Comparative Analysis of Energy Efficiency in Edge versus Cloud Computing Architectures

Thesis Abstract

The escalating demand for data processing and storage in contemporary information systems has prompted significant interest in evaluating the energy efficiency of emerging computing paradigms, particularly edge and cloud architectures. This study investigates and compares the energy consumption profiles of edge and cloud computing systems within a real-world urban IoT network context, aiming to inform sustainable deployment strategies. The specific objectives are to quantify the energy consumption patterns of both architectures under equivalent operational conditions, identify key factors influencing their energy efficiencies, and develop a comprehensive model to predict energy usage based on system parameters. Employing a mixed-methods approach, the research adopts an experimental quantitative design complemented by qualitative insights. The study population comprises 50 edge devices deployed in a metropolitan environmental monitoring network and 10 cloud data centers servicing the same network, with sampling stratified to ensure representative data collection across different operational loads. Data collection involves deploying specialized energy monitoring tools—such as power meters and system logs—to record real-time energy consumption over six months, ensuring coverage of peak and off-peak periods. Additionally, structured interviews with system operators are conducted to gather contextual insights into operational efficiencies and challenges. Quantitative data are analyzed using statistical techniques including descriptive statistics, t-tests, and multivariate regression analysis to compare mean energy consumptions and determine the predictors of efficiency within each architecture. A simulation-based analytical framework is further employed to model network traffic loads and evaluate their impact on energy profiles, based on the theoretical foundations of the Energy Consumption Model in distributed computing. The study leverages the Theory of Sustainable Computing to interpret system efficiency metrics and identify pathways for optimization. The anticipated findings suggest that edge computing architectures, due to their localized processing capabilities, consume significantly less energy per data processing task compared to centralized cloud systems; however, when considering aggregate network energy, the differences may vary depending on traffic volume and data aggregation strategies. Factors such as device hardware specifications, data processing complexity, and network latency are expected to substantially influence energy efficiency outcomes. The results are expected to reveal that optimized edge deployment can reduce overall energy consumption by approximately 30% compared to traditional cloud-centric approaches under similar operational conditions. The study's contribution to knowledge lies in providing a detailed empirical comparison of energy consumption patterns between edge and cloud systems, supplemented by a robust predictive model that policymakers and system designers can utilize for sustainable infrastructure planning. It bridges existing gaps by incorporating real-world operational data, diverse environmental scenarios, and a comprehensive analytical framework, thus extending the theoretical understanding of energy efficiency determinants in distributed computing. In conclusion, the research advocates for strategic integration of edge and cloud systems tailored to energy efficiency objectives, emphasizing the importance of context-aware deployment and operational optimizations. It recommends future investigations into the integration of renewable energy sources with distributed computing architectures and suggests expanding the study scope to include diverse geographical and technological settings. Overall, the findings aim to guide stakeholders toward adopting greener and more sustainable computing practices, aligning technological advancement with environmental conservation imperatives.