An Action Research Study towards the Use of Cloud Computing

Scenarios in Undergraduate Computer Science Courses

Heleno Cardoso da Silva Filho and Glauco de Figueiredo Carneiro

Universidade Salvador (UNIFACS), Salvador-BA, Brazil

Keywords:

Cloud Computing, Active Learning, Action Research.

Abstract:

Cloud computing has been a successful paradigm in its goal to provide remote computing resources in a com-

petitive and scalable way when compared to traditional computing scenarios. Companies have a growing

interest in migrating and using cloud services. However, the literature has reported difﬁculties and challen-

ges faced by companies while migrating their assets to the cloud. One of the possible reasons for this is the

difﬁculty in the identiﬁcation of qualiﬁed professionals to support companies to plan, perform and monitor

the migration of their legacy systems to the cloud. This paper presents an action-research study analyzing

the inclusion of cloud computing scenarios in the System Analysis and Design and Operating Systems under-

graduate courses at Salvador University (UNIFACS). The results of the action-research study provided initial

evidence that cloud computing resources integrated to the contents of the aforementioned courses can contri-

bute to motivate and engage students in activities. In addition, the knowledge and experience gained by these

students can improve their qualiﬁcation to facilitate access to the labor market.

1 INTRODUCTION

Cloud computing paradigm has the goal to provide

services and scalable resources at an accessible cost

with acceptable levels of elasticity and reliability

(Armbrust et al., 2010; Zhang et al., 2010). It is an

evolutionary step towards the effective use of com-

putational resources (Oliveira et al., 2014). Moreo-

ver, it can be a solution for companies to deal with

issues such as cost reduction as well as the possibi-

lity to change the conﬁguration and the allocation of

computational resources, including software and har-

dware, on demand (Armbrust et al., 2010; Oliveira

et al., 2014).

Researchers have identiﬁed key advantages and

challenges faced by practitioners. In terms of advan-

tages, we highlight elasticity (Armbrust et al., 2010),

scalability (Marston et al., 2011), storage capacity

(Bond, 2015), cost reduction (Marston et al., 2011),

and mobility (Fernando et al., 2013). In terms of po-

tential challenges that are bound to concerns faced du-

ring Cloud Computing adoption, we mention the fol-

lowing (de Paula et al., 2017; Sultan, 2010): security

, reliability (Sultan, 2010), privacy and conﬁdentia-

lity (Ryan, 2011), portability (Jones et al., 2017), and

www.cloudsecurityalliance.org

interoperability (Petcu and Vasilakos, 2014).

There is a tendency towards the use of cloud com-

puting in several areas and this is not an exception

for education (Lin et al., 2014; Smith et al., 2014).

The demand for professionals to conﬁgure and ma-

nage cloud computing resources is an opportunity for

new practitioners in cloud computing related activi-

ties. Despite this opportunity, there is still an open

question on how undergraduate students can be pre-

pared to deal with the cloud computing paradigm.

Studies have reported the need to engage Computer

Science undergraduate students in hands-on activities

(Hanna et al., 2015; Vaquero, 2011). The lack of mo-

tivation affects the learning process and therefore can

interfere in the execution of activities. For this rea-

son, it is advisable to include in the courses activities

that resemble real situations related to cloud issues in

an attempt to engage students (Lin et al., 2014). For

example, activities dealing with the identiﬁcation of

which cloud provider to choose, as well as a feasi-

bility analysis related to the migration of the assets

and potential services of an organization to the cloud

can be interesting scenarios to grasp students atten-

tion. The challenges faced by newcomers while exe-

cuting cloud activities may include the identiﬁcation

of which cloud provider to hire and the respective re-

sources to allocate to a new service (Oliveira et al.,

Filho, H. and Carneiro, G.

An Action Research Study towards the Use of Cloud Computing Scenarios in Undergraduate Computer Science Courses.

DOI: 10.5220/0006644800150025

In Proceedings of the 13th International Conference on Evaluation of Novel Approaches to Software Engineering (ENASE 2018), pages 15-25

ISBN: 978-989-758-300-1

2014; Sadiku et al., 2014).

Empirical instruments are an effective option to

analyze the effectiveness of the aforementioned acti-

vities. In fact, studies have reported the use of these

instruments in Computer Science undergraduate cour-

ses where students can experience real-life problems

that can motivate them towards positive learning out-

comes (Smith et al., 2014; Vaquero, 2011; Sultan,

2010).

Considering the scenario described above, we de-

ﬁned the following Research Question (RQ): ”Ana-

lyze the use of cloud computing scenarios for the pur-

pose of understanding its effectiveness with respect to

the adoption of these scenarios in ”Software Analysis

and Design” and ”Operating Systems” courses from

the viewpoint of students in the context of Compu-

ter Science undergraduate courses at Salvador Uni-

versity (UNIFACS)”. We intend to draw conclusions

from the results of this study on how to engage stu-

dents to be prepared and aware of the challenges and

opportunities of cloud computing in the market.

The rest of this paper is organized as follows. Next

section presents the context of this work. Section

3 describes the action research study and in Section

4 we analyze the data collected during the study.

Section 5 presents the conclusion, threats to validity

and scope for future research.

2 CONTEXT OF THIS WORK

Many researchers have argued that the traditional

classroom setting has key shortcomings and alterna-

tives such as the ﬂipped classroom have been used

with interesting results (Bishop and Verleger, 2013;

Williams et al., 2017).

The core idea of the ﬂipped classroom approach

is to ﬂip the lecture-based classroom instruction and

utilize interactive activities and reading assignments

in advance of class (Tucker, 2012). Class time is then

used to engage learners in problem-based, collabora-

tive learning and advancing concepts. After class, stu-

dents can check understanding and extend learning to

more complex tasks. Most importantly, the learner

has control of the pace and time it takes to learn the

material (Green et al., 2017). The Figure 1, adapted

from the Faculty Innovation Center

, illustrates this

scenario.

Flipping a class can be a worthwhile approach, es-

pecially in courses where the material or processes are

traditionally difﬁcult for students to grasp (Williams

et al., 2017). And this can be the case of Computer

https://facultyinnovate.utexas.edu/ﬂipped-classroom

Figure 1: Flipped Classroom (Faculty Innovation Center -

Univ. of Texas).

Science Undergraduate courses. Moreover, student-

centered, technology-based, and active learning ap-

proaches such as ﬂipped classroom relies on students

wanting to take control of their learning. This appro-

ach enables students to set their own goals, monitor

their own progress, facilitating their own and others

critical thinking and problem solving skills (Newman

et al., 2015).

Today with the advent of DevOps and Cloud Com-

puting, solid knowledge in operating systems is re-

quired (Bond, 2015). This reinforces the need for an

active learning approach in these courses. On the ot-

her hand, the quality of the software product depends

directly on the quality of the artifacts produced du-

ring the software development life cycle. For this rea-

son, the course of System Analysis and Design plays

an important role to consolidate the knowledge regar-

ding the construction of artifacts that describe the pro-

blem and register the solution proposed to this pro-

blem. Students usually have reported the challenge

to understand and register the problem accordingly in

artifacts such as use cases or user stories and to derive

this information in other artifacts throughout the soft-

ware life cycle such as class and sequence diagrams

(Bahill and Madni, 2017).

3 THE STUDY

In the action research term, action refers to impro-

ving practice and research refers to creating know-

ledge from the practice experience (McNiff, 2016).

When conducting an action research study, the rese-

archer is immersed in the target situation under in-

vestigation. The work unfolds in response to the si-

tuation and not only to the researchers requirements.

Descriptions and theories are built up as a result of

the iteration within the context in close collaboration

between researchers and participants (Holwell, 2004).

The action research steps are iterative and incremen-

tal (Hendricks, 2012). Figure 2 shows these steps in-

tegrated to the ﬂipped classroom approach.

We used the Goal Question Metric (GQM) appro-

ach to plan the action research study. From the re-

search question (RQ) presented before, we deﬁned

ENASE 2018 - 13th International Conference on Evaluation of Novel Approaches to Software Engineering

the goals of this study. The goal was then reﬁned

into questions that break down the issue into its ma-

jor components. Each question was in the sequence

reﬁned into metrics. The same metric can be used to

answer different questions under the same goal. In

Figure 3 we show diagrammatically the relationship

among goals, questions and metrics of this study. The

action research was conducted during the ﬁrst semes-

ter of the academic year of 2017.

Figure 2: Action Research Iterative Steps. Adapted from

(McNiff, 2016).

3.1 Action-Research Planning Phase

The planning phase has two key moments as descri-

bed below. In the ﬁrst moment, the course syllabus is

the input for the identiﬁcation of the course characte-

ristics. In the sequence, the set of key cloud resources

are indicated as requirements to decide which cloud

computing scenarios should be included in the acti-

vities. In the second moment, we plan the activities

based on the topics/components of the syllabus and

contextualized in the selected scenario. We then con-

ﬁgure the environment for the execution of the activi-

ties in the context of the selected scenarios. An im-

portant issue in this phase is the diagnosis of the cha-

racteristics of each course, especially the components

of the syllabus to promote an effective alignment of

the proposed activities with the course goals. It is im-

portant to mention that, following the characteristics

of an action-research study, the activities created to

each course can be adjusted based on feedback pro-

vided by the students, researcher or teacher. In these

cases, we should return to the Phase 2 of the planning

to promote the activities adjustments.

Figure 4 shows the spiral meta-model of the the

action-research approach we envisioned to be instan-

tiated in each course. This meta-model is compo-

sed of the following phases: planning (green rectang-

les), execution (yellow rectangle), monitoring, ana-

lysis and feedback phases (blue, gray and white rec-

tangles). When conducting the action research study,

we repeatedly passe through these phases in iterations

(called Spirals). This was needed particularly for the

execution of all topics of the course syllabus until the

end.

3.2 The Targeted Courses

We decided to conduct the action research study fo-

cusing on two courses: Operating Systems (OS) and

System Analysis and Design (SAD) for the following

two reasons. The ﬁrst is the potential to explore the

use of cloud resources (Mokhtar et al., 2013) in both

courses. The second is the opportunity to illustrate

how to perform real activities in the context of to-

pics from the syllabus of the two courses. An ope-

rating system is a program that manages a compu-

ters hardware. It also provides a basis for applica-

tion programs and acts as an intermediary between

the computer user and the computer hardware (Sil-

berschatz et al., 2014). The Ubuntu Linux distribu-

tion was adopted as the operating system to perform

the planned activities. This distribution is available in

several cloud providers such as Google Cloud

and

Amazon AWS

. Moreover, this distribution is well-

known and popular in the open source community

The topics related to process management, memory

management, input/output system, and ﬁle manage-

ment are related to operating systems concepts and

also enable the use of virtual machines and containers

in the cloud.

System Analysis and Design (SAD) deals with

planning the development of software systems based

on the understanding and speciﬁcation in detail of

what a system should do and how the components of

the system should be implemented and work together.

The discipline also focuses on identifying characteris-

tics of the software architecture and its components

and addresses the concepts of deploy and orchestra-

tion of applications in the IaaS layer (infrastructure

as a service). Both courses deal with contents that

for some extent can require the use of cloud compu-

ting capabilities. Therefore, the two courses can be

integrated with activities related to cloud computing

activities, such as those required for the migration of

legacy systems to the cloud.

3.3 Selecting Projects from Github

We used the following criteria to select software pro-

jects from Github: Stars, Forks, Contributors, Recent

https://cloud.google.com/compute/docs/images

https://aws.amazon.com/marketplace/pp/B01JBL2I8U

https://distrowatch.com/dwres.php?resource=popularity

An Action Research Study towards the Use of Cloud Computing Scenarios in Undergraduate Computer Science Courses

Figure 3: Action Research Goals, Questions and Metrics.

Figure 4: Action Research Meta-model.

Updates, and Open Source Version. These criteria are

the same as adopted by (Borges et al., 2016). These

projects are used as scenarios for the activities in each

course. The selection also considered that they have

ENASE 2018 - 13th International Conference on Evaluation of Novel Approaches to Software Engineering

an installation script and can be accessed/installed in

Ubuntu. Table 1 lists the selected projects (Web appli-

cations) based on the highest scores in each domain.

3.4 Cloud Scenarios and Activities

We used the virtualization and containers features in

the cloud computing scenarios. The virtualization ai-

med at simulating a remote server that provided cloud

services according to the Infrastructure as a Service

(IaaS) model.

Table 1: Projects Selected from Github.

Domain

Stars Forks Contributors Updates Version

Health- Ope-

nEmr

352 490 85 Yes GNU GPL

Conference

Talk- OpenCFP

313 130 50 Sim MIT

Wine- Nodecel-

lar

14 44 26 Yes APACHE

2.0

Taxi- App 2 8 5 No No

Network Moni-

toring - Zabbix

141 66 1 Yes GPL-2.0

Scenario 1 - Nodecellar Wine Store (virtual ma-

chine and container versions): For the execution of

this cloud computing scenario, we used the Cloudify

Gigaspaces platform to simulate an application run-

ning in the cloud through virtualization (Virtual Box

), according to an Infrastucture as a Service (IaaS)

model. The Nodecellar

is based on Node.js, a Ja-

vaScript runtime, to manage (retrieve, create, update,

delete) the wines in a wine cellar database

. This ap-

plication has been used by the open source commu-

nity to demonstrate the steps required to migrate, de-

ploy and orchestrate a legacy application to the cloud.

Scenario 2: OpenEMR (virtual machine and con-

tainer versions). OpenEMR

is an open source elec-

tronic health record and medical practice manage-

ment solution. It can run on Windows, Linux, Mac

OS X and several other platforms. Its source code has

PHP5+ as programming language, MySQL or Mari-

aDB as database, as well as Apache or other PHP re-

lated webserver.

Scenario 3: OpenCFP (virtual machine and con-

tainer versions). OpenCFP

is a PHP-based con-

ference talk submission system that enables the dis-

cussion of concepts related to the analysis (problem)

and project (solution) in a software system. It also

enables the discussion of process management, me-

mory management and input/output management in

the context of the operating system course.

https://www.virtualbox.org/

http://nodecellar.coenraets.org/

https://github.com/cloudify-cosmo/cloudify-

nodecellar-example

http://www.open-emr.org/

https://github.com/opencfp/opencfp

Scenario 4: Zabbix (virtual machine and contai-

ner versions). It is an open source enterprise-level

software designed for real-time monitoring of metrics

collected from remote servers, virtual machines and

network devices. This scenario was analyzed in the

Amazon Web Services Cloud Service (AWS) through

the use of containers (Docker) and Virtual Machines

(VM).

Scenario 5: Mobile Taxi Application (virtual ma-

chine and container versions). For use of this cloud

computing scenario, the artifacts of the mobile taxi-

app application are available in the GitHub

to dis-

cuss the concepts of reverse engineering and UML di-

agrams.

Scenario 6: Portal Amazon. It is a worldwide on-

line shopping of books, magazines, music, DVDs, vi-

deos and many other items.

As showed in Figure 2, the activities to be per-

formed by the students were posted and available in

the Blackboard portal of the university. It is a virtual

learning environment adopted at Salvador University

(UNIFACS) for both face-to-face and distance lear-

ning courses. The activities can be reached at the url

indicated in this footnote

3.5 Data Sources for the Analysis

We considered data provided by the following sour-

ces in this study: (a) Activities registered in Black-

board; (b) Questionnaire on Student proﬁle; (c) Que-

stionnaire for feedback about the adherence of cloud

computing scenarios; (d) Midterm and Final Assign-

ment; (e) Student Attendance; (f) Feedback from stu-

dents during classes and activities; (g) Research and

teacher perceptions.

The questionnaires used the Likert scale of 5. No

identiﬁcation was required to answer the questionnai-

res. Prior to the questionnaires application, we high-

lighted the importance of the answers to the study and

the beneﬁts that the results and ﬁndings from these

study would bring to better prepare students to the

cloud computing market. A total of 13 students in

the discipline of Operating systems and 11 students

in the discipline of System Analysis and Design com-

pleted the questionnaires. The questionnaire for the

System Analysis and Design (SAD) students had 35

questions (33 closed and 2 open). The questionnaire

for the Operating Systems students had 34 issues (32

closed and 2 open). The closed questions aimed at

obtaining the degree of knowledge in the Computer

Science ﬁeld. The open questions aimed at obtaining

https://github.com/mistryrn/taxi-app

https://cloudeduc.github.io/cloudeduc/

An Action Research Study towards the Use of Cloud Computing Scenarios in Undergraduate Computer Science Courses

Table 2: Goals, Questions and Metrics for SAD Discipline.

Research issues

Partici-

pation

Evalu-

ation

Activities

Difﬁ-

culties

Motivated

Commi-

tted

Qualiﬁ-

cation

Chall-

enges

Results

Partici-

pation

(64.32%)

(64.80%)

(47%)

Low Pro-

ductivity

–

Motivated

(61.73%)

Compro-

mised

(61.73%)

– — (59.92%)

Partici-

pation

(64.32%)

–

(47%)

Low Pro-

ductivity

–

Motivated

(61.73%)

– – – (57.68%)

Partici-

pation

(64.32%)

–

(47%)

Low Pro-

ductivity

– –

Compro-

mised

(61.73%)

– – (57.68%)

Partici-

pation

(64.32%)

– – – – – –

Little

Prepared

(56.61)

(60.47%)

Partici-

pation

(64.32%)

(64.80%)

(47%)

Low Pro-

ductivity

Difﬁ-

culties

(55.54%)

Motivated

(61.73%)

Compro-

mised

(61.73%)

– – (59.19%)

Partici-

pation

(64.32%)

–

(47%)

Low Pro-

ductivity

– – –

Very little

Qualiﬁed

(40.73)

– (50.68%)

the perceptions of the students regarding the engage-

ment and what was learned while performing cloud

computing related activities.

3.6 Proﬁle of the Students

Both classes have 16 students enrolled in the ﬁrst se-

mester of 2017. We used questionnaires in May of

2017 to obtain data related to the proﬁle of the stu-

dents that took part in the action research study. The

questionnaire was composed of 15 closed questions.

The questions focused on issues related to academic

background and knowledge related to cloud compu-

ting. The questionnaire can be reached at the same

url indicated in the previous footnote.

3.7 The Role of the Authors of the Study

Among the two authors of the study, the second one

took the role as the instructor of the two disciplines.

He developed lesson plans, conducted face-to-face

and online learning activities and participated in all of

the action research process. The ﬁrst author was one

of the members of the Monitoring and Support Team

(MST). The purpose of MST was to guide the instruc-

tor, conduct macro level analysis with the researcher,

assess the process, increase the validity and reliabi-

lity of data collection and analysis procedures, deve-

lop functional actions based on ﬁndings of the macro

level analysis, and to help make the research process

as much objective as possible. The MST consisted of

2 experts (the authors) from curriculum, instruction

and educational technologies.

4 ANALYZING THE DATA

In this section, we present the analysis of data obtai-

ned during the action research study to answer the

Research Question (RQ): ”Analyze the use of cloud

computing scenarios for the purpose of understan-

ding its effectiveness with respect to the adoption of

these scenarios in ”Software Analysis and Design”

and ”Operating Systems” courses from the view-

point of students in the context of Computer Science

undergraduate courses at Salvador University (UNI-

FACS)”. With this analysis we intend to draw conclu-

sions on how to engage students to face the challenges

and opportunities of cloud computing in the market.

In Figure 5, we show diagrammatically the relations-

hip among goals, questions and metrics. Based on this

relationship, we explain ﬁrst how the metrics were

calculated, then we show how they were combined to

answer the questions. Finally, we explain the goals.

Within one-semester learning process of the proposed

inclusion of cloud computing scenarios in the two dis-

ciplines, the way students reacted and performed the

activities was monitored, as well as the ways students

interacted with their peers and with the researchers.

Moreover, the ways of interaction emerged during the

activity process were carefully analyzed via content

analysis.

4.1 Analyzing Data for the SAD

Discipline

In the following paragraphs, we analyze data for the

System Analysis and Design (SAD) Discipline. We

also explain how the values were calculated. In Table

3, we present the values for metrics M1-M8.

Answering the Questions for the SAD Discipline.

In the following, we present the calculations for the

ENASE 2018 - 13th International Conference on Evaluation of Novel Approaches to Software Engineering

eight questions aiming at providing conditions to cha-

racterize the goals stated in this study.

4.1.1 Question Q1 - Participation of the

Students in SAD

According to Figures 3 and 5, the metrics M1-M5

were designated to answer Question Q1. In Table

4, we present the results of metrics M1, M2 and M3

to answer Q1. In Table 5, we present the results of

metrics M4 and M5 collected from questions 3.1 to

3.10 of the questionnaire of discipline System Analy-

sis and Design.

Table 3: Collected Metrics for the SAD Discipline.

Metric

Description

Value

(Average)

M1 Average of students attendance (Frequency) 89.66%

M2 Assignments Average Value 64.80%

M3 Average of activities/discussions posted on Blackboard 47.00%

M4 Students ’ motivational level in activities 61.73%

M5 Level of commitment to students in activities 61.73%

M6 Qualiﬁcation level of students to deal with cloud compu-

ting activities

40.73%

M7 Student preparedness level for cloud computing challenges 56.61%

M8 Percentage of difﬁculties students had in the activities with

cloud resources

55.54%

The value obtained for Q1 of 64.32% classiﬁed

students from the System and Analysis Design (SAD)

as Participatives. This is consistent with the feedback

students provided describing their motivation to per-

form the activities contextualized in cloud computing

scenarios.

Table 4: M1, M2 and M3 in SAD Discipline.

Metrics Description

Average

M1 Average of students attendance (Frequency) (Total 1088h

/ 112.50h Faltas)

89.66%

M2 Average Rating 64.80%

M3 Average of activities/discussions posted on Blackboard 47.00%

Table 5: M4 and M5 in SAD Discipline - Question 3.1 a

3.10.

Question

Not

Approa-

ched

Insufﬁcient

Little

Enough

Mote

than

Enough

3.1 0% 0% 9.09% 90.91% 0%

3.2 0% 0% 9.09% 81.82% 9.09%

3.3 0% 9.09% 0% 90.91% 0%

3.4 0% 9.09% 18.18% 72.73% 0%

3.5 0% 9.09% 45.45% 36.36% 9.09%

3.6 0% 9.09% 72.73% 18.18% 0%

3.7 0% 0% 63.64% 27.27% 9.09%

3.8 0% 45.45% 18.18% 27.27% 9.09%

3.9 0% 45.45% 18.18% 36.36% 0%

3.10 0% 18.18% 45.45% 36.36% 0%

Average 0% 14.54% 30.00% 51.82% 3.64%

Table 6: Answering Q1 through M1-M5 for SAD Disci-

pline.

Metric Average (%) Weight

Result (%)

M1 89.66% 3 268.98%

M2 64.80% 2 129.60%

M3 47.00% 3 141.00%

M4 and 5M 51.82% 2 103.64%

Average 64.32%

4.1.2 Question Q2 - Results of the

Assignments/Evaluations - SAD

According to Figures 3 and 5, only the metric M2 was

designated to answer Question Q2. In Table 7, we

present the results of metric M2 to answer Q2 for the

ﬁrst semester of 2017. We considered the medterm

and ﬁnal exam to calculate M2 to answer Q2. The

value of 64.80 is considered regular, near 70 that was

the expected result.

4.1.3 Question Q3 - Activities Posted in

Blackboard - SAD

According to Figures 3 and 5, the metric M3 was de-

signated to answer Question Q3. In Table 8, we pre-

sent the results of metric M3 to answer Q3. As can be

seen in the table, a participation of 47% in the activi-

ties of the course was an evidence that it can be im-

proved. The reason for this occurrence was obtained

by feedback in which students declared that also used

email and instant message services in their mobile for

discussions.

4.1.4 Question Q4 - Difﬁculties Faced by

Students - SAD

According to Figures 3 and 5, the metrics M1, M3 and

M8 were designated to answer Question Q4. Consi-

dering that M1 and M3 for the SAD discipline were

already discussed and listed in Table 4, we present

in Table 9 the result for metric M8 as 36.36% rela-

ted to difﬁculties faced by students while performing

cloud computing activities in the discipline. Conside-

ring together metrics M1, M3 and M8, in Table 10 we

present the result for Question Q4 as 55.54%. This

highlight the challenges faced by students to perform

cloud related activities in the context of the SAD dis-

cipline. However, analyzing evidence from the stu-

dents feedback, we realized that despite the difﬁcul-

ties, students strove to achieved the goals.

4.1.5 Question Q5 - Motivation to Perform

Cloud Activities - SAD

According to Figures 3 and 5, the metrics M1, M3

and M4 were designated to answer Question Q5. In

Table 11, we present the results of metrics M1, M3.

The metric M4 is presented in table 12. Considering

together metrics M1, M3 and M4, in Table 13 we pre-

sent the result for Question Q5 as 61.73%. This is

considered a regular level of motivation for this dis-

cipline with an increasing tendency as veriﬁed in the

feedback provided by the students.

An Action Research Study towards the Use of Cloud Computing Scenarios in Undergraduate Computer Science Courses

Figure 5: Relationship among Goals, Questions and Metrics.

4.1.6 Question Q6 - Commitment to Perform

Cloud Activities - SAD

According to Figures 3 and 5, the metrics M1, M3 and

M5 were designated to answer Question Q6. In Table

11, we present the results of metrics M1 and M3. In

Table 14, we present the results of metrics M5. Fi-

nally, in Table 15, we calculate the value for Q6 as

61.73%. This is an interesting evidence of commit-

ment, reporting that majority of students are engaged

in the cloud activities in the context of the discipline.

4.1.7 Question Q7 - Level of Qualiﬁcation - SAD

According to Figures 3 and 5, the metrics M1, M3 and

M6 were designated to answer Question Q7. In Table

11, we present the results of metrics M1 and M3. In

Table 16, we present the results of the metric M6. Fi-

nally, in Table 17, we calculate the value for Q7 as

40.73%. This value indicates that students have the

perception that they need to improve their qualiﬁca-

tion to deal with challenges and opportunities related

to cloud computing in the market. This is in fact an

evidence that students are aware of improvement op-

portunities they need to strive in order to be a qualiﬁed

professional.

Table 7: Answering Q2 through M2 for SAD Discipline.

k Average Frequency Frequency Relative Percentage (%)

1 4.0 1 0.06 6

2 4.5 1 0.06 6

3 5.5 1 0.06 6

4 5.8 1 0.06 6

5 6.0 2 0.13 12

6 6.1 1 0.06 6

7 6.2 1 0.06 6

8 6.4 1 0.06 6

9 6.5 1 0.06 6

10 6.8 1 0.06 6

11 7.5 2 0.13 12

12 8.0 1 0.06 6

13 8.4 2 0.13 12

Average 64.80

Table 8: Answering Q3 through M3 for SAD Discipline.

Data

Activities Scenarios

Participation

students(%)

13/03/2017 Activity 1 wiki.openmrs.org 100%

20/03/2017 Activity 2 wiki.openmrs.org 50%

21/03/2017 Activity 3 wiki.openmrs.org 25%

10/04/2017 Activity 4 Taxi App Android 13%

Average 47%

Table 9: Difﬁculties faced in Activities (M8 - SAD).

Elements Qty Percentage(%)

a) Understanding the proposed strategy for using

cloud-computing scenarios

4 36.36%

b) Master analysis techniques for Business ru-

les, speciﬁcation functional requirements, rai-

sed/identiﬁed/Elicitados

3 27.27%

c) Master analysis Techniques for speciﬁcation requi-

rements not Func., raised/identiﬁed/Elicitados

1 9.09%

d) Elaborate user story, usage cases, usage case dia-

gram, prototyping

2 18.18%

e) Elaborate class diagram 3 27.27%

f) Use virtualization Techniques (virtual machine,

Container, images)

5 45.45%

g) Understanding Software Architecture 4 36.36%

h) Elaborate sequence diagram, activity diagram 7 63.64%

i) Elaborate Compoenentes diagram, deployment dia-

gram

7 63.64%

j) Understand the steps of direct/reverse engineering 5 45.45%

k) Using Software engineering tools 3 27.27%

Average 36.36%

Table 10: Answering Q4 through M1, M3 and M8 for SAD

Discipline.

Metric Average (%) Weight

Result (%)

M1 89.66% 3 268.98%

M3 47.00% 3 141.00%

M8 36.36% 4 145.44%

Average 55.54%

Table 11: M1 and M3 in SAD Discipline.

Metrics Description

Value

M1 Frequency (Total 1088h / 112.50h Fouls) 89.66%

M3 Average of activities/discussions posted on

Blackboard

47%

4.1.8 Question Q8 - Level of Preparation - SAD

According to Figures 3 and 5, the metrics M1, M3

and M7 were designated to answer Question Q8.

In Table 11, we present the results of metrics M1

and M3. In Table 18, we present the results of the me-

tric M7. Finally, in Table 19, we calculate the value

ENASE 2018 - 13th International Conference on Evaluation of Novel Approaches to Software Engineering

Table 12: M4 in SAD Discipline.

Question

Not Ap-

proached

Insufﬁcient

Little

Enough

than

Enough

3.1 0% 0% 9.09% 90.91% 0%

3.2 0% 0% 9.09% 81.82% 9.09%

3.3 0% 9.09% 0% 90.91% 0%

3.4 0% 9.09% 18.18% 72.73% 0%

3.5 0% 9.09% 45.45% 36.36% 9.09%

3.6 0% 9.09% 72.73% 18.18% 0%

3.7 0% 0% 63.64% 27.27% 9.09%

3.8 0% 45.45% 18.18% 27.27% 9.09%

3.9 0% 45.45% 18.18% 36.36% 0%

3.10 0% 18.18% 45.45% 36.36% 0%

Average 0% 14.54% 30.00% 51.82% 3.64%

Table 13: Answering Q5 through M1, M3 and M4 for SAD

Discipline.

Metric Average (%) Weight

Result (%)

M1 89.66% 3 268.98%

M3 47.00% 3 141.00%

M4 51.82% 4 207.28%

Average 61.73%

Table 14: M5 in SAD Discipline.

Question

Not Dis-

cussed

Insufﬁcient

Little

Enough

More than

Enough

3.1 0% 0% 9.09% 90.91% 0%

3.2 0% 0% 9.09% 81.82% 9.09%

3.3 0% 9.09% 0% 90.91% 0%

3.4 0% 9.09% 18.18% 72.73% 0%

3.5 0% 9.09% 45.45% 36.36% 9.09%

3.6 0% 9.09% 72.73% 18.18% 0%

3.7 0% 0% 63.64% 27.27% 9.09%

3.8 0% 45.45% 18.18% 27.27% 9.09%

3.9 0% 45.45% 18.18% 36.36% 0%

3.10 0% 18.18% 45.45% 36.36% 0%

Average 0% 14.54% 30.00% 51.82% 3.64%

Table 15: Answering Q6 through M1, M3 and M5 for SAD

Discipline.

Metric Average (%) Weight

Result (%)

M1 89.66% 3 268.98%

M3 47.00% 3 141.00%

M5 51.82% 4 207.28%

Average 61.73%

Table 16: M6 in SAD Discipline.

Question

No Acti-

vity

Little

Activity

Performed

Activities

90% 0% 10%

Table 17: Answering Q7 through M1, M3 and M6 for SAD

Discipline.

Metric Average (%) Weight

Result (%)

M1 89.66% 2 179.32%

M3 47.00% 4 188.00%

M6 10.00% 4 40.00%

Average 40.73%

for Q8 as 56.61%. Similarly to Q7, this value indica-

tes that students have the perception that they need to

improve their preparation to deal with challenges and

opportunities related to cloud computing in the mar-

ket. This is in fact an evidence that students are aware

of improvement opportunities they need to strive in

order to perform effectively cloud related activities.

In Table 2, we summarize the results for all Goals,

Questions and Metrics for the SAD discipline.

Table 18: M7 in SAD Discipline.

Question

Not

prepa-

red

Very little

prepared

Little Pre-

pared

Prepared

Very

prepa-

red

5.1 36.36% 27.27% 27.27% 9.09% 0%

5.2 45.45% 18.18% 0% 36.36% 0%

5.3 36.36% 27.27% 27.27% 9.09% 0%

5.4 27.27% 27.27% 27.27% 18.18% 0%

5.5 36.36% 27.27% 9.09% 27.27% 0%

5.6 54.54% 9.09% 36.36% 0% 0%

5.7 36.36% 27.27% 18.18% 18.18% 0%

5.8 18.18% 36.36% 18.18% 27.27% 0%

Average 36.36% 25.00% 20.45% 18.18% 0.00%

Table 19: Answering Q8 through M1, M3 and M7 for SAD

Discipline.

Metric Average (%) Weight

Result (%)

1 89.66% 3 268.98%

3 47.00% 4 188.00%

7 36.36% 3 109.08%

Average 56.61%

4.2 Analyzing Data for the OS

Discipline

Considering that the Operating System discipline has

the same goals, questions and metrics as the SAD dis-

cipline and that it follows the sama analysis as presen-

ted in the previous subsection, we present Table 22

where we summarize the results for all Goals, Ques-

tions and Metrics for this discipline.

4.3 Analysis of the Results of APS and

OS Disciplines

In Table 21, we compare the results for metrics M1-

M8 for the disciplines SAD and OS. In Tables 2 and

22, we present a panoramic view for the goals, questi-

ons and metrics for the disciplines SAD and OS. Ana-

lyzing these values, we can conclude that both disci-

plines have similarities based on their Goals, Questi-

ons and Metrics. Hence, despite being different dis-

ciplines, they reacted uniformly to the proposed ap-

proach that includes cloud computing scenarios in the

activities of the disciplines.

4.4 The Engagement of the Students

As can be seen in the results presented in the ﬁelds

Q5 and Q6 (Motivated/Committed: 61.73% SAD/

61.91% OS) of Tables 2 and 22, students from both

disciplines manifested engagement and also commit-

ment to perform the activities contextualized in cloud

computing scenarios. This is an initial evidence that

in fact classes in Computer Science undergraduate

courses can be enriched with these scenarios when ap-

plied using active learning techniques such as ﬂipped

classroom to provide students the opportunity to be

the main participants in the learning process. On the

other hand, teachers can identify improvement oppor-

tunities in each student and guide them to ﬁll gaps in

An Action Research Study towards the Use of Cloud Computing Scenarios in Undergraduate Computer Science Courses

topics they are not so conﬁdent. This is an iterative

and incremental process that can be conducted with

real scenarios from the market.

Table 20: Collected Metrics for the OS Discipline.

Metric

Description

Value

(Average

M1 Average of students attendance (Frequency) 90.49%

M2 Assignments Average Value 62.50%

M3 Average of activities/discussions posted on Blackbo-

ard

46.35%

M4 Students ’ motivational level in activities 61.91%

M5 Level of commitment to students in the activities 61.91%

M6 Qualiﬁcation level of students to deal with cloud com-

puting activities

52.64%

M7 Student preparedness level for cloud computing chal-

lenges

56.36%

M8 Percentage of difﬁculties students had in the activities

with cloud resources

52.34%

Table 21: Comparing the Metrics of SAD and OS Discipli-

nes.

Metric

APS (%) SO (%)

M1 - Average of students attendance (frequency) 89.66% 90.49%

M1 - Average of Students Attendance 64.32% 63.98%

M2 - Average Students Grade 64.80% 62.50%

M3 - Average of the activities/discussions posted

at Blackboard

47.00% 46.35%

M4 - Level of Student Motivation in activities 61.73% 61.91%

M5 - Level of commitment of students in the

activities

61.73% 61.91%

M6 - Level of qualiﬁcation of pupils to act with

cloud computing (virtualization activities)

40.73% 52.64%

M7 - Level of preparation of students for the

challenges of cloud computing

56.61% 56.36%

M8 - Percentage of difﬁculties faced by students

in cloud computing activities

55.54% 52.34%

5 CONCLUSIONS

Software is a fundamental component of many sys-

tems, services, and products. Its development con-

sumes increasing amounts of resources. Moreover,

an infrastructure is needed to run and make availa-

ble software to users. The cloud computing paradigm

has increasingly provided a better cost-beneﬁt relati-

onship for both the industry and ﬁnal users. To deal

with this scenario, students should be prepared to per-

form tasks in this scenario. In this paper, we analyzed

the inclusion of cloud computing scenarios in two un-

dergraduate courses. As future work, there is the pos-

sibility of conducting a new version of this research

in collaboration with industry (Hanna et al., 2015), as

well as applying the described instruments of ﬂipped

classroom and action research approach in industry

for training purposes (Fagerholm et al., 2017).Anot-

her possibility is the conduction of this study in ot-

her universities to compare results and to obtain better

conditions to generalize results and ﬁndings.

ACKNOWLEDGMENTS

The ﬁrst author of this paper received a scholarship

from the Bahia Research Foundation (FAPESB) re-

gistered as BOL0731/2016.

REFERENCES

Armbrust, M., Fox, A., Grifﬁth, R., Joseph, A. D., Katz,

R., Konwinski, A., Lee, G., Patterson, D., Rabkin, A.,

Stoica, I., et al. (2010). A view of cloud computing.

Communications of the ACM, 53(4):50–58.

Bahill, A. T. and Madni, A. M. (2017). Discovering System

Requirements, pages 373–457. Springer International

Publishing, Cham.

Bishop, J. L. and Verleger, M. A. (2013). The ﬂipped class-

room: A survey of the research. In ASEE National

Conference Proceedings, Atlanta, GA, volume 30, pa-

ges 1–18.

Bond, J. (2015). The Enterprise Cloud: Best Practices for

Transforming Legacy IT. O’Reilly Media.

Borges, H., Hora, A., and Valente, M. T. (2016). Pre-

dicting the popularity of github repositories. In Pro-

ceedings of the The 12th International Conference

on Predictive Models and Data Analytics in Software

Engineering, PROMISE 2016, pages 9:1–9:10, New

York, NY, USA. ACM.

de Paula, A. C. M., de Figueiredo Carneiro, G., and Maciel,

R. S. P. (2017). A characterization of cloud compu-

ting adoption based on literature evidence. In ICEIS

2017 - Proceedings of the 19th International Confe-

rence on Enterprise Information Systems, Volume 1,

Porto, Portugal, April 26-29, 2017, pages 53–63.

Fagerholm, F., Kuhrmann, M., and M

unch, J. (2017). Gui-

delines for using empirical studies in software engi-

neering education. PeerJ Computer Science, 3:e131.

Fernando, N., Loke, S. W., and Rahayu, W. (2013). Mo-

bile cloud computing: A survey. Future generation

computer systems, 29(1):84–106.

Green, L. S., Banas, J. R., and Perkins, R. A. (2017). The

ﬂipped college classroom.

Hanna, P., Allen, A., Kane, R., Anderson, N., and McGo-

wan, A. (2015). Building professionalism and em-

ployability skills. Computer Science Education,

25(3):292–310.

Hendricks, C. C. (2012). Improving schools through action

research: A reﬂective practice approach. Pearson

Higher Ed.

Holwell, S. (2004). Themes, iteration, and recoverability

in action research. In Information Systems Research,

pages 353–362. Springer.

Jones, S., Irani, Z., Sivarajah, U., and Love, P. E. (2017).

Risks and rewards of cloud computing in the uk pu-

blic sector: A reﬂection on three organisational case

studies. Information Systems Frontiers, pages 1–24.

Lin, Y.-T., Wen, M.-L., Jou, M., and Wu, D.-W. (2014). A

cloud-based learning environment for developing stu-

ENASE 2018 - 13th International Conference on Evaluation of Novel Approaches to Software Engineering

Table 22: Goals, Questions and Metrics for OS Discipline.

Research issues

Partici-

pation

Evalu-

ation

Activities

Difﬁ-

culties

Motivated

Commi-

tted

Qualiﬁ-

cation

Chall-

enges

Results

Partici-

pation

(63.98%)

(62.50%)

(46.35%)

Low Produc-

tivity

–

Motivated

(61.91%)

Compro-

mised

(61.91%)

– — (59.33%)

Partici-

pation

(63.98%)

–

(46.35%)

Low Produc-

tivity

–

Motivated

(61.91%)

– – – (57.41%)

Partici-

pation

(63.98%)

–

(46.35%)

Low Produc-

tivity

– –

Compro-

mised

(61.91%)

– – (57.41%)

Partici-

pation

(63.98%)

– – – – – –

Little

Prepared

(56.36%)

(60.17%)

Partici-

pation

(63.98%)

(62.50%)

(46.35%)

Low Produc-

tivity

Difﬁ-

culties

(52.34%)

Motivated

(61.91%)

Compro-

mised

(61.91%)

– – (58.17%)

Partici-

pation

(63.98%)

–

(46.35%)

Low Produc-

tivity

– – –

Very little

Qualiﬁed

(52.64)

– (54.32%)

dent reﬂection abilities. Computers in Human Beha-

vior, 32:244–252.

Marston, S., Li, Z., Bandyopadhyay, S., Zhang, J., and

Ghalsasi, A. (2011). Cloud computingthe business

perspective. Decision support systems, 51(1):176–

189.

McNiff, J. (2016). You and your action research project.

Routledge.

Mokhtar, S. A., Ali, S. H. S., Al-Sharaﬁ, A., and Aborujilah,

A. (2013). Cloud computing in academic institutions.

In Proceedings of the 7th international Conference on

Ubiquitous information Management and Communi-

cation, page 2. ACM.

Newman, D. L., Deyoe, M. M., Connor, K. A., and Lamen-

dola, J. M. (2015). Flipping stem learning: Impact on

students process of learning. Curriculum Design and

Classroom Management, page 23.

Oliveira, T., Thomas, M., and Espadanal, M. (2014). Asses-

sing the determinants of cloud computing adoption.

Information & Management, 51(5):497–510.

Petcu, D. and Vasilakos, A. V. (2014). Portability in clouds:

approaches and research opportunities. Scalable Com-

puting: Practice and Experience, 15(3):251–270.

Ryan, M. D. (2011). Cloud computing privacy concerns on

our doorstep. Communications of the ACM, 54(1):36–

38.

Sadiku, M. N., Musa, S. M., and Momoh, O. D. (2014).

Cloud computing: opportunities and challenges. IEEE

potentials, 33(1):34–36.

Silberschatz, A., Galvin, P. B., and Gagne, G. (2014). Ope-

rating system concepts essentials. John Wiley & Sons,

Inc.

Smith, A., Bhogal, J., and Sharma, M. (2014). Cloud

computing: adoption considerations for business and

education. In Future Internet of Things and Cloud

(FiCloud), 2014 International Conference on, pages

302–307. IEEE.

Sultan, N. (2010). Cloud computing for education: A new

dawn? International Journal of Information Manage-

ment, 30(2):109–116.

Tucker, B. (2012). The ﬂipped classroom. Education next,

12(1).

Vaquero, L. M. (2011). Educloud: Paas versus iaas cloud

usage for an advanced computer science course. IEEE

Transactions on Education, 54(4):590–598.

Williams, A., Sun, Z., Xie, K., Garcia, E., and Ashby, I.

(2017). Flipping STEM, pages 149–186. Springer In-

ternational Publishing, Cham.

Zhang, Q., Cheng, L., and Boutaba, R. (2010). Cloud com-

puting: state-of-the-art and research challenges. Jour-

nal of internet services and applications, 1(1):7–18.

An Action Research Study towards the Use of Cloud Computing Scenarios in Undergraduate Computer Science Courses