Modeling and measuring attributes influencing DevOps implementation in an enterprise using structural equation modeling

Abstract

ContextDevOps refer to set of principles that advocate a tight integration between development and operation to achieve higher quality with faster turnaround. It is paramount to assess and measure the DevOps attributes in an enterprise. The literature provides references to these attributes but the detail assessment of these attributes and determination of the maturity of DevOps implementation is still a challenge. ObjectiveThis paper provides important insights for practitioners to assess and measure the DevOps attributes using statistical analysis and Two-way assessment. The proposed framework facilitates the detailed assessment of eighteen attributes to identify key independent attributes and measure them to determine the maturity of DevOps implementation in an enterprise. MethodThe relationship between eighteen attributes was examined; a structural model was established using Exploratory and Confirmatory Factor Analysis, the model was validated using Structural Equation Modelling. Key independent attributes were identified which influences other attributes and overall DevOps implementation. Using Two-way assessment, key independent attributes were measured and the maturity of the DevOps implementation was determined in an enterprise. ResultsUsing Exploratory and Confirmatory Factor Analysis, 18 attributes were categorized under 4 latent variables namely Automation, Source Control, Cohesive Teams and Continuous Delivery. Using Structural Equation Modelling, 10 key independent attributes were determined, that influenced other attributes and overall DevOps implementation. Two-way assessment was applied to measure the key independent attributes and it was found that 4 of these attributes were performing below threshold level. Corrective actions were taken by the management team, and the revised measurement of these attributes demonstrated 40% improvement in the maturity level of DevOps implementation. ConclusionThe proposed framework contributes significantly to the field of DevOps by enabling practitioners to conduct the detailed assessment and measurement of DevOps attributes to determine the maturity of DevOps implementation to achieve higher quality.