Architecting for sustainability of and in the cloud: A systematic literature review

Abstract

Context:The interest in the intersection between cloud computing and sustainability is naturally growing as the popularity of the former makes it in many cases the default model for delivering software functionalities to end users. Furthermore, software architecture offers a fundamental route to address sustainability, with the recent shift towards recognizing sustainability as a software quality. Objective:Approaching the intersection between sustainability and cloud computing from the perspective of the study of software architectures, in this work we aim to collect the data necessary for us to understand the relation between these two areas as reflected in the literature. Given the lack of suitable surveys for this purpose in this paper we report on our review of the relevant literature, designed to address the question of how architectural solutions specifically for the cloud are addressing sustainability. Methods:Following the steps prescribed by a well-known method for this purpose on running a systematic literature review, we answer our defined research questions using both (qualitative) analysis and synthesis of data. Results:As a result of the review process, we are able to identify 10 solution types, 8 recurring design decisions, 11 involved architectural entities, and 17 distinct perceptions of the cloud with respect to sustainability. The adoption of reusable architectural tactics and patterns has been observed to be less prevalent than anticipated. Furthermore, certain fundamental characteristics of cloud computing, including multi-tenancy, on-demand resource provisioning, and the pay-as-you-go model, have been identified across diverse cloud perceptions. Conclusion:Our findings point to the need for more systematic work required on developing architectural solutions specifically for cloud computing incorporating sustainability goals. We also suggest that achieving sustainability through cloud in software architecture may be feasible. Furthermore, we identify a persistent threat to further secondary studies on the same topic due to improper use of terms.