Diversifying Software Architecture for Sustainability: A Value-based Perspective
This work addresses sustainability in software architecture for architects, but it is incremental as it extends existing cost-benefit analysis methods.
The paper tackles the problem of evaluating software architectural diversification for sustainability by proposing a method based on real options pricing theory to assess long-term value under uncertainty, illustrated through a GridStix case study.
Although the concept of software diversity has been thoroughly adopted by software architects for many years, yet the advent of using diversity to achieve sustainability is overlooked. We argue that option thinking is an effective decision making tool to evaluate the trade-offs between architectural strategies and their long-term values under uncertainty. Our method extends cost-benefit analysis method CBAM. Unlike CBAM, our focus is on valuing the options which diversification can embed in the architecture and their corresponding value using real options pricing theory. The intuitive assumption is that the value of these options can provide the architect with insights on the long-term performance of these decisions in relation to some scenarios of interest and use them as the basis for reasoning about sustainability. The method aims to answer the following: (1) Is diversification of architectural decisions beneficial and when they can help in sustaining the software, (2) When, where and to what extent. The proposed model is illustrated and evaluated using a case study from the literature referred to as GridStix.