Energy Consumption in Microservices Architectures: A Systematic Literature Review

Cloud computing emerges as a paradigm that facilitates on-demand access to technological resources through the mechanism of service virtualization. This virtualization enables the partitioning of hardware resources among applications that are organized into distinct independent modules. The concept of microservice architecture takes advantage of virtualization capabilities to embrace a software architecture strategy focused on the development of applications as assemblies of several interdependent but loosely coupled modules. Nonetheless, the adoption of microservices architecture is accompanied by substantial energy demands to meet the desired standards of performance and availability. Existing research within the domain of microservices has explored various topics pertinent to energy consumption, including elasticity, reliability, performance, and availability. Yet, the diversity of challenges and solutions presents a complex landscape for identifying prevailing research trends and unaddressed gaps in the context of microservices. This study aims to methodically discern, evaluate, and juxtapose the existing research trends and voids concerning energy consumption within microservices. It elucidates a systematic review on the subject of energy consumption in microservices architectures, offering a compilation of references to facilitate more directed future investigations. The initial selection encompassed 3625 articles, which were subsequently narrowed down through three stages of refinement, resulting in 37 articles chosen for an exhaustive review. These selected studies were cataloged and analyzed based on various criteria, including metrics, evaluation methodologies, and architectural typologies, thus uncovering research gaps and emerging trends related to energy consumption in microservice architectures. Furthermore, this inquiry delineates significant research challenges and prospective directions, structured around the key metrics that underpin the reviewed studies: performance, elasticity, scalability, reliability, sustainability, and availability.