Developing Green and Sustainable Software using Agile Methods in
Global Software Development: Risk Factors for Vendors
Nasir Rashid and Siffat Ullah Khan
Software Engineering Research Group (SERG-UOM), Department of Computer Science & Information Technology,
University of Malakand, KPK, Khyber-Pakhtunkhwa, Pakistan
Keywords: Systematic Literature Review, Green and Sustainable Software, Agile Software Development, Risk Factor,
Green Agile, GSD Vendors.
Abstract: Global software development (GSD) is gaining momentum due to the potential benefits it offers. GSD aims
at delivering remarkable software through a widely distributed pool of experts, with reduced efforts,
minimum cost and time. In recent years, GSD developers have reshaped the development processes and
have adopted agile techniques and green engineering principles to cope with the frequent changes in
requirements, accelerate the development in short increments and to produce energy efficient and
sustainable software. However, the adoption of agile methods for developing sustainable software possesses
a number of challenges. This paper presents a list of potential challenges/risks identified through systematic
literature review (SLR) that need to be avoided by the GSD vendors using agile methods for the
development of green and sustainable software. Our findings reveal eight risk factors that are faced by GSD
vendors in the development of green and sustainable software using agile methods. GSD vendors are
encouraged to address properly all the identified factors in general and the most-frequently cited critical
risks in particular, such as in-sufficient system documentation, limited support for real-time systems and
large systems, management overhead, lack of customer’s presence, lack of formal communication and lack
of long term planning.
1 INTRODUCTION
Global software development (GSD) has been
grown with recent improvements in ICTs. In GSD
various software engineers collaborate over
temporal, geographical, cultural and linguistic
distances. The major motivations for GSD are to
achieve improvements in resources at low cost, high
quality software, round the clock development and
time to market efficiency (Alsudairi and Dwivedi,
2010). However, GSD benefits will not be
achievable unless the associated risks are not
managed (Noll et al., 2010). Some of the potential
risks of GSD include hidden agreement costs,
maintenance, lack of awareness of existing tools and
lack of support for collaboration (Khan and Azeem,
2014).
In order to reduce some major risks in traditional
software development, such as resources cost,
frequent request for changes and timely delivery of
software, the software engineering community have
proposed some flexible methods called agile
methods (Altameem, 2015). Some of the agile
methods are Scrum, Extreme programming (XP),
Crystal, Dynamic Systems Development Method
(DSDM) and Lean Software Development (LSD).
These methods help software developers to have
more focus on requests for rapid changes in
requirements, iterative development, collaborations
among the developers and efficient resources (Al-
Saleem and Ullah, 2015). It is obviously clear that
agile methods have a positive impact on software
development life cycle as it enhances the efficiency
of developers and results in energy efficient
software.
Agile methods promise to scale down the
potential risks, detect faulty code, magnify software
production and embrace frequent changes (Wrubel
and Gross, 2015). Using agile methods in distributed
environment also yield many benefits like smart
communications, rigorous integration, sufficient
document production and scheduled software (Singh
et al., 2015).
This is not surprising that none of the agile
Rashid, N. and Khan, S.
Developing Green and Sustainable Software using Agile Methods in Global Software Development: Risk Factors for Vendors.
In Proceedings of the 11th International Conference on Evaluation of Novel Software Approaches to Software Engineering (ENASE 2016), pages 247-253
ISBN: 978-989-758-189-2
Copyright
c
2016 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
247
method is a silver bullet. It also poses a number of
risks in various environments like limited support
for distributed development, limited support for
development of safety critical software, management
overhead, reliance on the tacit knowledge of
developers and in-effective communication among
the customer and developers (Omar et al., 2011).
Research in the area of green agility is
flourishing with increase in demand for the
development of green and sustainable software.
Green software engineering (GSE) is a prominent
paradigm that has attracted the enthusiastic software
developers, aims to develop, design and use the
software with minimum economic, societal and
ecological impacts (Raturi, Tomlinson et al. 2015).
GSE aims at reducing the depletion of energy and
natural resources and to scale down the direct and
indirect negative effects on human being due to the
development and use of software systems (Naumann
et al., 2011).
In recent years, potential growth in green agile
research has been noticed. Considerable amount of
publications shed light on the issues in agile
methods and in GSD as well, that motivated us for
systematic review to identify risk factors in this way.
The expected findings will be the first of its kind of
study that will be helpful in the development of
green agile maturity model in GSD that aims to
measure green-agile maturity of vendor
organisations. The preliminary structure of our
proposed model has been published (Rashid and
Khan, 2014). We have followed the concept of
CMMI, IMM and SOVRM (Khan et al., 2010;
Niazi, 2007; Silva et al., 2015)
This paper tries to answer the following research
questions.
RQ1. What are the risks involved, as mentioned in
the literature, to be avoided by agile software
developers in GSD for the development of green and
sustainable software using agile methods?
RQ2. Do the identified risks in green agile
development, vary geographically (continent-wise)?
In the next section we describe related work, in
section 3 we discuss our research methodology, and
in section 4 we present the results. Section 5 discuss
summary of our research, section 6 reports some
limitations while section 7 presents conclusion and
future work.
2 RELATED WORK
Agile methods have gained popularity in recent
years and are working out to revamp the key
processes involved in software development to
ensure the delivery of green and sustainable software
(Misra et al., 2012). However agile methods poses
some limitations towards sustainability like test
automation, backlog management, volatility of
requirements and un-availability of customer’s
representative (Hushalini et al., 2014).
There is an emerging trend of introducing agile
methods in GSD to achieve the maximum benefits
of however this combination raises new challenges.
Agile in GSD means to practice the principles of
agile methods for energy efficient software delivery
(Singh et al., 2015).
However it is certain that agile distributed
development raises some challenges like task
management, delays in project delivery, team
management and communication (Jawad and Taira,
2015).
(Mohammad et al., 2013) demonstrate the
strengths and weaknesses of agile development in
GSD projects. Beside the voluminous benefits it
offers, agile methods still face some challenges
when it comes into practice, such as inconsistency in
customers interaction, difficulty of managing large
teams and lack of long term planning.
Current agile methods in practice produce less
implementation that meet only short-term needs and
thus assuredly produce software not green in nature
and lack long-term suitability. (Mohan et al., 2010).
The literature described earlier explains some
potential risks in the context of sustainable software
development using agile methods. However, none of
these factors have been identified through SLR. Our
findings affirm that no research has been carried out
to explore risks in sustainable software development
using agile methods. The results obtained from this
research work will assist the GSD agile developers
to avoid the identified.
3 METHODOLOGICAL
APROACH
This study adopts SLR methodology (Kitchenham
and Charters 2007) as a research methodology as
used by other researchers (Verner et al., 2012; Khan
et al., 2015; Alzoubi et al., 2016). SLR is a
systematic method for identifying, evaluating and
interpreting all currently available research relevant
to a particular research questions or area of interest.
(Zhang et al., 2011). The following subsections
describe the procedure specified in our protocol
ENASE 2016 - 11th International Conference on Evaluation of Novel Software Approaches to Software Engineering
248
(Rashid and Khan, 2015).
3.1 Searching Process
Developing an appropriate search strategy in SLR is
quite essential. A generous was conducted to cover a
broad range of relevant publications. The search
process used five search engines, i.e. IEEEXplore,
Science Direct, Google Scholar, Springer Link and
ACM Digital Library. Manual searches for the
relevant papers were conducted on the mentioned
libraries using snowballing approach to increase the
number of relevant research articles as used by
(Wohlin, 2014).
Table 1: List of identified risk factors.
S.
No
Risks Factors
Frequency
N=42
Percentage
1
Insufficient system
documentation
26 62
2
Limited support for
real-time systems and
large systems
19 45
3
Management
overhead
25 60
4
Lack of customer’s
presence
26 62
5
Lack of formal
communication
18 43
6
Limited support for
reusability
07 17
7
Insufficient
knowledge of the
customer
15 36
8
Lack of long term
planning
20 48
Search terms used were ((Global OR GSD OR
Distributed) AND (Agile OR ‘‘Agile methods’’ OR
“Green Agile” OR “Agile approaches”) AND
(Green OR Sustainable OR “Green Software
Engineering” OR “Green software”) AND
(Limitations OR Challenges OR Issues OR Risks)).
Adjustments were made where needed to fix the
syntax of different search engines used.
3.2 Publication Selection
Papers resulting from the generous search were
reviewed, after considering the title if required, the
abstract, rejected all other papers that were
obviously irrelevant. This resulted in 42 final
publications.
3.3 Data Extraction and Synthesis
Relevant data was extracted to answer the research
questions. A conscientious study carried out by the
author catalogued 25 groups of risk factors from 42
publications. After validation, we found 08 risks to
be decisive as shown in Table 3. Among the 08 risk
factors, 06 are remarkable due to high frequency and
are considered as critical risk factors (CRFs) for
having frequency >=40%. Criteria for criticality
have been acquired from (Khan et al., 2009).
4 RESULTS
The following subsections report the results of SLR
to answer the research questions.
Table 2: Continent wise paper frequency.
Continent
Frequency
N=42
Percentage
Asia 12 29
N. America 09 21
Europe 18 43
Mixed 03 7
4.1 Risk Factors while using Agile
Methods for Green and Sustainable
Software Development (Research
Question 1)
Table 1 illustrates that ‘insufficient system
documentation’ is the most common risk factor
(62%). Agile methods provide only a few
implementation details of software implementation,
and because of this “low-level” approach they may
build the software that meet short-term individual
project needs, but that do not necessarily lead to
sustainable software systems (Stammel et al., 2011;
Hall, 2014).
Our findings reveal that ‘Lack of customer’s
presence’ (62%) have the same severity to the
previous mentioned risk factor. Consistent presence
of the customer is most essential practice of agile
methods. This reduces the efforts to complete the
software within defined time frame and leads to
sustainable development (Mahmoud and Ahmad
2013).
In absence of customer, the software builds take
a longer time to complete and are more likely to fail.
Coherent, self-organizing agile teamwork and strong
communication supports a green process throughout
software development HSIEH and CHEN 2015).
Management of time and computing resources is
an integral component that promotes green software
development in all types of software development
environments. However, agile methods experience
Developing Green and Sustainable Software using Agile Methods in Global Software Development: Risk Factors for Vendors
249
Table 3: Distribution of risk factors continent wise.
RiskFactors
SampleSize
N=42
ChiSquaretest
(Linear‐by‐Linear
Association)
α=0.5
Asia
N=12
N.America
N=09
Europe
N=18
Mixed
N=03
Frequency % Frequency % Frequency % Frequency % X
2
d
f
p
Insufficient
system
documentation
09 75 06 67 10 56 03 100 2.105 1 0.147
Limitedsupport
forrealtime
systemsandlarge
systems
06 50 03 33 10 56 0 0 .208 1 .648
Management
overhead
10 83 02 22 10 56 03 100 .137 1 .711
Lackof
customer’s
presence
09 75 06 67 10 56 01 33 2.105 1 .147
Lackofformal
communication
04 33 04 44 08 44 02 67 .844 1 .358
Limitedsupport
forreusability
01 8 02 22 04 22 0 0 .128 1 .669
Insufficient
knowledgeofthe
customer
06 50 04 44 04 22 01 33 2.025 1 .155
Lackoflongterm
planning
08 67 04 44 07 39 01 33 2.25 1 1.133
the risk of ‘management overhead’ (60%), which
may lead to over-budget, time over run and
maximum use of available computing resources for
software development. This can greatly influence the
sustainable and green approach towards software
development (Altameem, 2015).
Intelligent and long- term planning is a crucial
factor to be considered for social, economic and
environmental sustainability of the software (Calero
et al., 2015). The ‘lack of long-term planning’
(48%)’ is listed as the fourth highest risk concern
with the use of agile methods in order to produce
sustainable software. The importance of
sustainability is increasingly recognized in terms
ofsoftware development which needs a long term
planning (Venters et al., 2015). However agile
methods focus more on the immediate delivery of
software according to the current needs of customers
without taking into considerations its long term
impact on human and society as a whole.
4.2 Assessment of the Risk Factors,
Continent Wise Analysis (Research
Question 2)
In order to answer the second research question,
Table 3 highlights a list of 8 risk factors retrieved
through SLR in distinct continents while Table 2
presents the research data from various continents.
North America, Europe and Asia are considered
among the continents for analysis of data whereas
the rest are merged into ‘Mixed category’ due to
small sample size. The content wise analysis
diagnoses the existence of compelling variances
among the risk factors. In order to observe the
variances, we apply linear by linear association Chi-
square (X
2
) test. For the evaluation of derivations
between ordinal variables, X
2
linear by linear
association is used for analysis, which is considered
more prevalent as compared to Pearson Chi-square
test.
Table 3 interprets the frequency of publications
in various. Few variations have been noticed among
the identified risk factors across different continents
as illustrated in Table 3. All the listed risk factors
have been found in the same number in different
continents except for ‘Limited support for real-time
systems and large systems’ and ‘limited support for
re-usability’, which do not exist in mixed continent
category.
Some risk factors show high rise up in frequency
in particular continents, such as ‘management
overhead’, ‘lack of customer presence’, ‘insufficient
system documentation’, ‘limited support for real
time systems and large systems’ and ‘lack of long
term planning’. This shows that in certain continents
the agile maturity may not have reached up to the
level that can better mitigate the risks and to
ENASE 2016 - 11th International Conference on Evaluation of Novel Software Approaches to Software Engineering
250
contribute towards green software development.
Table 3 also highlights the high ranked risk
factors among the identified ones in specific
continents. In Asia ‘insufficient system
documentation’, ‘management overhead’, ‘lack of
customer presence’ and ‘lack of long term planning’
have got high frequency. In North America only the
‘insufficient system documentation’ and ‘lack of
customer’s presence’ have got significant values.
While in Europe the severity of risks is the same as
in Asia except for ‘lack of long term planning’,
which shows low frequency. The three mentioned
continents present slight different approach in
applicability of agile methods in software
development.
The findings in Table 3 shows the risks
confronted when using agile methods for the
development of green and sustainable software in
different continents. The variation in frequency of
risks across the continents may be the difference in
nature and complexity of the software productions,
difference in technical expertise of agile developers,
selection of agile technique used or may be that
some factors are more concerned regarding the
milestones defined for organisations in different
continents that give rise to these potential risks.
5 SUMMARY AND DISCUSSION
This study indicates a number of potential risks
identified through SLR that need to be avoided by
the agilists when intend to develop green and
sustainable software in GSD environment. The
identified risks expose some key process areas of
agile methods that need a magnitude of agilists’
focus in the way of greener software development.
In order to pinpoint the severity of the identified
risks, we take into account the frequency percentage
to be >=40, as the criteria set by other researchers
(Darwish and Rizk, 2015) for the identification of
critical risk factors. We adopted the above
mentioned criteria for criticality of the identified risk
factors on RQ1, which resulted in 06 critical risks.
6 THREATS TO VALIDITY
With respect to internal validity, the threat comes
from the fact that authors of selected publications
may not have sufficient knowledge of the subject
areas to describe the identified risks in depth. There
may be an inclination towards reporting some risks
in certain papers as well. Concerning external
validity, first threat is the undeniable limited number
of publications, as search strings are compiled and
interpreted differently by various search engines.
Secondly, most of the papers we found are authored
by academicians who may lack the practical
experience of agile methods.
Finally, regarding construct validity, a pilot
questionnaire is made prior to the final
implementation of the construct, which aims to
ensure the improvement and assurance of the
associated documentation. Probably it will result in
several changes to the identified risk factors.
7 CONCLUSION AND FUTURE
DIRECTIONS
This study presents a distinct approach to evaluate
the use of agile methods for the development
sustainable software using SLR and to find the risks
along this way. This research identified 08 risk
factors as listed in Table 1, from a sample of 42
papers, out of which some risks have been declared
as critical due to high frequency.
Our findings suggest some key processes for
future work in agile GSD industry: (1) to conduct an
empirical study to validate the our findings (2) to
explore some unforeseen risks apart from the
identified ones, if any and (3) to conduct empirical
study for the identification of relevant practices for
the mitigation of the identified critical risks. Similar
method has been followed by other software
engineers and researchers (Garousi et al., 2015).
This study is a vital step in the long way
approaching the development of our proposed green-
agile maturity model that aims to measure the agile
capability of vendor organization in the context of
green software development. Our proposed research
design consists of SLR, empirical study in GSD
industry and case studies at vendor organizations. A
similar research design has been used by other
researchers (Niazi et al., 2015). This study also
presents a continent wise analysis of the identified
risks as shown in Table 3. The mentioned analysis
may be useful in providing a deep insight to agile
software developers about the most critical risks that
need in depth focus to be avoided.
REFERENCES
Al-Saleem, S. M. and H. Ullah (2015). "A Comparative
Developing Green and Sustainable Software using Agile Methods in Global Software Development: Risk Factors for Vendors
251
Analysis and Evaluation of Different Agile Software
Development Methodologies." IJCSNS 15(7): 39.
Alsudairi, M. and Y. K. Dwivedi (2010). "A multi-
disciplinary profile of IS/IT outsourcing research."
Journal of Enterprise Information Management 23(2):
215-258.
Altameem, E. (2015). "Impact of Agile Methodology on
Software Development?" Computer and Information
Science 8(2): p9.
Khan, S. U., & Azeem, M. I. (2014). Intercultural
challenges in offshore software development
outsourcing relationships: an exploratory study using a
systematic literature review. Software, IET, 8(4), 161-
173.
Calero, C., M. Á. Moraga, M. F. Bertoa and L. Duboc
(2015). Green Software and Software Quality. Green
in Software Engineering, Springer: 231-260.
Darwish, N. R. and N. M. Rizk (2015). "Multi-
Dimensional Success Factors of Agile Software
Development Projects." International Journal of
Computer Applications 118(15).
Garousi, G., V. Garousi-Yusifoğlu, G. Ruhe, J. Zhi, M.
Moussavi and B. Smith (2015). "Usage and usefulness
of technical software documentation: An industrial
case study." Information and Software Technology 57:
664-682.
Hall, J. J. (2014). "Quality Assurance in Business
Simulation Design." Developments in Business
Simulation and Experiential Learning 41.
HSIEH, C.-Y. and C.-T. CHEN (2015). "Patterns for
Continuous Integration Builds in Cross-Platform Agile
Software Development." JOURNAL OF
INFORMATION SCIENCE AND ENGINEERING
31: 897-924.
Hushalini, S., R. Randunu, R. Maddumahewa and C.
Manawadu (2014). "Software Test Automation in
Practice: Empirical Study from Sri Lanka." Compusoft
3(11): 1232.
Jawad, Z. and M. Taira (2015). "Improving Global
Software Development Tools to Facilitate Distributed
Agile Development." Software Engineering thesis,
University of Gothemburg, Sweden.
Khan, S. U., M. Niazi and R. Ahmad (2009). Critical
success factors for offshore software development
outsourcing vendors: A systematic literature review.
Fourth IEEE International Conference on Global
Software Egineering, 2009. ICGSE 2009. Limerick,
Ireland.207-2016.
Mahmoud, S. S. and I. Ahmad (2013). "A green model for
sustainable software engineering." International
Journal of Software Engineering and Its Applications
7(4): 55-74.
Misra, S., V. Kumar, U. Kumar, K. Fantazy and M.
Akhter (2012). "Agile software development practices:
evolution, principles, and criticisms." International
Journal of Quality & Reliability Management 29(9):
972-980.
Mohammad, A. H., T. Alwada’n and J. M. A. Ababneh
(2013). "Agile Software Methodologies: Strength and
Weakness." International Journal of Engineering
Science and Technology (IJEST) 5(03): 455-459.
Mohan, K., B. Ramesh and V. Sugumaran (2010).
"Integrating software product line engineering and
agile development." Software, IEEE 27(3): 48-55.
Naumann, S., M. Dick, E. Kern and T. Johann (2011).
"The greensoft model: A reference model for green
and sustainable software and its engineering."
Sustainable Computing: Informatics and Systems 1(4):
294-304.
Niazi, M., S. Mahmood, M. Alshayeb and A. Hroub
(2015). "Empirical investigation of the challenges of
the existing tools used in global software development
projects." Software, IET 9(5): 135-143.
Noll, J., S. Beecham and I. Richardson (2010). "Global
software development and collaboration: barriers and
solutions." ACM Inroads 1(3): 66-78.
Omar, M., S.-L. Syed-Abdullah and A. Yasin (2011). "The
impact of agile approach on software engineering
teams." American Journal of Economics and Business
Administration 3(1): 12.
Rashid, N., & Khan, S. U. (2014)."Green-Agile Maturity
Model (GAMM) for Global Software Development
(GSD) Vendors" (2014) The 1st Agile Conference
Pakistan (Islamabad), Sci. Int.(Lahore),vol. 26(5),pp.
2041 -2043.
Raturi, A., B. Tomlinson and D. Richardson (2015). Green
Software Engineering Environments. Green in
Software Engineering, Springer: 31-59.
Singh, A., K. Singh and N. Sharma (2015). "Agile in
global software engineering: an exploratory
experience." International Journal of Agile Systems
and Management 8(1): 23-38.
Stammel, J., Z. Durdik, K. Krogmann, R. Weiss and H.
Koziolek (2011). Software evolution for industrial
automation systems: literature overview, KIT, Fakultät
für Informatik.
Venters, C., C. Becker, S. Betz, R. Chitchyan, L. Duboc,
S. Easterbrook, B. Penzenstadler, G. Rodriguez-Navas
and N. Seyff (2015). "Mind the Gap: Bridging the
Sustainable Software Systems Research Divide."
Verner, J. M., O. P. Brereton, B. Kitchenham, M. Turner
and M. Niazi (2012). "Systematic literature reviews in
global software development: A tertiary study."
Wohlin, C. (2014). Guidelines for snowballing in
systematic literature studies and a replication in
software engineering. Proceedings of the 18th
International Conference on Evaluation and
Assessment in Software Engineering, ACM. p.38.
Wrubel, E. and J. Gross (2015). "Contracting for Agile
Software Development in the Department of Defense:
An Introduction.", Technical report (CMU/SEI2015
TN006),CarneigeMellonUniversity.
Zhang, H., M. A. Babar and P. Tell (2011). "Identifying
relevant studies in software engineering." Information
and Software Technology 53(6): 625-637.
Khan, S. U., & Niazi, M. (2010, June). A preliminary
structure of software outsourcing vendors' readiness
model. In Proceedings of the 11th International
Conference on Product Focused Software (pp. 76-79).
ACM.
ENASE 2016 - 11th International Conference on Evaluation of Novel Software Approaches to Software Engineering
252
Niazi, M., A Framework for Assisting the Design of
Effective Software Process Improvement
Implementation Strategies. Journal of Systems and
Software.2007,78(2):204-222.
Rashid, N., & Khan, S. U. (2015). Green Agility for
Global Software Development Vendors: A Systematic
Literature Review Protocol. Proceedings of the
Pakistan Academy of Sciences 52(4): pp.301-313.
Khan, R. U., Khan, S. U., Khan, R. A., & Ali, S. (2015).
Motivators in Green IT-outsourcing from Vendor’s
Perspective: A Systematic Literature Review.
Proceedings of the Pakistan Academy of Sciences
52(4): pp.345-360.
Alzoubi, Y. I., Gill, A. Q., & Al-Ani, A. (2016). Empirical
studies of geographically distributed agile
development communication challenges: A systematic
review. Information & Management, 53(1):pp. 22-37.
Kitchenham, B. A., & Charters, S. (2007). Procedures for
performing systematic literature reviews in software
engineering. Keele University & Durham University,
UK.
Silva, F. S., Soares, F. S. F., Peres, A. L., de Azevedo, I.
M., Vasconcelos, A. P. L., Kamei, F. K., & de Lemos
Meira, S. R. (2015). Using CMMI together with agile
software development: A systematic
review. Information and Software Technology, 58, 20-
43.
Developing Green and Sustainable Software using Agile Methods in Global Software Development: Risk Factors for Vendors
253