Software Architecture Decision-Making Practices and Challenges: An Industrial Case Study

TL;DR

This paper investigates industrial software architecture decision-making practices through a case study, highlighting challenges faced by practitioners and proposing knowledge management as a key improvement area.

Contribution

It provides empirical insights into current practices, challenges, and potential improvements in software architecture decision-making in industry.

Findings

Practitioners follow diverse decision-making practices.

Challenges include knowledge management and method inconsistencies.

Improving knowledge management can enhance decision quality.

Abstract

Software architecture decision-making is critical to the success of a software system as software architecture sets the structure of the system, determines its qualities, and has far-reaching consequences throughout the system life cycle. The complex nature of the software development context and the importance of the problem has led the research community to develop several techniques, tools, and processes to assist software architects in making better decisions. Despite these effort, the adoption of such systematic approaches appears to be quite limited in practice. In addition, the practitioners are also facing new challenges as different software development methods suggest different approaches for architecture design. In this paper, we study the current software architecture decision-making practices in the industry using a case study conducted among professional software architects in three different companies in Europe. As a result, we identified different software architecture decision-making practices followed by the software teams as well as their reasons for following them, the challenges associated with them, and the possible improvements from the software architects' point of view. Based on that, we recognized that improving software architecture knowledge management can address most of the identified challenges and would result in better software architecture decision-making.