DevOps Ontology - An ontology to support the understanding of DevOps in the academy and the software industry

Abstract

Currently, the degree of knowledge about what DevOps really means and what it entails is still limited. This can result in an informal and even incorrect implementation in many cases. Although several proposals related to DevOps adoption can be found, confusion is not uncommon and terminology conflict between the proposals is still evident. This article proposes DevOps Ontology, a semi-formal ontology that proposes a generic, consistent, and clear language to enable the dissemination of information related to implementing DevOps in software development. The ontology presented in this article facilitates the understanding of DevOps by identifying the relationships between software process elements and the agile principles/values that may be related to them. The DevOps Ontology has been defined considering the following aspects: the REFSENO formalism that uses the representation in UML was used and the language OWL language using Prótegé and HermiT Reasoner to evaluate the consistency of its structure. Likewise, it was satisfactorily evaluated in three application cases: a theoretical validation; instantiation of the continuous integration and deployment practices proposed by the company GitLab. Furthermore, a mobile app was created to retrieve information from the DevOps Ontology using the SPARQL protocol and RDF language. The app also evaluated the Ontology’s proficiency in responding to knowledge-based questions using SPARQL. The results showed that DevOps Ontology is consistent, complete, and concise, i.e.: to say: the consistency could be observed in the ability to be able to infer knowledge from the ontology, ensuring that the ontology is complete by checking for any incompleteness and verifying that all necessary definitions and inferences are well-established. Additionally, the ontology was assessed for conciseness to ensure that it doesn't contain redundant or unnecessary definitions. Furthermore, it has the potential for improvement by incorporating new concepts and relationships as needed. The newly suggested ontology creates a set of terms that provide a systematic and structured approach to organizing the existing knowledge in the field. This helps to minimize the confusion, inconsistency, and heterogeneity of the terminologies and concepts in the area of interest.