Applying system descriptors to address ambiguity on deployment diagrams

TL;DR

This paper proposes using system descriptors to eliminate ambiguity in deployment diagrams, enhancing communication and consistency throughout system evolution in DevOps environments.

Contribution

It introduces a novel approach linking system descriptors and deployment diagrams to ensure unambiguous, consistent representations, supported by a case study on Netflix deployment diagrams.

Findings

Generated diagrams are graphically equivalent to descriptors.

The approach reduces ambiguity in deployment diagrams.

Preliminary results support the hypotheses.

Abstract

Communication between practitioners is essential for product quality in the DevOps context. This communication often takes place through deployment diagrams of a system under development. However, it is common diagrams to become ambiguous or inconsistent as the system progresses and goes to a continuous delivery pipeline or production. Moreover, diagrams could not follow the evolution of systems, and it is challenging to associate diagrams to production. In this paper, we propose the use of system descriptors to address the ambiguity of deployment diagrams. We state three main hypotheses (1) if a deployment diagram is generated from a valid system descriptor then the diagram is unambiguous; (2) if a valid system descriptor is generated from a deployment diagram then the descriptor is unambiguous; (3) if a diagram $\mu$ generated from a descriptor $A$ is unambiguous and if a descriptor $B$ is generated from the diagram $\mu$ equally unambiguous then descriptors $A$ and $B$ are equivalent. We report a case study to test our hypotheses. We constructed a system descriptor from Netflix deployment diagram, and we applied our tool to generate a new deployment diagram. Finally, we compare the original and generated diagrams to evaluate our proposal. Our case study shows the generated deployment diagrams are graphically equivalent to system descriptors and eliminated ambiguous aspects of the original diagram. Thus, our preliminary results lead to further evaluation in controlled and empirical experiments to test our hypotheses conclusively.