Aspects of Modelling Requirements in Very-Large Agile Systems Engineering

TL;DR

This paper explores how requirements models are successfully used in an agile systems engineering context at Ericsson, highlighting practical approaches to integrate modeling with agile workflows.

Contribution

It provides empirical insights into effective requirements modeling practices in agile systems engineering, addressing a gap in existing research.

Findings

Models help understand requirement relationships and track product evolution

Text-based notations improve model maintenance and integration into workflows

Trade-offs include reduced control over diagram layout

Abstract

Using models for requirements engineering (RE) is uncommon in systems engineering, despite the widespread use of model-based engineering in general. One reason for this lack of use is that formal models do not match well the trend to move towards agile developing methods. While there exists work that investigates challenges in the adoption of requirements modeling and agile methods in systems engineering, there is a lack of work studying successful approaches of using requirements modelling in agile systems engineering. To address this gap, we conducted a case study investigating the application of requirements models at Ericsson AB, a Swedish telecommunications company. We studied a department using requirements models to bridge agile development and plan-driven development aspects. We find that models are used to understand how requirements relate to each other, and to keep track with the product's evolution. To cope with the effort to maintain models over time, study participants suggest to rely on text-based notations that bring the models closer to developers and allow integration into existing software development workflows. This results in tool trade-offs, e.g., losing the possibility to control diagram layout.