An Empirical Study of Energy Efficiency Monitoring for Container-Based Data Exchange Systems

Optimizing energy efficiency in software systems has become an increasingly critical concern. The ICT sector contributes significantly to global greenhouse gas emissions, with cloud computing and microservice-based architectures playing a substantial role in this energy demand. Moreover, with the advent of various data exchange technologies, data-centric multi-domain applications are growing. We address this challenge by investigating energy efficiency optimizations for data exchange archetypes within policy-aware dynamically adaptive microservice-based middleware. We introduce an energy consumption monitoring pipeline applicable to microservice-based policy-aware middleware designed to support data exchanges, such as DYNAMOS (Dynamically Adaptive Microservice-based OS). The monitoring pipeline is integrated in the system's architecture and utilizes monitoring tools to retrieve energy consumption data and statistical tools to identify which stages of the execution contributed the most to energy consumption. Using this analysis, we design and implement two optimizations that demonstrate reductions in energy usage. Both optimisations are deployed in both centralized and decentralized setups.