This article showed that, although dependency models provide a lot of advantages to service management, they are not widely used. The main reason for this is the lack of generally available service models (respectively sufficiently exact dependency descriptions) and the huge effort needed to generate such models manually.
A method was presented that enables the creation of such models for various use cases in a -- to a considerable extent -- automated way.
There are still unsolved problems and question that have to be investigated further, like to what number of services and applications the method still works well enough. Or what polling intervals are suitable for the data collection. For this, a compromise between reducing time and bandwidth' for the management purposes and the quality of the dependency detection must be found.
To prove and to improve the robustness of the method, especially because it is based on sometimes mistrusted neural networks, the plans for the future also include tests in several real world environments and for various management tasks.