10.1 Introduction

Development and operations (DevOps) is one of the hottest and most favorable practices in the software delivery process in recent years. Software organizations are under pressure these days because they want an immediate supply of products to market [1]. DevOps aims to increase software production with speed and quality by means of new infrastructure, procedures and potency of automation [2]. DevOps is a set of practices and cultural values that aim to improve collaboration and resolve obstacles between development personnel and IT operation personnel [3-5]. DevOps is a shifting culture whose main objective is to integrate the development and operation to gain business benefits by means of appropriate technology [6]. The word “DevOps” describes the objective to break down the wall between the development team and operation team and bring together both teams by collaborative culture [7]. Time and quality are the two main constraints infecting software development [2]. At a DevOps organization, an independent team takes care of both development as well as IT operations department [8]. The key objectives of DevOps are to mitigate time, boost automation, resolve silos and user feedback between development and IT operation during the software delivery process without affecting the software quality [9, 10]. It aims to integrate teams with new practices and improve communication and automation techniques between the development team and operations teams [11].

Since DevOps is a new trendy and innovative approach in software engineering, its popularity is growing day by day. Most software organizations achieve their business goals by means of this new practice. Today’s most popular software companies, like Google, Facebook, Amazon, Netflix and LinkedIn, adopt DevOps practices that enable fast software production, reduce time and reply quickly to user requests [12]. DevOps greatest emphasis is on productivity and its business-oriented ability to pave the way for organizational culture change and rapid learning [13]. Adoption of DevOps has many advantages, it allows greater collaboration and communication, and reduces the time it takes to market and release new updates of software [14]. However, adoption of this new approach is not an easy task. Several organizations have faced difficulties while changing from legacy infrastructure to new DevOps infrastructure. Transitioning from traditional organization to DevOps organization is a very costly, time-consuming, tedious process requiring great effort [4]. The DevOps practices need a solid and continuous link between the development department and IT operation in order to create a cross-functional team where every member is expected to do the task of the other member [15].

Various literature has been studied and published about DevOps practices with the main objective of finding an ultimate definition of DevOps [7]. Recent and previous scientific literature reveals that the critical challenge is that there is no standard and formal definition for it [2, 16]. A systematic literature review shows that DevOps rapidly spread from 2014 to 2015, and the related papers are low ranking in quality [10]. A systematic literature review (SLR) and empirical study (ES) is required to adopt DevOps practices to improve product quality, automation of operation services and fast delivery of software at low cost.

Since DevOps is a union of development and operation, and development members and operation members work together, there are always conflicts between them which affect the DevOps environment in an organization. Many critical challenges are faced by the development team and operation team. One of the most critical issues is “culture.” DevOps culture focuses on the collaboration and integration of development and operation teams, which enhance the maturity of the software process, and quality and knowledge sharing among teams [17]. One main reason which can cause the entire process to fail is the temporary employee in an organization who is not part of the long-established DevOps culture [18]. Another reason is the geographical distribution, which creates issues like communication, time zone, and social relationship, which are core elements for organizational culture [19].

10.2 Background

Software organizations are shifting from traditional approaches to new practices to deliver fast and smooth applications in the market. Conventional software processes are linear and performed sequentially. Each stage is completed after the completion of the previous stage. Traditional or waterfall model software is monolithic and executes in an isolated environment [20]. Traditional approaches become more challenging and create instability in the software development process [21]. One main disadvantage of traditional methodologies is that they delay the release of new software updates, causing users to become disheartened when errors are found in the updated version of software or it doesn’t work as expected [16].

Conversely, the Agile model is an iterative and incremental software development process. The Agile software development model first appeared in 2001, and created a better relationship between software programers and business organizations [22]. The agile approach was developed by software companies due to the inefficiency of traditional methods to fulfill the desired goals of both software developers and business management. The agile concept has a set of guidelines for the software development process and focuses on the delivery of software in increments. Scrum is an agile application which has been broadly used in software development processes [23].

The term “DevOps” was invented by Petrick Debois in 2009, and he is also known as the father of DevOps [24]. From previous studies it is revealed that issues, such as lack of collaboration, and conflicts between development and operation, adversely impact on the quality, time and delivery of software production [25]. DevOps is a recent practice which breaks down the wall between the development team and operation team. It aims to integrate all team members into a collaborative organization.

DevOps is an evolution of agile approaches in the software development [2,7,26]. It does not have approaches like Scrum and extreme programming (XP), therefore organizations can apply DevOps practices by training their personnel [12]. DevOps is a continuous loop of integrated processes. Every process has its own specific tool which automates deployment and operations processes. The fast growth of DevOps ensures remarkable benefits of automation and acceleration of delivery and deployment of applications in the software industry [27]. Industrial DevOps is a vision to enable sustainable adaptation and improvement in industrial manufacturing by making industrial data available to multiple stakeholders [28]. The DevOps environment needs skillful technical staff as well as a wide variety of tools to maintain a continuous integration pipeline [9].

Despite the numerous advantages of DevOps, every organization cannot adopt this approach. It is not an easy job to adopt DevOps practices. DevOps is a culture shift and it requires various practices, tools, personnel and technology changes and upgradation. According to the literature, researchers have found many challenges while adopting DevOps. One of the most important and critical challenges faced by many organization is “Culture.” The core elements of DevOps culture are face-to-face communication, motivation, trust, responsibility, and respect [29].

• Hüttermann et al. [19] describe DevOps as a multifaceted word with four major characteristics: culture, automation, measurement and sharing.
• Rembetsy et al. [9] stated that Etsy, an e-commerce company, converted his company culture into a DevOps culture.
• Luz et al. [5] stated that the key category for DevOps adoption is a collaborative culture which resolves the conflicts between the development team and IT operation team and carries out all tasks from the first day.
• Luz et al. [3] built a model based on grounded theory, which highlighted that for DevOps adoption a collaborative culture is a central element for practitioners.

From an organizational point of view, the top priorities are company culture and people and how they use DevOps practices in an organization [30]. Those in the DevOps culture need to have the correct mindset to learn new techniques, tools, and additional new tasks; and every member in the organization needs to closely work with colleagues to communicate and understand the new architecture design. Respect and trust are core elements in DevOps culture. Blaming each other can cause the failure of the entire project.

From the literature it is revealed that many researchers have tried to overcome the challenge of the DevOps culture. Researchers have created theories, theoretical models, maturity models, competency models and other techniques to resolve the culture issue in DevOps. In spite of that, it is a new approach in which some studies are carried out using the SLR protocol on the DevOps culture. However, many studies are limited and do not address the culture issues.

10.3 Systematic Literature Review Protocol

This study deals with designing the systematic literature review protocol for DevOps’ culture challenges model (DC2M) by using the guideline of Kitchenham [31]. Systematic literature reviews are different from general literature surveys; the results are more accurate, authoritative and less biased than general literature surveys [32,33]. SLR is a procedure of collecting knowledge about a specific topic and its main objective is to be as unbiased, repeatable and auditable [34]. From the studies it has been demonstrated that SLR assists in organizing and focusing results related to a specific research question in software engineering [35,36]. The SLR is based on the three main phases of systematic review, i.e., planning the review, conducting the review and reporting the review [35]. We first performed the SLR procedure to identify key challenges for DC2M and develop a protocol. We have defined our search strings, related digital libraries, the inclusion and exclusion criteria, and data extraction process [37].

Research Questions:

• RQ1: What are the culture challenges that should be avoided by vendor organizations in DevOps development process?
• RQ2: What are the practices discussed in the literature to overcome culture challenges in DevOps process?
• RQ3: What are the real-world practices that should be adopted to overcome culture challenges?

10.4 Creating the Search String

After constructing the research questions, the following terminologies assist in the design for creating search terms. Table 10.1 show the search terms construction.

Experimental Design: Empirical studies, systematic literature review, theoretical studies, experts’ opinions, and AHP techniques are used in experimental design.

Table 10.2, 10.3, and Table 10.4 specify the search terms construction for RQ1, RQ2, and RQ3 respectively.

10.5 Search Strategies

10.6 Final Search String Construction

In order to find the most relevant literature, we chose the following five scientific databases:

• Google Scholar¹
• ACM Digital Library²
• ScienceDirect³
• SpringerLink⁴
• IEEE Xplore⁵

We constructed a single search string from all three research questions. Within databases the search string was not filtered and also the date range was not selected for the purpose of getting a wide overview. We used the following final search string from the three research questions.

((DevOps OR “continuous integration” OR “software automation” OR “cross-function collaboration” OR “continuous deployment”) AND (culture OR values OR literature) AND (challenges OR issues OR barriers) AND (vendor OR supplier OR trader) AND (practices OR methods OR implementation))

10.7 Selection Criteria and Search Process

In order to find the relevant data to the research questions, the primary selection first takes place on the basis of research paper title, research paper abstract and research paper keywords. The following Table 10.6 shows the primary selection of papers by search results.

The search results gave us the primary selection and then we used the following inclusion and exclusion criteria for selecting papers for the final publications.

10.8 Assessment of Publication Quality

Quality criteria are used to ensure the quality of the final publication and the assessment of quality is performed in parallel at the time of data extraction. The quality of publication is carried out purely on the basis of the following questions:

• Has the author noticeably recognized the challenge which affects the DevOps culture in software organizations?
• What practices are adopted by the author to overcome culture challenges in DevOps?

The abovementioned questions will be marked as ‘YES’ or ‘NO’ or ‘NA’.

The supervisor (secondary reviewer) will also contribute in scoring of small subset for the validation.

10.9 Data Extraction Stage

10.9.2 Presentation of Data Extraction

The extracted data will be presented in the following format:

10.10 Data Synthesis

In the data synthesis stage, one summary table is constructed having columns as well as S No., DevOps Culture Challenges, Frequency, Percentage, etc., which spotlights the list of all challenges in DevOps culture with their frequencies and percentages.

10.11 Discussion

We have conducted this study to explore the various challenges in the DevOps development process faced by vendor organizations. We have used the SLR method to identify key challenges from the literature as given in Table 10.8.

Table 10.8 lists a total of 10 critical challenges identified by the SLR. The challenge “lack of collaboration and communication” (68%) has the highest frequency. The distance of cooperation between communication, development and work will lead to problems such as team conflict and failure [38]. When the conversation is in English, which is not the mother tongue of the seller or the client, it creates natural challenges for communication [39]. Lwakatare et al. describe the problem of poor communication between teams in an organization, which can be identified by the limited common knowledge between people [40]. If there is a lack of communication and collaboration between IT managers and vendor organizations, then it creates misunderstandings and confusion, which results in a lot of work and time [41].

“Lack of skill and knowledge” (56%) is the second major challenge on this list. Implementing successful DevOps techniques in an organization requires knowledgeable and skilled staff in both development and operations departments [38]. Many organizations do not have enough skilled staff to create the right environment for IT projects and this is a key factor in project failure [42].

“Criticism practices,” which has a reported 50% frequency in the literature, is a major challenge that has a negative impact on DevOps culture. If there is a culture of blaming and pointing fingers and people do not change their behavior, we will never learn anything. The literature shows that people resist changing their attitudes, which is the biggest obstacle to the success of organizational change [43].

In our study, the “lack of DevOps approach” (47%) has been identified as a key challenge in the implementation of DevOps culture. The biggest challenge in adopting the DevOps approach is learning new methods and tools as it requires changing mindsets and attitudes in specific contexts [44]. Due to its multifaceted nature, it is difficult to adopt the DevOps approach, because its vague definitions and goals confuse organizations and individuals, leading to a negative perception of resistance to the adoption of DevOps [45].

In every organization, management plays a vital role in achieving goals and objectives. “Lack of management” (45%) is also a critical challenge in our list. The management and DevOps teams do not have a common goal, which results in resource allocation as well as delays in delivery dates, which has a negative impact on stakeholders [46]. If there is friction between the management level as well as the members of other levels, then an undesirable atmosphere is created [47].

“Trust and confidence problems” (45%) also affects the DevOps culture and is a challenge for adoption of DevOps practices. Achieving goals is impossible if there is a lack of trust among team members. Lack of trust is easily established when the development team and the operations team cannot trust each other [48]. People generally resist and fear the loss of changing culture and jobs because they have insufficient expertise in both development and operation areas [45].

“Complicated infrastructure” (35%) is a serious problem that affects the overall performance of the organization. DevOps is a movement from a legacy structure to a highly integrated system, making it more difficult for IT services to become modern, reliable and efficient [44]. It is very difficult to be agile when a complex organization and culture is unintentionally slow [49].

Another challenge on the list is “poor quality” (33%). Poor quality is a major obstacle to project success due to the unavailability of a proper user company. When the development and operations team pays too much attention to non-functional requirements, they release substandard products into the market [39]. The high turnover of team members in an organization also affects the quality and volume of work [47].

Security is the backbone of any organization’s success. In our literature, we have found that “security issues” are a major problem in the adoption of DevOps and our list of key challenges includes 29% frequency in security matters. When an organization does not have security experts, the development team focuses more on production than on “big fixes,” which leads to products with poor quality as well as error-prone areas [50]. Another major mistake of the DevOps organization is that it does not pay attention to the security aspect and the employee has access to the database and can possibly install malware in the system and steal confidential information [46].

“Legacy infrastructure” is the last major challenge on our list and is identified by its 23% frequency in the literature. Because DevOps is a new practice that transforms the traditional structure into a new complex and integrated infrastructure, legacy infrastructure does not support new technologies, new software versions, and new ways of working with tools [51-55]. The traditional system creates resistance to the current way of working to adapt the organization’s role, structure, decision-making and DevOps [40].

10.12 Validation of Review Protocol

This protocol will be submitted to the secondary reviewer (Supervisor, Dr. Abdul Wahid Khan) for the purpose of reviewing the protocol.

10.13 Limitation

A limitation to this study is that only real-world practices from RQ3 (questionnaire survey, interviews, etc.) will be generated. Another limitation to this study is that since DevOps is a new concept, the literature on DevOps culture is limited. Moreover, Google Scholar shows only 184 results out of 6580.