Agility of Security Practices and Agile Process Models: An Evaluation
of Cost for Incorporating Security in Agile Process Models
H. Maria Maqsood and Andrea Bondavalli
Department of Mathematics and Informatics, University of Florence, Morgagni, Florence, Italy
Keywords: Software and Systems Development Methodologies, Software Process Improvement, Empirical Software
Engineering, Agile Methodologies, Agility, Process Models, XP, Scrum, Agile Process Models.
Abstract: Agile process models are widely used today for software development. There has been an immense increase
in use of agile methodologies due to their major focus on delivering working software and accommodating
changes in requirements. However, use of agile methodologies for developing secure systems still poses
many challenges. This research, addresses the issue of observing the effect on agility of process models
while security practices are applied in them. An approach is proposed which calculates level of agility of six
agile process models (XP, Scrum, FDD, ASD, DSDM, and Crystal) and security practices against four
fundamental parameters of agility. When security practices are applied to process models they lower the
degree of agility. We propose a method to see this effect based on factor of agility and also that the degree
of agility of process model can be adjusted at desired level by including or excluding security practices.
1 INTRODUCTION
Agile focuses on rapid development and follows the
rule of delivering working software in short
intervals.
One of the major advancement in the field of
software is that from past decade developers and all
stakeholders consider security as a proper issue
(McGraw, Allen, Barnum, & Ellison, 2008). The
current concern is we do not have very compact
solutions to address the problem.
This paper is structured into five major sections.
Section 1 gives introduction, section 2 gives brief
overview of related work to understand the problem
in hand, section 3 describes the methodology to
calculate agility of process models and lists details
of a research survey that we have performed to
calculate the agility of security practices in process
models. Section 4 gives the formula to calculate the
agility of models after the application of selected
security practices.Section 5 concludes the work done
and Section 6 describes future work directions.
2 RELATED WORK
According to Jacobson (Jacobson, 2002), an agile
team is very responsive to changes since adapting to
change is what agile software development is all
about. It is very important for an agile team to
understand that software is developed by teamwork,
and collaboration is the heart of success (Boström,
Gustav, & et al., 2006).
Software engineers gathered forces and began to
classify agile processes in early 2001 (K.Beck,
2001). Agile Alliance stated the agile manifesto as
(K.Beck, 2001)
“Individuals and interactions over processes and
tools, Working software over comprehensive
documentation, Customer collaboration over
contract negotiation; Responding to change over
following a plan
Security is an afterthought during software
development, it is addressed either very late in
development or even after it (Kravchenko, Elena, &
E. W., 2017).
Now the question arises what will we call a
secure product? According to Microsoft (Howard &
Lipner, 2006), a secure product is the one that can
handle the integrity, confidentiality and customer
Maqsood, H. and Bondavalli, A.
Agility of Security Practices and Agile Process Models: An Evaluation of Cost for Incorporating Security in Agile Process Models.
DOI: 10.5220/0009356403310338
In Proceedings of the 15th International Conference on Evaluation of Novel Approaches to Software Engineering (ENASE 2020), pages 331-338
ISBN: 978-989-758-421-3
Copyright
c
2020 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
331
information along with the confidentiality of
processing resources under administrator.
Many experts give advice on using agile
methodologies for the development of secure
systems (Moyon, Beckers, & Kleppe, 2018) as there
are reported benefits of using agile for software
development; however, security cycles are in
contrast to agile approaches with the consequence of
compromising the level of agility when
incorporating security practices (Alnatheer, Ahmed,
Gravell, Andrew and Argles, & David, 2010). There
is no concrete method to monitor how agility will be
affected. Security in its very own nature as a non-
functional requirement is not easy to cater to. Still,
it is the experience in this area that counts the most
(K., S., & V., August, 2017) (Ashraf, S., & Aftab, &
S., 2017).
3 METHODOLOGY
We provide a method to assess (numerically) agility
of process models and security practices. There are
two major contributions of this work, first is related
to the degree of agility for six process models and
second is about the degree of agility of twelve
security practices in Scrum and XP. The agility of
security practices varies from one process model to
another; hence, there are dedicated tables to show
values of security practices for both of our chosen
process models.
The agility of a security activity or a process
model refers to how an activity/process model
behaves against basic parameters given by agile
manifesto. The capacity of a security
activity/process model to be flexible, lean,
responsive and speedy defines how agile it is. If it
possesses higher values of these attributes, it has a
high degree of agility and vice versa. We have
further assessed the agility of process models after
including selected security practices by using a
formula. The variation in the degree of agility of a
process model before and after application of
security practices shows how much agility is
compromised to incorporate security.
The four parameters of agility that we have used
in our work are defined by agile manifesto (K.Beck,
2001) as
Flexibility
Ability to adapt to expected or unexpected
changes at any time.
Leanness
It refers to the improvement of products and
services based on the feedback of customers in
terms of what they value.
Responsiveness
Responsiveness refers to appropriate reaction
against expected or unexpected changes.
Speed
Speed refers to rapid and iterative
development for small releases.
It is important to understand that they are not
ranked in any Order. A process must have all of
these attributes to be called an agile process.
3.1 Agility of Process Models
We have evaluated the agility of six agile process
models against four basic parameters of agility based
on the work of (Qumer, A., Henderson-Sellers, & B.,
2008). They proposed an approach to calculate the
agility of process models in terms of their phases
and practices however, We have taken both practices
and phases into account and computed a single value
that represents the degree of agility of certain model.
Following are six process models that we have
considered.
XP(Extreme Programming),
Scrum,
ASD(Adaptive Software Development),
DSDM(Dynamic System Development
Method),
FDD(Feature driven development) and
Crystal.
However, we have used XP and Scrum only for
further evaluations, since we were not able to have
significant amount of responses for rest of the
process models through our survey.
Extreme programming (XP) is an agile process
model, which intends to improve responsiveness and
software quality and cater to flexible requirements
(Beck & Kent, 2004).There are four basic phases of
XP
Planning,
Designing,
Coding and
Testing.
Scrum (Ken Schwaber & Beedle, 2002) has
major focus on courage, respect, openness and
commitment. Scrum has the following basic
activities
Product backlog,
Building teams,
Scheduled meetings,
Sprint and
Sprint review.
ENASE 2020 - 15th International Conference on Evaluation of Novel Approaches to Software Engineering
332
The calculation of agility of process models is
based on the work of Qumer B.Henderson (Qumer,
A., Henderson-Sellers, & B., 2008). They derived a
formula to present the agility of each process
model’s phases and practices within the range 0 to 1
where 1 represents that the process model is highly
agile. In this paper, we have calculated the agility of
a process model as whole, including practices and
phases.
We will use following equation to perform
calculations,
Degree of agility of process model = Sum
of agile factors / (no. of practices of process
model * no. of agile attributes)
(1)
For example, let us consider the process model
XP.
Sum of agile factors in XP [flexibility + speed
+ leanness + responsiveness] = [15+13+6+15]
=49
Number of practices + phases in XP = 18
Number of agile attributes = 4
By applying the formula (1) we get
49÷ (18*4)=0.68
Table 1: Process Models.
Agile
Attributes
Process
Models
Flexibility
Leanness
Speed
Responsiveness
Sum of agile factors
No. of practice + phases in a
process model
Total
Agility
of
process
model
XP 15 6 13 15 49 18 0.68
Scrum 9 0 10 9 28 10 0.7
ASD 10 0 12 10 32 12 0.66
FDD 10 0 10 10 30 13 0.57
Crystal 9 0 11 11 32 11 0.70
DSDM 10 0 11 11 32 15 0.53
It is evident that some process models have
higher agility as compared to others. As given in
Table 1 Crystal and Scrum have highest degrees of
agility and DSDM has lowest, which refers to the
fact that Crystal and Scrum are more agile, and
DSDM is least agile in this set of process models.
However, our focal point is not comparison
between process models; our focus is to see how
agility of any one-process model is affected after
including security practices.
3.2 Agility of Security Activities in
Process Models
We have considered the degree of agility of few of
the most widely used security activities against four
agile parameters (as we did for process models).
This approach is based on the work of (Hossein
keramati & Mirian-Hosseinabadi, 2008). The values
of security attributes are based on survey from
industry personnel. There was a need to perform this
survey because as per our knowledge there is no
numerical data available for such reasoning. We
could only find theoretical reasoning through
literature, hence we conducted a survey to provide
grounds for empirical analysis.
3.2.1 Research Survey
We performed a questionnaire-based survey to
observe the agility of 12 most common security
practices against four agile parameters in six process
models. The research survey serves as descriptive
survey and it provides a descriptive analysis. As
Oppenheim, (Oppenheim, 2000) describes a
descriptive survey provides descriptive analysis
only, which refers to frequencies and cross
tabulation. According to Oppenheim, (Oppenheim,
2000) descriptive surveys are not meant to describe
causal relationship of variables instead their focus is
on describing what proportions of sample represent
certain opinion or what is the frequency of
occurrence of certain events/values.
While selecting the organizations for responses
we used purposive sampling. Nardi (Nardi, 2014)
describes this method as collecting samples from
respondents based on some specific trait, which is
important for the study. In this research survey, we
have involved organizations that have experience of
working with agile AND are using security
techniques in agile methods.
The research was designed to get responses of
individuals’, project managers or developers
working with agile methodologies. One person can
give response for more than one process model
according to his experience in relevant
model/models. All responses were collected through
a web-based survey using Google forms. (Alreck,
P.L., Settle, & R.B, 1995) Mentions three major
methods for collecting data when performing
questionnaire-based survey. First is Personal
interviewing second is mail data collection and third
is telephone interviewing. However, now days there
is very popular method of conducting web based
survey. There are many reasons for selecting web-
Agility of Security Practices and Agile Process Models: An Evaluation of Cost for Incorporating Security in Agile Process Models
333
based surveys. All the data you get through them is
already in electronic form so it is easy and fast to
access the data. In addition, it prevents the chances
of errors during manual entries of data as Nardi
(Nardi, 2014) suggests. This method is speedy and
cost effective too when compared to mail based
surveys. It also proves to have higher response rate,
which is one of the major issues with other methods
of surveys.
3.2.2 Design of Research Survey
We conducted a survey through questionnaire by
using 5 point Likert scale (Likert, 1932) method.
The responses were collected through web-based
questionnaire. Sample is given in Appendix A. We
asked industry experts to rank the security activities
against four parameters of agility on the scale of
(Alreck, P.L., Settle, & R.B, 1995) (Rea, L.M., &
R.A., 1997)
5= Strongly Agree,
4=Agree,
3=Neutral,
2=Disagree and
1=Strongly Disagree.
Each expert has chosen the process model of his
or her expertise and ranked security attributes for
that particular process model (Jon A. Krosnick &
Presser, 2010).
For example, an expert of XP ranks the security
practice “attack identification” as 5 (strongly agree)
against agile attribute “flexibility”, this represents
that expert strongly agrees that attack identification
is highly flexible in XP. OR in other case, if he
assigns 1 (strongly disagree) this refers to not
flexible at all in this case. Flexibility and other three
agile parameters are defined in accordance to agile
manifesto (described in detail in section 3).
Figure 1
represents sample sources.
In our questionnaire, first two questions were
dichotomous questions that represent an exclusive
disjunction. Our first question, “I have experience of
agile process models for more than 5 years”.
Question 2 states, “I have more than 3 and less than
5 years of experience with agile process models and
security practices”. The detail of responses is shown
in Figure 2. In total, we received 56 responses.We
received 18 responses for Scrum out of which 10
had experience with agile and security for less than 3
years, 6 had more than 5 years of experience in agile
whereas 2 had more than 3 years of agile experience
with security practices.
For XP we got 20 responses, 10 of them have
worked
with security practices in agile
for
less
than
Figure 1: Participants with Experience in Specific Process
Model.
3 years, 8 had greater than 5 years of agile
experience and 2 had more than 3 years of agile
experience. For Feature Driven Development we
encountered five responses, three of them had
experience with security attributes and agile for less
than 3 years whereas 1 had agile experience of more
than 5 years and last respondent has experience of
more than 3 years. For Adaptive Software
Development we received 6 responses 3 of which
have worked with security and agile processes for
less than 3 years and 2 of them had experience of
more than 5 years in agile development whereas 1
had experience of more than 3 years with agile and
security practices. For DSDM (Dynamic System
Development Method) we had 5 respondents, 3 of
them had experience with agile process models and
security attributes for less than 3 years, 1 had more
than five years of agile experience and last one had
more than three years of agile and security
experience solely. For Crystal process model we
received two responses 1 had experience with agile
and security attributes for more than 3 years and 1
had experience with agile for less than 3 years.
Further evidence is required for process models,
to deduce more reliable conclusions about their
agility. However, in this paper we provide the
application of our methodology on two process
models namely Scrum and XP, since we have better
number of responses against these models as
compared to rest of the four process models.
Furtherwork is required to verify and validate the
results of process models with lower number of
responses.
3.2.3 Results
The tables in this section represent the final values
of security practices against agile parameters for
Scrum and XP based on the expert’s opinions taken
through a survey. We have processed the data of the
survey by using a formula given in (2). First we have
assigned weights to the responses.
ENASE 2020 - 15th International Conference on Evaluation of Novel Approaches to Software Engineering
334
Weightage 3 is assigned to category A which has
people with more than five years of experience,
weightage 2 is assigned to category B which has
people with more than 3 and less than 5 years of
experience and finally category C has weightage 1
for responses from people with less than 3 years of
experience.
The value of each agile attribute against each
security practice is calculated by using the following
formula,
= (sum of values by category A*3) + (sum
of values by category B*2) + (sum of values
by category C*1) ÷ (Total no. of responses of
category A*3) + (Total no. of responses of
category B*2) + (Total no. of responses of
category C*1)
(2)
For example, let us consider the process model
Scrum. For calculating flexibility of Attack
Identification (security practice), we have (data by
experts through survey).
Responses in Category A = 6
Figure 2: Detail in Terms of Percentage of Recorded
Responses.
Responses in Category B = 2
Responses in Category C = 10
Sum of values (for ‘flexibility’ of ‘attack
identification’) marked by experts in Category A=21
Sum of values marked by experts in Category
B=9
Sum of values given by experts in Category
C=36
By applying (2) we get,
=(21 * 3) + (9 * 2) + (36 * 1) ÷ ( (6 * 3) + (2 * 2) +
(10 * 1) )
Flexibility of Attack Identification in Scrum is =
3.6 (Table 2)
(Since we want to find agility of each security
practice)
Number of agile attributes = 4
Total agility of Attack Identification is sum of
flexibility, speed, leanness, responsiveness.
= [3.6+3.5+3.2+3.7] = 14
Number of practices considered = 1
(Since we want to find agility of each security
practice) Number of agile attributes = 4
Table 2: Agility of Security Practices in Scrum.
Agile Attributes
Security Practices
Flexibility
Speed
Leanness
Responsiveness
Total
Agilit
y of a
Secur
ity
practi
ce
Attack Identification 3.6 3.5 3.2 3.7 0.87
Threat Modelling 3.9 3.5 3.2 3.6 0.88
Security Requirement
Analysis
3.8 3.1 3.1 3.3 0.83
Security Education and
Awareness
3.4 2.9 2.8 3.0 0.75
Build Security Team 3.6 3.3 3.4 3.4 0.85
Resource Identification 3.6 3.4 3.5 3.2 0.85
Roles Identification 3.9 3.3 3.7 3.4 0.89
Review Design Security 3.7 3.3 3.0 3.8 0.86
Static Code Analysis 3.8 3.1 3.3 3.5 0.85
Penetration Testing 3.7 3.2 3.4 3.5 0.86
Incident Response
Planning
3.5 3.5 2.7 3.6 0.81
Table 3: Agility of Security Practices in XP.
Agile Attributes
Security Practices
Flexibility
Speed
Leanness
Responsiveness
Total
Agility
of a
Securit
y
practice
Attack Identification 3.8 3.8 3.4 3.7 0.91
Threat Modelling 3.8 3.3 3.0 3.3 0.8
Security Requirement
Analysis
3.6 3.1 2.8 3.6 0.81
Security Education
and Awareness
3.4 3.0 2.6 3.2 0.76
Build Security Team 3.3 2.7 3.0 3.2 0.77
Resource
Identification
3.5 3.2 2.9 3.1 0.79
Roles Identification 3.5 3.1 3.4 3.2 0.82
Review Design
Security
3.7 3.1 2.8 3.6 0.85
Static Code Analysis 3.8 2.9 3 3.4 0.8
Penetration Testing 3.6 3.0 3.0 3.2 0.85
Incident Response
Planning
3.4 3.5 2.7 3.3 0.80
Agility of Security Practices and Agile Process Models: An Evaluation of Cost for Incorporating Security in Agile Process Models
335
By applying (1) we get, [14÷(1 x 4)]
Total agility of Attack Identification =3.5
Since we need to see the effect of this value on
agility of any process model, we must have it
between the ranges 0-1.
Hence, Total agility of Attack Identification
=3.5÷4=0.87
There are different practices and phases in each
model and so the agility of security activities differs
for each model. For example, Build Security Team
Roles has value of 0.85 in Scrum whereas in XP it is
0.77. This is explained by the fact that Scrum has an
inherent process of building teams in terms of
making Scrum teams thus making the activity more
flexible, speedy, lean and responsive whereas, XP
has no such inherent process. However, further study
is required to understand the changing behaviour of
security practices in different process models.
4 APPLYING SELECTED
SECURITY PRACTICES TO
PROCESS MODELS
The main purpose of calculating agility of process
models and security practices is to see how security
practices will affect the agility of process models.
Here we will perform the calculations and analysis.
Few abbreviations will be used
AOM Actual Agility of Model (Calculated
previously as Degree of Agility of Process Model)
(Range 0-1).
ART
Agility Reduction Tolerance (Calculated
Previously as Degree of Agility of Security
practices).
(Range 0-1)
AAAS Agility after Applying Security
practices (Calculated in this section).
(Range 0 – AOM)
Note we call agility of security practice, as
“Agility Reduction Tolerance” because this value
represents the cost a process model will bear for
including a security practice. Hence, ART (Agility
Reduction tolerance) is the factor responsible for
reducing agility of process models.
4.1 Formula
AAS=[((ART activity
1
+ ART
activity
2……
+ART activity
n
) ÷ n) x AOM]
(3)
In this formula, we sum up the agility of selected
security activities and divide it by total number of
selected activities then multiply the obtained value
by agility of selected process model. This would
give us new agility of process model after taking out
the cost of including security practices. Further
explanation and application of this formula is
provided with examples.
It is evident that some process models have
higher agility as compared to others. As given in
Table 1. Further we have applied the approach to
Scrum and XP, keeping in view that their data is
more unswerving with greater number of responses
as compared to other process models.
4.1.1 Scrum
Let us see the effect of including following security
practices in Scrum. Let us take the values for ARTof
these activities from (Since we want to find agility
of each security practice) Number of agile attributes
= 4.
By applying (1) we get, [14÷(1 x 4)]
Total agility of Attack Identification =3.5. Since we
need to see the effect of this value on agility of any
process model, we must have it between the ranges
0-1.
Hence, Total agility of Attack Identification
=3.5÷4=0.87 (
Table 2)
ART of Attacks Identifications =0.87
ART of Review design security =0.86
ART of Static code analysis=0.85
Here,
n=3
AOM (Actual Agility of Model
Table 1) = 0.7
By putting the values in (3) we get,
Agility After Application of Security (AAAS)
= [((0.87+0.86+0.85) ÷3) × 0.7]
AAAS=0.62
Here, the new agility of Scrum is 0.62 whereas. It
is obvious that we will bear cost of including
security in Scrum; this work provides an empirical
way to see that cost. This can serve as one major
factor, (there can be other factors like time (Ayalew,
Kidane, & Carlsson, 2013)) for selection of security
practices.
4.1.2 XP
Let us see the effect of including same security
practices (Attacks Identifications, Review design
security and Static code analysis) in XP. Let us take
the values for ART of these activities from (Table 3)
ART of Attacks Identifications =0.91
ENASE 2020 - 15th International Conference on Evaluation of Novel Approaches to Software Engineering
336
ART of Review design security =0.85
ART of Static code analysis=0.8
Here,
n=3
AOM (Actual Agility of Model
Table 1 ) = 0.68
By putting the values in (3) we get,
Agility After Application of Security AAAS
= [((0.91+0.85+0.8) ÷3) × 0.68]
AAAS=0.58
In this case the AAAS of XP becomes 0.58
which is lower than its original value (0.68). The
effect of including selected security activities in XP
is visible in terms of reduced degree of agility. This
represents the cost that one has to bear in terms of
agility for including security practices.
5 CONCLUSIONS
The effect of including selected security activities
can be seen in both process models. This leads to
two conclusions. Firstly you are firm to use certain
security practices, let us say as your prime factor in
this case you can perform the calculations to see the
effect of your decision on agility of different process
models. Secondly, you are firm to use certain
process model and you are ready to adjust security
practices keeping the degree of agility of process
model as prime factor. Both of above-mentioned
approaches can be handled by proposed method.
6 FUTURE WORK
This research can serve as one major parameter for
selection of security practices. However, further
study is required to learn about different values of
same security activities in different process models.
Second area of further work on this topic is to
investigate other dimensions (for example time and
monetary factors) that can help in selection of
security activities. We will be looking into these
dimensions along with Agility Reduction Tolerance
of security activities in future in perspective of agile
process models.
REFERENCES
Alnatheer, Ahmed, Gravell, Andrew and Argles, & David.
(2010). Agile Secuirty Issues. International
Symposium on Empirical Software Engineeringl and
Measurement. Italy: ACM/IEEE.
Alreck, P.L., Settle, & R.B. (1995). The survey research
handbook:guidelines and strategies for conducting a
survey. IRWIN Professional Publishing.
Ashraf, S., & Aftab, & S. (2017). IScrum: An improved
scrum process model. Ashraf, S., & Aftab, S. (2017).
IScrum: A International Journal of Modern Education
and Computer Science (IJMECS), Ashraf, S., & Aftab,
S. (2017). IScrum: An improved scrum process model.
9(8), 16-24.
Ayalew, T., Kidane, T., & Carlsson, B. (2013).
Identification and Evaluation of Security Activities In
Agile Projects. Springer-Verlag Berlin Heidelberg
(pp. 139-153). Springer.
Beck, & Kent. (2004). Extreme Programming Explained,
Embrace Change. Addison-Wesley.
Boström, Gustav, & et al. (2006). Extending XP practices
to support security requirements engineering.
International workshop on Software engineering for
secure systems. ACM.
Hossein keramati, & Mirian-Hosseinabadi, S.-H. (2008).
Integrating Software development Security Activities
with Agile Methodologies. International Conference
on Computer Systems and Applications. Doha, Qatar:
IEEE.
Howard, M., & L. S. (2006). The Security Development
Lifecycle - SDL: A Process for Developing
Demonstrably More Secure Software. Microsoft Press.
Jacobson. (2002). A resounding 'Yes' to agile processes -
But also more. Cutter IT Journal, 15.
Jon A. Krosnick, & Presser, S. (2010). Handbook of
Survey Research. Emerald Group Publishing Limited.
K., R., S., H., & V., L. (August, 2017). Busting a myth:
Review of agile security engineering methods. In
Proceedings of the 12th International Conference on
Availability, Reliability and Security., 1-10.
K.Beck, M. A. (2001). The Agile Manifesto. Retrieved
from www.agie.alliance.org.
Ken Schwaber, & Beedle, M. (2002). Agile Software
Development with Scrum (Vol. 1). Upper Saddle
River: Prentice Hall.
Kravchenko, Elena, & E. W. (2017). Integrating Security
in Agile projects. Belfast: OWASP.
L. R. (1932). A Technique for the Measurement of
Attitudes. Archives of Psychology, Vol. 22, No. 140,
pp. 1-55.
McGraw, G., Allen, J. H., Barnum, S., & Ellison, R. J.
(2008). Why Is Security a Software Issue?, Software
Security Engineering: A Guide for Project Managers.
The Addison-Wesley Software Security Series.
Moyon, F., Beckers, K., & Kleppe, S. (2018). Towards
Continuous Security Compliance in Agile Software
Development at Scale. International Workshop on
Rapid Continuous Software Engineering. Gothenburg,
Sweden: ACM/IEEE.
Nardi, P. M. (2014). Doing survey research : a guide to
quantitative methods. London: Paradigm Publishers.
Oppenheim, A. N. (2000). Questionnaire design,
interviewing and attitude measurement. Bloomsbury
Publishing.
Agility of Security Practices and Agile Process Models: An Evaluation of Cost for Incorporating Security in Agile Process Models
337
Qumer, A., Henderson-Sellers, & B. (2008). An
Evaluation of the Degree of Agility in Six Agile
Methods and its Applicability for Method
Engineering. Information and Software Technology,
50(4), 280-295.
Rea, L.M., & R.A., P. (1997). Designing and conducting
survey research: a comprehensive guide (2nd ed.).
Jossey-Bass, Inc.
APPENDIX
Matrix Marked by Experts for Each Process Model
Agile
Attributes
Security Practices
Flexibility
Speed
Leanness
Responsiveness
Attacks identification
[1,2,3,4,5] [1,2,3,4,5] [1,2,3,4,5] [1,2,3,4,5]
Threat Modelling
Security requirement
Analysis
Security Education
and Awareness
Build Security Team
Resource
Identification
Roles Identification
Review Design
Security
Static Code Analysis
Penetration Testing
Incident Response
Planning
ENASE 2020 - 15th International Conference on Evaluation of Novel Approaches to Software Engineering
338