, K. 1999. Embracing Change with Extreme Programming.
Computer, vol.32, 70-77. DOI=10.1109/2.796139
[20] Beck, K., et al. 2001., Manifesto for Agile Software
Development. http://Agilemanifesto.org/
[21] Carvalho, F., et al.. 2013. Service Agile Development Using XP,
SOSE, IEEE Conference Publication, 254
259.DOI=10.1109/SOSE.2013.25
[22] Murphy, B., et al.. 2013. Have Agile Techniques been the Silver
Bullet for Software Development at Microsoft?,
Empirical Software Engineering and Measurement, ACM /
IEEE International Symposium,(2013), 75-84.
DOI=10.1109/ESEM.2013.21
[23] Hasteer, N., Bansal, A., Murthy, B.K. 2013. Pragmatic
Assessment of Research Intensive Areas in Cloud : A
Systematic Review, In ACM SIGSOFT Software Engineering
Notes, Vol. 38, no.3 (May 2013 ), ACM New York, NY, USA,
1-6.
[24] Willie, 2011. Reinforcing Agile Software Development in the
Cloud, White Paper CollabNet Inc.
https://www.open.collab.net/media/pdfs/CollabNet%20Whitepa
per_Reinforcing%20Agile%20Dev%20in%20the%20Cloud.pdf
?_=d
[25] Munshi, D. 2009. Adopting Agile Methodology in a Large Scale
Distributed Development Environment, TCS White Papers
(2009).http://www.tcs.com/sitecollectiondocuments/white%20p
apers/GCP_white-
paper_Agile_Methodology_Large_Scale_Distributed_Developm
ent_Environment_01_2010.pdf
[26] Qumer, A., Henderson-Sellers, B. 2006. Evaluation of XP and
Scrum using the 4D analytical tool (4-DAT), European and
Mediterranean Conference on Information Systems (J uly 6-7
2006).epress.lib.uts.edu.au/research/bitstream/handle/.../200600
5499.pdf?
[27] Ramanujam, R., et al.. 2011. Collaborative and Competitive
Strategies for Agile Scrum Development. In proceedings of 7
th

International conference on Network Computing and Advanced
Information Management (NMC), IEEE Conference
Publication, 123 127.





ACM SIGSOFT Software Engineering Notes Page 6 July 2014 Volume 39 Number 4
DOI: 10.1145/2632434.2632447 http://doi.acm.org/10.1145/2632434.2632447