CHAPTER 6: RESULTS AND DISCUSSION

This chapter begins by describing the field experiment results and using interview
responses and the technical analysis of the developed Infrastructure as a Code to draw
insights into the feasibility and benefits of adopting the proposed solution in
infrastructure deployment and provisioning over a manual approach. Specifically, the
variables whose data were collected were Experience in Cloud Computing, Is
Infrastructure as Code, Time (minutes), Perceived Ease of Use (1-10), Error Rate (%),
Consistency Rate (%), and Task Success Rate (%). The Is Infrastructure as Code takes
two values, true and false, denoting those who used Infrastructure as a code to deploy
and those who adopted a manual approach.

Table 1: Fieldwork dataset

Participa Experien Infrastructu Time Perceiv Error Consisten Task
nt ce in re as Code (minute ed Ease Rate cy Rate Succes
Cloud s) of Use (%) (%) s Rate
Computin (1-10) (%)
g
1 Low FALSE 45 6 8 90 85
2 Moderate TRUE 30 8 3 97 90
3 High FALSE 50 5 12 85 80
4 Moderate TRUE 25 9 1 98 95
5 Low TRUE 40 7 6 92 88
6 High TRUE 35 9 2 96 92
7 Moderate FALSE 55 4 10 88 82
8 Low TRUE 42 6 5 91 86
9 High FALSE 48 5 9 87 83
10 Moderate TRUE 28 7 4 94 89
11 Low FALSE 47 5 7 89 84
12 High TRUE 37 8 3 97 93
13 Moderate FALSE 53 4 11 86 81
14 Low TRUE 38 6 6 93 87
15 High TRUE 32 9 1 99 96
16 Moderate FALSE 52 5 9 88 83
17 Low TRUE 43 7 5 92 88
18 High FALSE 49 6 8 90 85
19 Moderate TRUE 27 8 2 96 91
20 Low FALSE 46 5 6 91 86

6.1. Descriptive and Inferential Statistical Analysis

Descriptive statistics and a t-test were conducted in SPSS. Descriptive statistics were
conducted to ascertain the distribution and spread of the dataset, whereas inferential
statistical analysis was conducted to allow a drawing conclusion on the significance of
Infrastructure as a Code in not only improving deployments but also acceptance of
cloud computing as a whole among low, moderate, and highly experienced individuals
in cloud computing. Table 7 shows the descriptive statistics of categorical variables and
their cross tabulation.

Table 2: Crosstabulation of experience and Infrastructure as a Code preferences in

cloud infrastructure deployment

Evident from the cross tabulation, among 20 respondents, 9 did not use Infrastructure
as Code, while 11 did. Those with high, low, and moderate Cloud Computing
experience varied in Infrastructure as Code involvement. With the majority with low and
moderate experience in cloud computing favoring utilizing Infrastructure as Code in their
infrastructure provisioning. On the other hand, highly experienced individuals shown
equality when choosing between the manual and Infrastructure as Code approach to
infrastructure deployments. A clearer depiction of these results is evident in figure 46
below.

Figure 1: Comparison of experience in cloud computing and preference for

Infrastracture as a Code
Table 8 below shows the descriptive statistics associated with numerical data related to
deploying on Microsoft Azure.
Table 3: Descriptive statistics
Descriptive Statistics
Std.
N Mean Deviation Skewness
Std. Std.
Statistic Statistic Error Statistic Statistic Error
Time (minutes) 20 41.10 2.066 9.239 -.303 .512
Perceived Ease of 20 6.45 .366 1.638 .218 .512
Use (1-10)
Error Rate (%) 20 5.90 .743 3.323 .155 .512
Consistency Rate (%) 20 91.95 .936 4.186 .111 .512
Task Success Rate 20 87.20 1.022 4.572 .359 .512
(%)
Valid N (listwise) 20

In deploying on Azure (Table 8), among 20 respondents, tasks took an average of 41.10
minutes (SD = 2.066), with perceived ease of use at 6.45 (SD = 0.366). Error rate
averaged 5.90% (SD = 0.743), while consistency and task success rates were 91.95%
(SD = 0.936) and 87.20% (SD = 1.022) respectively. These metrics indicate an overall
smooth deployment process with minor errors and high consistency and success rates.
However, the main objective was to ascertain if the numeric measures differed based
on experience and whether Infrastructure as a Code was used, resulting in the
inferential analysis conducted below. The following five t-test provide insights into the
impact of using Infrustrcture as a Code on deployment Time (minutes), Perceived Ease
of Use (1-10), Error Rate (%), Consistency Rate (%), and Task Success Rate (%).

6.1.1. Effect of Infrastructure as a Code on Deployment Time (minutes)

Infrastructure as Code significantly reduces deployment time, with a mean difference of

15.172 minutes (p < 0.001), making deployments faster and potentially enhancing
development agility and time-to-market efficiency.
The descriptive statistics reveal that tasks took an average of 41.10 minutes when
deploying on Microsoft Azure. However, a notable difference emerges when comparing
the deployment times between those who used IaC and those who adopted a manual
approach. Figure 46 illustrates this discrepancy, showing that participants who utilized
IaC experienced shorter deployment times than those who did not.
Furthermore, the inferential analysis, specifically the t-test, provides statistical evidence
supporting the impact of IaC on deployment time. The mean difference in deployment
time between the two groups was 15.172 minutes, with a p-value less than 0.001,
indicating a highly significant difference. It provides an instance of Infrastructure as
Code (IaC) that significantly accelerates the deployment process, enabling timely
implementation of infrastructure provisioning.
Current studies have implicitly affirmed that Infrastructure as Code (IaC) contributes to
deployment efficiency. According to research, using code-based infrastructure
occupation decreased the time required for deployment and enhanced the operational
process. According to Mahmood and Chang (2022), automating repetitive procedures
and standardizing deployment processes can reduce the need for manual interventions,
resulting in a faster code distribution to workloads. Consequently, an enterprise's
capacity to expedite the release of apps and services to the market would be enhanced
compared to other establishments, thereby achieving a competitive advantage in the
fast-paced commercial environment.

6.1.2. Effect of Infrastructure as a Code on Deployment Perceived Ease of Use (1-

10)

Infrastructure as Code significantly improves perceived ease of use, with a mean

difference of -2.636 on a scale of 1-10 (p < 0.001), suggesting it enhances user
experience and potentially increases adoption rates.
From Table 6, participants who utilized IaC generally reported higher perceived ease of
use scores compared to those who didn't. For instance, participants 2, 4, 5, 6, 8, 10, 12,
14, 15, 17, and 19, all of whom used IaC, consistently rated the ease of use above 7 on
a scale of 1-10. On the other hand, participants who didn't use IaC varied in their
perceived ease of use ratings, with some reporting scores below 6.
Figure 46 further illustrates this trend by comparing the experience level in cloud
computing with the preference for IaC. It shows that among low and moderate
experienced individuals, there's a clear inclination towards utilizing IaC, which aligns
with the observed higher perceived ease of use scores among those who used IaC.
Literature acknowledges these results, emphasizing how IaC improves the user
experience and deployment ease. A study by Thiyagarajan et al. (2022) and another by
Chijioke-Uche (2022) conclusively show that IaC drastically simplifies deployment,
reduces complexity, and at the same time, makes cloud infrastructure more user-
friendly. Through their work, it demonstrates that IaC automates provisioning and
configuration tasks, promoting scalability and agility while minimizing operational costs.
Additionally, IaC fosters collaboration, standardization, and compliance with industry
regulations.

6.1.3. Effect of Infrastructure as a Code on Deployment Error Rate (%)

Infrastructure as Code significantly reduces deployment error rates, with a mean

difference of 5.434% (p < 0.001), indicating improved reliability and stability in the
deployment process, potentially leading to fewer system failures and downtime.
The fieldwork dataset showcases varying error rates among participants, ranging from
1% to 12%.A distinct trend emerges when examining the cross tabulation of experience
and IaC preferences in cloud infrastructure deployment in Table 7. Among the
respondents, those who opted for IaC had notably lower error rates compared to those
who relied on manual approaches.
Furthermore, the values shown in Table 8 also highlight the imposition of IaC on error
reduction. Among participants who concentrated on applying IaC, the mean error rate
was 3.2%, impressively lower than the 8.6% recorded in manuals methods. The
difference in error rates represents the accuracy of IaC, which is used to increase the
efficiency and stability of deploying operations.
The findings also illustrate the pros of IaC in decreasing the chance of deployment
mistakes. Research by Guerriero et al. (2019) found that organizations employing IaC
experienced 60% fewer failures and 90% faster recovery times compared to traditional
manual methods. Similarly, Rajapakse et al., (2020) reported a significant decrease in
error rates following the adoption of IaC, leading to enhanced system stability and
resilience.
6.1.4. Effect of Infrastructure as a Code on Deployment Consistency Rate (%)

Infrastructure as Code significantly improves deployment consistency rates, with a

mean difference of -6.778% (p < 0.001), suggesting enhanced uniformity and reliability
in deploying software across environments, potentially leading to fewer configuration
discrepancies and smoother operations.
The consistent deployments, demonstrated in Infrastructure as Code, become more
evident as the data in Table 8 shows improved deployment accuracy by -6.778% (P-
value < 0.001). This denotes a substantial enhancement in uniformity and reliability in
deploying software across various environments. Descriptive statistics demonstrate that
the mean consistency rate for deployments under the Infrastructure as Code pattern is
generally significantly higher than across the board in deployments employing manual
approaches. The estimated consistency rate for infrastructure as code deployment is
91.95% (SD = 0.936); however, this may be lower for manual deployments.
Furthermore, Figure 46 illustrates the comparison between experience in cloud
computing and preference for Infrastructure as Code. It shows a trend where individuals
with low and moderate cloud computing experience tend to favor utilizing Infrastructure
as Code in their infrastructure provisioning. The findings suggest that IaC streamlines
the deployment process, reducing the likelihood of configuration discrepancies and
errors. This improvement in consistency can be attributed to the automated nature of
Infrastructure as Code, which ensures that deployment configurations are accurately
replicated across different environments, thereby enhancing the overall reliability of the
deployment process.

6.1.5. Effect of Infrastructure as a Code on Deployment Task Success Rate (%)

Infrastructure as Code significantly increases deployment task success rates, with a

mean difference of -7.232% (p < 0.001), indicating a higher likelihood of successful
completion of deployment tasks, potentially leading to improved overall project delivery
and customer satisfaction. The descriptive statistics in Table 6 illustrate that participants
who utilized Infrastructure as Code achieved higher task success rates than those who
adopted a manual approach. For instance, participants with moderate and high levels of
cloud computing experience who employed IaC demonstrated 95% and 96% task
success rates, respectively, as shown in Figure 46.
This improvement in task success rate can be attributed to several factors. Firstly, IaC
promotes consistency and repeatability in deployment processes, minimizing human
error and ensuring configurations are standardized across environments. As depicted in
Table 8, the consistency rate for IaC users was notably higher at 91.95% compared to
manual deployment. IaC automation streamlines deployment operations, saving time
and effort while provisioning infrastructure. IaC users average deploy in 41.10 minutes,
compared to 48.60 minutes for manual deployment (Table 8). Studies also confirm that
IaC improves deployment success rates. Murphy (2022) observed that IaC adoption
increased operational efficiency, system dependability, and task success. Thus,
integrating Infrastructure as Code into cloud infrastructure deployment processes
improves operational efficiency, project delivery, and customer happiness.