Global Continuous Delivery in a Financial Organization

After moving to Agile methodologies the next step in innovating the ING IT landscape has been the transformation to Continuous Delivery and DevOps. At beginning this transformation was local to each different team. Every team could define and implement their own continuous delivery pipeline, integrating the ops activities with the dev activities. However, this approach caused a lot of repetitive work among the different teams to implement and maintain their own pipeline. Moreover, in the case of global improvements needed in any step of the continuous delivery process (e.g., Quality Assurance improvements) all the teams should apply different changes to integrate the same improvement into their different pipelines.


To avoid these problems we developed a global continuous delivery pipeline that is available as service to all ING DevOps teams. Having one unique continuous delivery pipeline has different benefits. First, repetitive development activities have been removed giving this responsibility to a dedicated team. Second, having a global continuous delivery pipeline enabled the seamlessly integration of improvements to all teams using it. This is particularly relevant to assure always high quality and compliance of ING software systems. Finally, having all applications going through a unique pipeline makes the pipeline the single point of truth. The continuous delivery pipeline is not only a technical support to ease the release of software systems. It is the ideal means to apply the scientific approach and, hence, methodologies like Six Sigma and Lean Analytics. Currently, the pipeline is becoming a framework to apply the DMAIC (Define, Measure, Analyze, Improve, and Control) data-driven improvement cycle. Since all the software systems goes through the same pipeline, managers and devops are enabled to solve problems following a data-driven approach. Once a problem is defined (e.g., improve productivity, reduce software problems) metrics can be easily measured mining all the components of the pipeline (e.g., versioning control systems, quality assurance systems). The values of these metrics provide relevant insights to, first, act on a problem and, then, to measure the impact of the proposed solutions. As a consequence, the pipeline is becoming a framework to speed up the resolution of organizational and technical problems and not only the release of software.


However, the globalization of the pipeline can hide critical risks for the agility of the entire organization. First, it’s challenging to have a unique perfect pipeline for the needs of hundreds of different teams causing frustration in teams. Second, having a centralized team dedicated to the pipeline isolate its engineers from the rest of the IT landscape. The main risk is that the CD development teams take away the teams’ responsibility of mastering the tools and the entire delivery process bringing back the organization to a pre-devops era causing the silo-oriented organizational problems. To eliminate these risks the Global Continuous Pipeline at ING is becoming a framework that everyone can extend and contribute to. In this scenario the central continuous delivery team is the integration team that integrate all contributions coming from all teams while assuring the concept-of-one and


Dr. Daniele Romano, Product Owner Continuous Delivery as a Service, ING

DD

Dr. Daniele Romano

Product Owner Continuous Delivery as a Service, ING