Content area
More and more, modern software systems are modular and distributed, built by different teams that work independently and run in complicated environments. In this situation, keeping track of dependencies between components is very important, especially when it comes to keeping things stable, lowering technical debt, and allowing for quick changes. This dissertation looks into how graph databases, especially Neo4j, can help with managing dependencies in a real-world setting on a large scale, like the Vodafone TV (VTV) platform made by Celfocus.
The study's main goal is to use a graph-based model to keep track of software dependencies between three separate teams working on the VTV project. After looking at the problems with current methods, a solution was suggested that used a centralized Neo4j instance for high-level inter-team dependencies and team-specific instances for more detailed internal modeling. A version control system based on Git was created to keep graph data safe and make it easier to make updates all the time, in line with agile workflows.
We used a mix of questionnaires and informal interviews to test the system, focusing on the team that was the first to use Neo4j. The results showed that the tool made things easier to see, encouraged proactive communication, and helped with impact analysis during development and deployment. But it was still hard to make sure that new users were trained and that modeling was done at the right level of detail.
This dissertation comes to the conclusion that using graph-based dependency management in the daily lives of agile teams can greatly improve the stability of software and the ability of teams to work together. It also shows how important it is for organizations to work together and keep records to keep long-term adoption going.