Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/334854501

Global Software Development and Capability Maturity Model Integration: A


Systematic Literature Review

Conference Paper · October 2018


DOI: 10.1109/IAC.2018.8780489

CITATIONS READS

10 304

4 authors, including:

Anita Hidayati Betty Purwandari


Politeknik Negeri Jakarta University of Indonesia
31 PUBLICATIONS 65 CITATIONS 105 PUBLICATIONS 1,070 CITATIONS

SEE PROFILE SEE PROFILE

Iis Solichah
University of Indonesia
19 PUBLICATIONS 179 CITATIONS

SEE PROFILE

All content following this page was uploaded by Anita Hidayati on 17 November 2020.

The user has requested enhancement of the downloaded file.


Global Software Development and
Capability Maturity Model Integration:
a Systematic Literature Review
Anita Hidayati, Betty Purwandari, Eko K. Budiardjo, Iis Solichah
Faculty of Computer Science
Universitas Indonesia
Depok, Indonesia
anita.hidayati@ui.ac.id, bettyp@cs.ui.ac.id, eko@cs.ui.ac.id, iis@cs.ui.ac.id

Abstract—Global Software Development (GSD) offers world. GSD can accelerate product development with high
many benefits. On the other hand, it also has several challenges quality and low-cost products, which are closely associated
such as its impact on software quality. The implementation of with overseas markets [3], as well as the need to share risks.
software processes improvement, such as the Capability GSD also provides benefits for vendors, including
Maturity Model Integration (CMMI), can improve quality of
enhancing the growth of developing countries [4], offering
software products. However, CMMI as a process improvement
model has been rarely discussed in GSD context. To fill this new motivations such as opportunities for traveling and
gap a systematic literature study was conducted by using interacting with different cultures [5]. The most important
Kitchenham’s method. It investigated GSD research from thing is the technology transfer. As a result, it increases the
several databases between 2013-2018. At the end there are 12 ability and experience for IT personnel involved in the GSD
selected papers. The findings indicate that CMMI is essential team.
for the success of GSD projects. However, it requires multiple GSD has unique characteristics, because it is done in a
customizations and combinations with other standards. geographically dispersed location across national
Concept Process Areas (PAs) and maturity level from CMMI boundaries. It involves communication for information
are used in new standards Software Process Improvement
exchange, as well as group coordination, activities and
(SPI) with different approaches. The results of this study are
potential as a reference to develop customized CMMI to artifacts to achieve the overall goal. The coordination
increase success rate in GSD projects. involves real time (synchronous) and asynchronous
interactions, group control (following goals and policies),
Keywords—Global Software Development (GSD), Capability and artefacts (quality, visibility and management).
Maturity Model Integration (CMMI), Systematic Literature GSD is different from Open Source Software (OSS),
Review (SLR), software process improvement, Process Area (PA) Traditional Software Development, and Multi National
Corporation (MNC) as depicted in Fig. 1. More detailed
I. INTRODUCTION comparisons are shown in Table 1 to 3.
Global Software Development (GSD) is also called
Global Software Engineering (GSE), Distributed Software
Development (DSD) [1], IT-outsourcing [2], Offshore
Outsourcing or Overseas Outsourcing [3]. It is an
outsourcing technique with team members coming from
different nationalities and cultural backgrounds. Outsourcing
is an activity in which an organization, which is called a
vendor or service provider, provides services to other
organizations as their customers or clients at predetermined
prices. There is agreement on quality criteria and specified
time schedules [1]. Overseas or offshore outsourcing is a
form of outsourcing conducted outside the country of the
origin of the organization, contracting out work to third
parties in other countries. Outsourced activities include the
transfer of non-core activities to suppliers for the purpose of Fig. 1. Software Development Paradigms
securing special skills and lower costs [3].
TABLE 1. GSD vs. OSS [6]
GSD becomes more widely applied, because it has many GSD OSS
advantages for both the vendors and clients. Client There is a target time, performance. Nothing. It depends on the
companies do not have to send power from their own liveliness of the community.
There is a contract to bind the client There is freedom for anyone
country to branches abroad. They access world-class and the vendor involved. to engage and contribute.
capabilities including new technologies, tools, There are rules, procedures, tools and Simple communication tool.
methodologies and procedures, which do not exist in the ways of communicating sophisticated
organization [1]. It optimizes and reduces costs by finding a and diverse.
zone, where there are more ready labors and expert
TABLE 2. GSD vs. TRADITIONAL SOFTWARE DEVELOPMENT [7]
resources at the least cost. It provides the easiest approach to GSD Traditional Software
24/7 customers known as "Follow the Sun (FTS)" [4]. It Development
especially happens if the branches are spread around the

XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE

Authorized licensed use limited to: Universitas Indonesia. Downloaded on November 17,2020 at 08:32:03 UTC from IEEE Xplore. Restrictions apply.
It must use communication It is easy to communicate RQ2: Are there any maturity levels in GSD and how to
techniques, which can overcome the directly. measure?
barriers of distance, time, language
and culture.

GSD Traditional Software The aim of these research questions is to map GSD
Development studies with the CMMI to find out how the CMMI concepts
It requires huge consequences, when a It is easy to make project applied in GSD. This can be a reference to advance software
project changes in the middle of the changes. process quality in GSD.
road.
All employees are actively involved in Employees are unaware of
innovation and project development. innovation and project II. CMMI
development.
Software process is defined as a set of activities,
TABLE 3. GSD vs. MNC [3] methods, practices and transformations that people use to
GSD MNC build and maintain software and related products [16]. There
It consists of 1 client and 1 It belongs to one company. It can be are several process improvement models such as CMMI,
vendor or more. in multiple locations. CMM, ISO 9000 and ISO/IEC 15504. By applying one of
The relationship is the working There is a company's organizational
relationship between client and structure.
the process improvement models, the process of making
vendor. software products will take a shorter time but the quality is
not decreased, even increased.
The uniqueness of GSD creates challenges to adopt it, CMMI is one of the maturity models used to improve the
such as how to manage its impact on software quality [8]. In process (process environment) within the organization.
general, an organization must improve its software process CMMI is a systematic and structured collection of best
to advance the quality of its software product. The SPI practices for process improvement and assessment [12]. The
(Software Process Improvement) model is potential to main objective of CMMI is to reduce the cost of implement
increase productivity and efficiency of the implementation improvements in processes. By eliminating inconsistencies
within the organization in efficient and effective ways [9]. It and establishing guidelines, CMMI reduces the cost of
also leads software organizations to develop good products implement improvement in processes.
[10]. The SPI plays a significant role to make GSD projects In CMMI there are 22 Process Areas (PAs) grouped in
successful [11]. Deployment of SPI in GSD context is more four functional groups as follows [17]:
challenging [12]. CMMI is a process model framework for
SPI. There are several studies on the relation between GSD • Process Management: Organizational Process Focus
and CMMI. Several models and standards of software (OPF), Organizational Process Definition (OPD),
process improvement have been designed including CMMI Organizational Training (OT), Organizational Process
for enhancing software processes. Performance (OPP), Organizational Performance
There have been researchers, who developed state of the Management (OPM)
art of SPI, e.g. CMMI and ISO/IEC [9]. They gave future • Project Management: Project Planning (PP), Project
work to study on SPI in the field of GSD. Richardson et al Monitoring and Control (PMC), Supplier Agreement
[13] looked for a gap analysis between CMMI and the Management (SAM), Risk Management (RSKM),
application of GSD. Afterwards they proposed a new project Integrated Project Management (IPM), Quantitative
management process area called Global Teaming (GT) to Project Management (QPM), Requirement
address specific problems relating to cultural, geographic, Management (REQM)
temporal, and linguistic distance, which meet with the
complex and changing needs of GSD. These can be used as a • Engineering: Requirement Development (RD),
guidance to implement GSD. Technical Solution (TS), Product Integration (PI),
Verification (VER), Validation (VAL)
Dayyala studied key elements affecting the performance
of software development projects based on software process • Support: Configuration Management (CM), Process
components [14]. He investigated team characteristics that and Product Quality Assurance (PPQA),
affect the output of global software projects based on the Measurement and Analysis (MA), Decision Analysis
CMMI level. Moreover, Persson and Schlichter discussed and Resolution (DAR), Causal Analysis and
how and why managers were willing to accept clients from Resolution (CAR)
companies with level 5 CMII maturity [15]. They also CMMI defines five-level maturity that range from initial
studied the implications on theory and practice in software (1), managed (2), defined (3), and quantitatively managed (4)
development offshoring. to optimizing (5) at the highest level as seen on Fig. 2. The
In four previous studies, there are correlations between maturity level provides a way to characterize its capability
GSD and CMMI. There is an effort to develop process area and performance. Maturity levels represent a staged path for
from the existing CMMI. There are also special discussion performance and process improvement effort of
about the characteristics of the team and the reason in the organization. CMMI is successful when operating in a local
selection of clients based on CMMI. However, more studies environment, but it does not discuss the complexity of GSD
are required to understand the interplay between GSD and in detail [18] whereas GSD is fundamentally different to
CMMI. Hence, this study will address the following research collocated development.
questions:
RQ1: How does GSD relate to CMMI or its variant?

Authorized licensed use limited to: Universitas Indonesia. Downloaded on November 17,2020 at 08:32:03 UTC from IEEE Xplore. Restrictions apply.
DATABASES

Running search string

- Indexed Journal and N=178


Proceeding
- English Version

N=86

Title, keyword and


abstract
N=17
Fig. 2. CMMI Maturity Levels [19] Full text

III. RESEARCH METHODOLOGY N=8+4(snowballing)


= 12
The research methodology of this article is based on
guidelines proposed by Kitchenham [20] in identifying and Fig. 3. The Selection Process
evaluating all sources that related with research question.
Kitchenham is one of the most widely used SLR methods in
software engineering. To find the literature sources used IV. RESULTS AND FINDINGS
some academic databases such as SpringerLink, ProQuest, This section reports the results and findings from the
ScienceDirect, IEEE Xplore, EBSCOhost and ACM Digital research questions for this study. Global summary of selected
Library. The literature studied is from 2013 and 2018 papers can be seen in Table 5.
semester 1. 178 total papers used in this study with the
composition in Table 4. EBSCOhost dominates with most RQ1: How does GSD relate to CMMI or its variant?
papers totaling 69, whereas in ACM has no paper. There are still many developing countries which are
For article selection, use of terms: global software reluctant to use CMMI. In Turkey and Nigeria, CMMI is
development, global software engineering, distributed only followed if it is explicitly specified as a requirement of
software development, GSD, GSE and CMMI. The query project. CMMI is only as a brand or a label and just for 'show
string used is: off', the product qualities are not necessary [21]. If
organizations have obtained CMMI, they do not follow their
("GLOBAL SOFTWARE DEVELOPMENT" OR prescribed directives. CMMI is considered to slow the
"GLOBAL SOFTWARE ENGINEERING" OR development process and it is not so efficient in practice for
"DISTRIBUTED SOFTWARE DEVELOPMENT" OR small software companies. The majority of software
"GSD" OR "GSE") AND ("CMMI" OR “CAPABILITY development firms in Barbados, Guyana, Jamaica and
MATURITY MODEL”) Trinidad (English-Speaking Caribbean), were not aware of
This study is included based on criteria: nor using CMMI [22]. With less attention to CMMI, it
results in poor of quality software being delivered in
• Study describes about CMMI in GSD context. developing countries. Meanwhile in a developed country,
Finland, SMEs companies do not adopt CMMI because its
• Study is published in a Scopus-indexed journal or benefits are not clear [23].
conference.
Having international quality certifications such as CMMI
Studies are excluded based on criteria: is important for the success of offshore software industries
• The domain is not in the field of computer science. [24]. The standard model, such as CMMI, is still needed in
GSD, one important is documentation [25]. With CMMI
• Not in English version. certification, GSD vendors maintain stringent documentation
• The terms are not explicit in title, keywords or of projects undertaken by them. All artifacts (data flow
abstract. charts, software code, test scripts, etc.) have strict
documentations. These help understanding of partners about
The last exclude is based on content of full text. The common project expectations.
result is 8 full-reviewed papers. There 4 papers obtained by
snowballing method, so total papers are 12. So that CMMI can be relevant in the context of GSD and
widely applied, it is devoted to develop customization [26].
TABLE 4. DATA SOURCES Customization of the area process must be performed to suit
E-Databases Total of First Second Final different GSD characteristics with collocated, especially
Articles Selection Selection Paper
when associated with culture, coordination, agile concept,
SpringerLink 25 9 2 1
ProQuest 46 22 5 2 et.al. New standards such as Global Teaming [27][28],
ScienceDirect 32 30 7 4 SPIIMM [29], SOPM [30], PAM [31] and an ontological
IEEE Xplore 6 6 2 0 framework [32] have been built to accommodate the needs of
EBSCOhost 69 29 1 1 GSD. Although these standards are not purely CMMI, they
ACM Digital Library 0 0 0 0 can be called as CMMI variants, but they still use the basic
Total 178 86 17 12 (8+4) concept of CMII, such as process areas and maturity levels.

Authorized licensed use limited to: Universitas Indonesia. Downloaded on November 17,2020 at 08:32:03 UTC from IEEE Xplore. Restrictions apply.
TABLE 5. RELATION GSD WITH CMMI taken from CMMI. In some areas, there are replacements of
Sources Related with CMMI PAs. This is because PAs from other standards are more
[21] CMMI is only followed if it explicitly specified as a appropriate and have wider possible definitions, so they can
requirement, and just for ‘showoff’.
In ESC, key determinants of process maturity are
fit the context GSD much better.
[22]
verification & validation and project monitoring & control. RQ2: Are there any maturity levels in GSD and how to
[23] Benefit of CMMI is not clear for small company.
measure it?
[24] CMMI is important for the success of offshoring.
[25] Certifications from CMMI can push maintaining stringent In PAM, there is no maturity level of organization like
documentation of all projects. CMMI. The maturity which is meant in PAM is maturity of
[26] More pragmatic approach is more likely, in Scandinavia the PA. The average rating of a PA is used to compare PAs
toward to “agile” era, so need to customize CMMI.
and show the maturity of the PA. The higher the value, the
[27] Global Teaming is made based on empirical study in GSD
with deriving CMMI template for governance. better the maturity.
[28] DSS for global teaming. SPIIMM has maturity levels indicates the capability of
[29] SPIIMM is based on existing process development models, an organization in process improvement. SPIIMM maturity
including CMMI. levels is based on CMMI and two others. Each maturity
[30] SOPM categorizing vendors’ capability level based on level consists of different Critical Success Factors (CSFs)
CMMI maturity levels, and on partnership CSFs based on
Pas
and Critical Barriers (CBs). Its assessment component is
[31] PAM takes all top-level PAs from CMMI and replace PAs used to evaluate maturity levels of an organization. But
with other standards as a way that fits context of GSD. SPIMM does not have a detailed explanation about the
[32] Ontologicall framework based on a multi-model approach, measurement of maturity level, because it is still a model
including CMMI concept. design.

RQ2: Are there any maturity levels in GSD and how to V. CONCLUSIONS
measure? From this research can be drawn conclusion as follows:
An initial Software Process Improvement • CMMI as a SPI is still essential for the success of
Implementation Management Model (SPIIMM) was GSD projects. However, it has not been able to
developed to assist GSD organizations to measure and answer the challenges in GSD. Multiple
improve their SPI activities. It will tackle the problems customizations and combinations with other standards
associated with implementation of SPI, which is vital to the are required to make it be more flexible and able to fit
GSD organizations success and progression. It used CMMI the needs and relevance of GSD.
and other existing process development models to structure
SPIIMM, as shown in Fig. 4. CMMI as one of references to • There are some new standards SPI as variants of
determine maturity levels. Critical Success Factors (CSFs) CMMI. They referred to the CMMI concept in terms
dan Critical Barriers (CBs) are also identified based on of PAs and maturity levels.
CMII. • There are maturity levels in GSD that are derived
Process Area Map (PAM) is one example reference from CMMI, for example SPIIMM, with combination
frameworks and standards like CMMI. It structures and of CSFs and CBs. But there is no measurement of
categorizes available solutions (tool, process, model, maturity levels.
technique, and so on) to tackle challenges in GSD into main • There are still open opportunities for research in GSD
and sub-process areas. It is used to evaluate and improve especially in improving product quality through
development environment in a structure way. Fig. 5 shows customization of existing process models.
the structure of PAM looks like a tree. It was constructed
using several sources, with CMMI as the basis of the The results of this study will be developed for further
structure. All top-level PAs (Process Management, Project research to standardize new models by combining existing
Management, Engineering, Support) and most of the PAs are standards and adoption to the GSD.

Fig. 4. Architecture of SPIMM Model [29]

Authorized licensed use limited to: Universitas Indonesia. Downloaded on November 17,2020 at 08:32:03 UTC from IEEE Xplore. Restrictions apply.
Fig. 5. Structure of PAM [31]

[9] M. Kuhrmann, P. Diebold, and J. Münch, “Software process


ACKNOWLEDGEMENT improvement : a systematic mapping study on the state of the
This research was supported by a PITTA grant at the art,” no. 2012, 2016.
University of Indonesia (1868/UN2.R3.1/HKP.05.00/2018). [10] A. A. Khan and J. Keung, “Systematic review of success factors
and barriers for software process improvement in global software
REFERENCES development,” IET Softw., vol. 10, no. 5, pp. 125–135, 2016.
[1] K. Siakas, “Cultural and Organisational Diversity Evaluation ( [11] A. A. Khan, J. Keung, M. Niazi, and S. Hussain, “Towards a
CODE ) A tool for improving global transactions,” 2015. hypothetical framework of humans related success factors for
[2] J. Palvia and C. Shailendra, “Outsourcing issues and challenges process improvement in global software development,” Proc.
facing CEOs and CIOs,” J. Inf. Technol. Cases Appl., vol. 7, no. Symp. Appl. Comput. - SAC ’17, pp. 180–186, 2017.
4, pp. 1–6, 2005. [12] M. Niazi, M. A. Babar, and J. M. Verner, “Software process
[3] P. Hong, “Strategic outsourcing practices of multi-national improvement barriers: a cross-cultural comparison,” Inf. Softw.
corporations ( MNCs ) in China,” 2014. Technol, vol. 52, pp. 1204–1216, 2010.
[4] J. Kroll, I. Richardson, R. Prikladnicki, and J. L. N. Audy, [13] I. Richardson, V. Casey, F. McCaffery, J. Burton, and S.
“Empirical evidence in follow the Sun software development : A Beecham, “A process framework for global software engineering
systematic mapping study,” vol. 93, pp. 30–44, 2018. teams,” Inf. Softw. Technol., vol. 54, no. 11, pp. 1175–1191,
[5] S. Beecham et al., “Challenges and Recommendations for the 2012.
Design and Conduct of Global Software Engineering Courses : A [14] N. Dayyala, “Essays On Sofwtare Development Projects: Impact
Systematic Review Protocol Contact Information :,” 2015. Of Social And Technological Factors On Project Performance
[6] P. Mohagheghi, “Global software development: Issues, solutions, And Co-Diffusion Of Software Sourcing Arrangements,” The
challenges,” Retrieved April, no. September, pp. 1–26, 2004. University Of Texas At El Paso, 2017.

[7] G. Gaughan, B. Fitzgerald, and M. Shaikh, “An Examination of [15] J. Stouby and B. Rerup, “Managing Risk Areas in Software
the use of Open Source Software Processes as a Global Software Development Developm Offshoring : A CMMI Level 5 C Case,”
Development Solution for Commercial Software Engineering .,” J. Inf. Technol. theory Appl., vol. 16, no. 1, p. 5=24, 2015.
2009. [16] P. Mark, C. V. Weber, C. Bill, and M. B. Chrissis, The capability
[8] S. Mahmood, S. Anwer, M. Niazi, M. Alshayeb, and I. maturity model: Guidelines for improving the software process.
Richardson, “Key factors that influence task allocation in global Addison-Wesley, 1995.
software development,” Inf. Softw. Technol., vol. 91, pp. 102– [17] D. Proença and J. Borbinha, “Formalizing ISO/IEC 15504-5 and
122, 2017. SEI CMMI v1.3 – Enabling automatic inference of maturity and

Authorized licensed use limited to: Universitas Indonesia. Downloaded on November 17,2020 at 08:32:03 UTC from IEEE Xplore. Restrictions apply.
capability levels,” Comput. Stand. Interfaces, vol. Volume 60, pp. [26] M. Kuhrmann and P. Diebold, “How Does Software Process
13–25, 2018. Improvement Address Global Software Engineering ?,” in IEEE
[18] I. Richardson, V. Casey, J. Burton, and F. Mccaffery, “Global 11th International Conference on Global Software Engineering
Software Engineering : A Software Process Approach,” in How, 2016, pp. 89–98.
Collaborative Software Engineering, Springer, 2010, pp. 35–57. [27] J. Noll, S. Beecham, I. Richardson, and C. N. Canna, “A Global
[19] S. Godfrey, “What is CMMI?,” NASA Presentation. 2008. Teaming Model for Global Software Development Governance:
A Case Study,” 2016 IEEE 11th Int. Conf. Glob. Softw. Eng., pp.
[20] B. Kitchenham, O. Pearl Brereton, D. Budgen, M. Turner, J.
179–188, 2016.
Bailey, and S. Linkman, “Systematic literature reviews in
software engineering - A systematic literature review,” Inf. Softw. [28] S. Beecham, J. Noll, and I. Richardson, “A Decision Support
Technol., vol. 51, no. 1, pp. 7–15, 2009. System for Global Software Development,” in Sixth IEEE
International Conference on Global Software Engineering
[21] O. Y. Sowunmi, S. Misra, L. F. Sanz, B. Crawford, and R. Soto,
Workshops, 2011, pp. 48–53.
“An empirical evaluation of software quality assurance practices
and challenges in a developing country : a comparison of Nigeria [29] A. A. Khan, J. Keung, M. Niazi, S. Hussain, and A. Ahmad,
and Turkey,” Springerplus, pp. 1–14, 2016. “Systematic literature review and empirical investigation of
barriers to process improvement in global software development:
[22] E. C. Firms, “Key Factors of Process Maturity in English-
Speaking Caribbean Firms,” pp. 322–330, 2014. Client–vendor perspective,” Inf. Softw. Technol., vol. 87, pp.
180–205, 2017.
[23] S. Wagner, M. K. M. Felderer, T. Conte, M. C. D. Greer, and P.
[30] S. Ali and S. U. Khan, “Software outsourcing partnership model:
M. A. Vetr, “Naming the pain in requirements engineering
An evaluation framework for vendor organizations,” J. Syst.
Contemporary problems , causes , and effects in practice,” pp.
Softw., vol. 117, pp. 402–425, 2016.
2298–2338, 2017.
[31] S. Schneider, R. Torkar, and T. Gorschek, “Solutions in global
[24] G. P. Sudhakar, “A Review of Critical Success Factors for
software engineering: A systematic literature review,” Int. J. Inf.
Offshore Software Development Projects,” vol. 46, no. 6, pp.
Manage., vol. 33, no. 1, pp. 119–132, 2013.
282–296, 2013.
[32] J. Mejia, E. Muñoz, and M. Muñoz, “Reinforcing the
[25] A. Mathrani and S. Mathrani, “Test strategies in distributed
applicability of multi-model environments for software process
software development environments,” Comput. Ind., vol. 64, no.
1, pp. 1–9, 2013. improvement using knowledge management,” Sci. Comput.
Program., vol. 121, pp. 3–15, 2016.

Authorized licensed use limited to: Universitas Indonesia. Downloaded on November 17,2020 at 08:32:03 UTC from IEEE Xplore. Restrictions apply.
View publication stats

You might also like