MODEL AND METHODOLOGY OF MEASURING AND FORECASTING OF ENERGY IMPACT OF SECURITY RISKS MITIGATION APPROACHES IN DISTRIBUTED COMPUTING SYSTEMS ON PERSONAL MOBILE DEVICES

Abstract

The rapid growth of cloud computing and artificial intelligence has drastically increased energy consumption in data centers. Volunteer computing using personal mobile devices represents a promising solution by potentially reducing energy demands. However, such systems face critical challenges related to reliability, security, and sustainability. The subject of this study is the mitigation of three principal risks – node outage, result cheating, and data fishing – within mobile volunteer computing environments, and the evaluation of their energy overhead. The goal of the research is to extend a unified model capable of forecasting energy consumption under different security mitigation strategies and to quantify their trade-offs between efficiency and trustworthiness. The methodology combines mathematical modeling, simulation, and empirical measurements conducted on networks of mobile devices. Case studies employing ray tracing workloads were used to compare baseline energy efficiency with scenarios involving node outages, malicious result injection, and data leakage attempts. Results demonstrate that each mitigation strategy entails distinct energy implications. Node outage protection introduces moderate and predictable overheads that scale with failure probability.Cheating mitigation, based on redundancy and trust evaluation mechanisms, is identified as the most energy-expensive approach, with overhead rising exponentially in adversarial environments and, in extreme cases, collapsing system performance. Data fishing mitigation produces minimal additional energy consumption but restricts task allocation, potentially leading to incomplete job execution if trusted nodes are unavailable. The scientific novelty of this work lies in the integrated framework that unifies security and reliability risk mitigation with quantitative energy modeling in distributed computing. Unlike prior studies that address risks in isolation, this research systematically compares their energy impacts, establishing the taxonomy of overhead profiles and providing a foundation for balancing efficiency and security in future large-scale distributed systems.