A model for DevOps implementation in South Africa.

Abstract

Software is a vital component of a digital economy. Organisations use software for operational efficiency, to gain competitive advantage and to retain customers. Software systems have become the focal point for competitive advantage and organisations have embraced software development methodologies that have the capacity to handle changing user requirements in a quick and nondisruptive manner. The advent of agile methodologies ensured that there was a universally endorsed software development methodology that could deliver on the promise of speedy development while maintaining sufficient quality standards to ensure end-user satisfaction. However, the euphoria that surrounded the perception of success with regard to agile methodology was soon dispelled because of challenges related to the phases of software release and deployment into the functional (“live/operational”) environment. Hence, the intended benefits of agile methodology became somewhat limited thereby compromising the methodology’s acceptance on a universal platform. To overcome this impediment, the DevOps approach was introduced to obviate the disconnect between the development and deployment teams so that software features can be rapidly deployed. The DevOps approach did, however, come with its own set of challenges. Software practitioners were challenged to acquire an understanding of the factors that achieve acceptance and enable DevOps success. Although these factors have been extensively researched in the context of agile methodology, there is a lack of empirical evidence that attests to the acceptance and success factors for DevOps implementation. The current study was undertaken to provide a resolution to this gap that has been identified in the body of knowledge pertaining to software development methodology. The study leverages comprehensive empirical evidence from the professional software development community in South Africa to enable the researcher to converge the study’s analysis and discussion to a cogent model of DevOps acceptance and success factors. From a research design perspective, a sequential exploratory research method consisting of qualitative and quantitative phases of research was employed in this study. The qualitative method used a phenomenological approach to examine DevOps professionals' lived experience of DevOps to validate success and acceptance factors that have been identified in the academic literature. The outputs from the qualitative phase of the study were integrated with existing literature to create a conceptual model that was then subjected to quantitative inquiry using the technique of structural equation modelling. The results from the quantitative inquiry identified the factors of automation, Devops tools and technology and organisational culture as being critical to DevOps success. From a methodology acceptance perspective, performance expectancy, effort expectancy and social influence were identified as crucial determinants of the behavioural intention to use DevOps. The outcome of this study is the generation of a concise list of factors that are structured along the dimensions of DevOps success and DevOps acceptance. These factors have been empirically validated from the academic and practitioner perspectives thereby providing practitioners with an ideal resource to understand DevOps usage and enhance DevOps adoption in the professional sector. It also provides sufficient academic content to enable researchers to use the knowledge generated in the study as a platform from which to sustain the evolution of agile methodology and DevOps knowledge.