A Persona-Based Approach for Exploring Architecturally Significant Requirements in Agile Projects

[Context and motivation]

Architecturally significant requirements (ASRs) drive and constrain many aspects of the architecture. It is therefore beneficial to elicit and analyze these requirements in early phases of a project so that they can be taken into consideration during the architectural design of the system. Unfortunately failure to invest upfront effort in exploring stakeholders quality concerns, can lead to the need for significant refactoring efforts to accommodate emerging requirements. This problem is particularly evident in agile projects which are inherently incremental.

[Question/Problem]

Existing techniques for early discovery of ASRs, such as Win-Win and i*, are typically rejected by agile development teams as being somewhat heavy-weight. A light-weight approach is therefore needed to help developers identify and explore critical architectural concerns early in the project.

[Principal ideas/results]

In this paper we present the use of Architecturally-Savvy Personas (ASP-Lite). The personas are used to emerge and analyze stakeholders’ quality concerns and to drive and validate the architectural design. ASP-Lite emerged from our experiences working with the requirements and architectural design of the TraceLab project. The approach proved effective for discovering, analyzing, and managing architecturally significant requirements, and then for designing a high-level architectural solution which was designed to satisfy requirements despite significant interdependencies and tradeoffs.

[Contributions]

This paper presents the ASP-Lite approach and describes its support for architectural design in the US$2 Million TraceLab project.