I'm absolutely thrilled to finally share the highlights of my Ph.D thesis journey: Achieving Successful DevOps Adoption in IT Organizations. It has been a fascinating deep dive into DevOps’ adoption, untamed ocean.
For years, I've been passionate about this DevOps problem. From arguing about siloed teams to geeking out at open-source meetups, I've seen firsthand the virtues – and the difficulties – of trying to bring Dev and Ops together. And let's be honest, while everyone and their cousins seem to be talking about DevOps benefits like swifter releases and more satisfied customers, the reality is that many organizations are still scratching their heads, seeing inconsistent success. That’s the elephant in the room, isn’t it? Despite all the hype, there's been a real lack of solid evidence about what realistically drives successful DevOps adoption.
So, I set out on a quest – a proper academic adventure spanning six articles – to figure out how we can make DevOps adoption less of a gamble and more of a predictable outcome. My main aim was pretty straightforward: to investigate achieving successful DevOps adoption in IT organizations.
I had to examine a massive volume of already published material using my "detective" skills to identify the connections. In order to obtain a clear perspective from both scholarly papers and the frequently overlooked "gray literature," such as blogs and white papers, which required two multivocal literature reviews. This assisted me in determining the fundamental concepts of a three-vector problem-solving strategy:
- DevOps Capabilities: The fundamental skills and practices that result in successful DevOps. Consider the "how-to" of DevOps, which includes cross-team collaboration, monitoring, automation, releasing, and more. 37 key capabilities were found and grouped into four categories by my research: cultural, measurement, process, and technical.
- DevOps metrics: Are the most important indicators for checking if DevOps capabilities are going in the right direction. They let you know if those features are truly improving the way you deliver software. Our research revealed 24 main DevOps metrics. Divided into four categories: business, change, operating, and cultural.
- Life Cycle Processes: To really understand how DevOps fits into the nitty-gritty of software development, I mapped the 30 DevOps life cycle processes from the IEEE 2675-2021 standard with a systematic literature review. This provided a structured way to see where and how DevOps capabilities can have the most impact. LCPs are grouped into Technical, Technical Management, Organizational Project-Enabling, and Agreement Processes.
Now for the innovate part! In addition, I carried out a Design Science research that generated a Capability Evaluation Matrix. This is a very useful artifact that assists businesses determine where to concentrate their IT investment and which capabilities, when properly measured, can yield long-term gains. It all comes down to taking informed steps instead of impulsively adhering to the newest trends or buzzes.
But I didn’t stop there. To really help organizations eliminate those frustrating adoption challenges, I reviewed case studies, analyzed documentation, metrics, conducted many interviews with experts and finally developed a Framework for Improving DevOps Adoption Success. Picture it as a systematic roadmap. It uses capabilities, metrics, and life cycle processes as key building blocks to address specific hurdles, like complicated integrations and resisting to change. The framework is iterative, entailing it allows a continuous feedback cycle and dedicated assessment to react to an organization’s unique journey.
What were some important takeaways? Well, cross-team collaboration consistently popped up as absolutely vital. You can't have DevOps without breaking down those silos! Aside from measuring, and having the right metrics in place is about offering the insights you need to drive continuous success. And, do not undervalue the importance of transformational leadership and a generative organization, since they are necessary to create an organizational environment inside which DevOps might flourish and realize what is expected.
I anticipate that this research's practical implications will be significant for every one of you in the trenches. The framework helps to identify and overcome particular obstacles. It shows you how to use capabilities and metrics effectively. Ultimately, it's about moving towards a mindset of collaboration, common ownership, and efficiency.
This lengthy journey has been about bringing a bit more rigor, evidence and consensus to the often-chaotic DevOps adoption world. It's about providing a framework and solidified practical strategies from research to help IT organizations not just do DevOps, but to achieve truly successful and sustainable DevOps adoption.
So, there you have it – a Hitchhiker's Guide to my +200 page thesis. I truly believe the findings offer a significant contribution to understand both the problem from the academic angle and the practical application of DevOps.
This stands as the foundation of an upcoming DevOps course focused on applying DevOps with use of Open-Source tooling, considering Kubernetes, and various Cloud Native software, starting in November at Técnico+. It aims to provide skills to implement DevOps effectively in real-world scenarios. Emphasizing on understanding DevOps capabilities, tools, metrics, processes, and the culture of collaboration, along with its preparation for CNCF certification, and strategizing adoption in practice.
Sources:
Amaro, R., Pereira, R. and Mira da Silva, M. (2022) ‘Capabilities and Practices in DevOps: A Multivocal Literature Review’, IEEE Transactions on Software Engineering, 1, p. 20. Available at: https://doi.org/10.1109/TSE.2022.3166626.
Amaro, R., Pereira, R. and da Silva, M.M. (2024) ‘DevOps Metrics and KPIs: A Multivocal Literature Review’, ACM Computing Surveys [Preprint]. Available at: https://doi.org/10.1145/3652508.
Amaro, R., Pereira, R. and da Silva, M.M. (2023) ‘Capabilities and Metrics in DevOps: A Design Science Study’, Information & Management, p. 32. Available at: https://doi.org/10.1016/j.im.2023.103809.
Amaro, R., Pereira, R. and da Silva, M.M. (2024) ‘Mapping DevOps capabilities to the software life cycle: A systematic literature review’, Information and Software Technology, p. 107583. Available at: https://doi.org/10.1016/j.infsof.2024.107583.
Faustino, J. et al. (2022) ‘DevOps benefits: A systematic literature review’, Software: Practice and Experience, 52(9), pp. 1905–1926. Available at: https://doi.org/10.1002/spe.3096.