Comparative Analysis of Power Efficiency in Edge versus Cloud Computing Architectures

Thesis Abstract

The rapid proliferation of Internet of Things (IoT) devices and data-intensive applications has intensified the need for efficient computing architectures, prompting a comparative investigation into the power consumption profiles of edge and cloud computing systems. This study aims to evaluate and contrast the power efficiency of edge and cloud architectures, providing actionable insights for optimizing energy consumption in large-scale distributed computing environments. The specific objectives include quantifying the energy consumption of representative edge and cloud setups, identifying the key factors influencing their power efficiency, and proposing a hybrid model that leverages the advantages of both architectures to minimize overall power use. Employing a mixed-methods research design, the quantitative component involved the measurement of power consumption across two simulated environments—an edge deployment comprising 10 Raspberry Pi 4 devices and a cloud-based system utilizing a cluster of 20 virtual machines hosted on Amazon Web Services (AWS). The sample size consisted of three repeated trials for each setup, with data collected through specialized power monitoring tools such as Watts Up Pro meters and AWS CloudWatch metrics, over a continuous operational period of 72 hours to account for variability in workload demands. The qualitative component encompassed semi-structured interviews with 15 system administrators and energy management experts to explore contextual factors affecting power efficiency and to validate quantitative findings. Data analysis employed descriptive statistics to illustrate basic energy consumption patterns, followed by inferential techniques including multiple regression analysis to determine the influence of workload types and data transmission volumes on power use, and ANOVA tests to compare differences between the two architectures under similar operational conditions. Theoretical underpinning drew on the Theory of Energy Efficiency in Distributed Systems, complemented by the Technology Acceptance Model to interpret user perceptions impacting system configurations. A hybrid analytical framework combining quantitative metrics with thematic analysis of interview transcripts was adopted to generate comprehensive insights. Expected findings indicate that, under typical IoT workloads, edge computing demonstrates superior power efficiency primarily due to reduced data transmission and minimized data processing latency. However, the cloud exhibits higher energy consumption due to centralized resource intensity, especially in scenarios involving large-scale data analytics requiring substantial computational capacity. The analysis anticipates establishing a quantifiable relationship between workload characteristics and energy consumption, highlighting key factors such as data size, system uptime, and network bandwidth utilization. These findings are expected to contribute substantively to the emerging body of knowledge on sustainable computing, emphasizing context-dependent energy optimization strategies applicable across diverse operational environments. The study's main contribution lies in proposing a hybrid framework that dynamically allocates computational tasks between edge and cloud nodes to optimize overall power efficiency, informed by real-time assessment of workload demands and energy costs. By providing a detailed comparison supported by empirical data, this research informs practitioners, policymakers, and system designers seeking energy-efficient solutions in distributed computing. The study concludes with recommendations for implementing adaptive load balancing algorithms, integrating renewable energy sources, and refining energy monitoring practices to facilitate sustainable deployment of edge-cloud infrastructures. Suggested avenues for further research include longitudinal studies to assess the long-term impacts of hybrid architectures on power consumption, and investigations into emerging technologies such as fog computing and AI-enabled energy management systems.