Towards Sustainable Computing: Exploring Energy Consumption Efficiency of Alternative Configurations and Workloads in an Open Source Messaging System

2025-06-12

AI Summary

High energy consumption of message systems in cloud environments remains a critical challenge for sustainable infrastructure. Method: This paper presents the first systematic quantification of how RabbitMQ architectural configurations and workload characteristics jointly affect energy efficiency. Leveraging a real-world deployed cluster, we design a multi-configuration, multi-workload benchmarking framework covering microservices and IoT scenarios, integrating custom monitoring, standardized load generation, and multi-dimensional power measurement tooling. We propose a reproducible methodology for evaluating message broker energy efficiency. Contribution/Results: Our analysis uncovers sensitivity patterns between key configuration parameters (e.g., persistence policies, queue topology, QoS settings) and workload features (e.g., throughput, message size, concurrency) with respect to energy consumption. Experiments achieve up to 31% power reduction. We release the first open-source, fine-grained RabbitMQ energy dataset—enabling green architecture selection, sustainable cloud cost modeling, and low-carbon infrastructure design.

Technology Category

• Machine Learning: Efficient ML / Green AI
• Data Mining & Knowledge Management: Scalability, Parallel & Distributed Systems
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• Security and Privacy: Large-scale security measurements
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Energy consumption in current large scale computing infrastructures is becoming a critical issue, especially with the growing demand for centralized systems such as cloud environments. With the advancement of microservice architectures and the Internet of Things, messaging systems have become an integral and mainstream part of modern computing infrastructures, carrying out significant workload in a majority of applications. In this paper, we describe an experimental process to explore energy-based benchmarking for RabbitMQ, one of the main open source messaging frameworks. The involved system is described, as well as required components, and setup scenarios, involving different workloads and configurations among the tests as well as messaging system use cases. Alternative architectures are investigated and compared from an energy consumption point of view, for different message rates and consumer numbers. Differences in architectural selection have been quantified and can lead to up to 31% reduction in power consumption. The resulting dataset is made publicly available and can thus prove helpful for architectures' comparison, energy-based cost modeling, and beyond.

Problem

• Assessing energy efficiency of messaging system configurations
• Comparing power consumption in alternative RabbitMQ architectures
• Quantifying energy savings up to 31% for different workloads

Innovation

• Energy-based benchmarking for RabbitMQ
• Comparing alternative architectures for efficiency
• Quantifying power consumption reduction up to 31%