Agile Development and Operation of Complex Systems in Multi-technology and Multi-company Environments: Following a DevOps Approach

Abstract

Big innovation and research projects usually require merging contributions form organizations with expertise in different domains. Managing and participating in multi-company projects that use multiple state-of-the-art technologies constitute a challenging activity due to many factors such as integration inexperience, evolving components, tentative requirements, independent teams or independent management centers. In the late nineties and in the early years of 2000, several methodologies arose with the focus on fast releases of working software, commonly known as agile, that aimed to address many of the challenges that this kind of projects face. However, in most cases, these methodologies were not fully adoptable as the automation investment was too high and it was not recoverable during the duration of the project. The global servitization trend and the appearance of approaches, such as DevOps, to support the continuous and fast adjustment of those services to stay in business has also impacted innovation and research projects. On one hand, matured technologies that reduce the automation investment have arisen. On the other hand, whenever it makes sense, services which benefit from the application of DevOps approaches are required to be implemented. This paper explains the implementation of DevOps approaches to support the agile development in the context of innovation and research projects. It also describes two practical implementation cases where such approaches were implemented and how they evolved in the course of the time.

A systematic mapping study about DevOps

“…According to the analyzed studies, it was also observed that DevOps represents: an approach [35,52,55], a cultural movement [17,57], a set of practices [23,27], an emerging paradigm [38,56], a mixture of two words [39], a phenomenon [40], an interdisciplinary theme [22], a set of principles [41], a conceptual framework [42], a development methodology [51], an artificial word [44], a philosophy [47] and a neologism [36]. From the above, it can be evidenced that DevOps has been interpreted in several ways because it is a relatively new subject.…”

“…Table 6 shows only the number of tools proposed by each author because the technological tools that support DevOps are addressed in depth in the research question Q6. 12 Github [21,27,47,50], Bitbucket [21,22,27], GitLab [21,56], Mercurial [22], Apache Subversion [21,56], SonarQube [50,56] where the creation of services to support practices related to the adoption of DevOps over local networks is proposed. Only in [46] describes artifacts and specific practices for the adoption of DevOps in mobile development environments, the other proposals are focused on Web environments 3.6.…”

DevOps benefits: A systematic literature review

“…70 Likewise, the environment process configuration and tools used by each team should be standardized, avoiding common situations like "it was working on my machine." 71 With the mixing of IT teams by making developers and operators work together, the competences of these resources will be increased. 25 Developers will be able to get abilities that most often regard to operations, while operators will get abilities on areas of development.…”

“…This allows developers to have development and preproduction environments, which aids in the discovery of issues early in the SDLC 70 . Likewise, the environment process configuration and tools used by each team should be standardized, avoiding common situations like “it was working on my machine.” 71 …”

DevOps model in practice: Applying a novel reference model to support and encourage the adoption of DevOps in a software development company as case study

“…Some of them (56%) [21], [22], [40]- [43], [24]- [26], [31]- [35] highlight the benefits and challenges. Likewise, 60% [23], [25], [37], [40], [43], [45], [46], [26], [27], [29], [30], [32]- [34], [36] of the studies suggest some process elements that could be considered for the adoption of DevOps. The remaining 40% [24], [27]- [33], [35], [46] mention some technological tools that could support DevOps practices.…”